Systematic Review of Hierarchical and Multi-Agent Optimization Strategies for P2P Energy Management and Electric Machines in Microgrids

Abstract

The growing complexity of distributed energy systems and the rise of peer-to-peer energy markets demand innovative solutions for efficient, resilient, and sustainable energy management. However, existing research often remains fragmented, with limited integration between control strategies, optimization frameworks, and practical implementation. This paper presents a comprehensive systematic review, following the PRISMA methodology, that synthesizes findings from 94 high-quality studies and addresses the lack of consolidated insights across technical, operational, and architectural layers. This review highlights advancements in six key areas: optimization and modeling, multi-agent systems, simulations, blockchain and smart contracts, robust frameworks, and electric machines. Despite progress, several studies reveal challenges related to scalability, data privacy, computational complexity, and system adaptability, particularly in dynamic and decentralized environments. Stochastic–robust optimization and multi-agent systems improve decentralized coordination, while blockchain enhances security and automation in peer-to-peer trading. Simulations validate energy strategies, bridging theory and practice, and electric machines support renewable integration and grid flexibility. The synthesis underscores the need for unified frameworks that combine artificial intelligence, predictive control, and secure communication protocols. This review aims to provide a roadmap for advancing distributed energy systems toward scalable, resilient, and sustainable energy solutions.

1. Introduction

The global energy sector is undergoing a profound transformation, driven by the urgent need to transition from traditional, centralized energy systems to decentralized networks dominated by Renewable Energy Sources (RES). Escalating energy demands, environmental imperatives, and the growing importance of energy resilience in the face of disruptions fuel this shift. Distributed Energy Systems (DES), which incorporate RES, energy storage technologies, and Electric Vehicles (EVs), are at the forefront of this transformation, offering opportunities to enhance energy sustainability and flexibility. Within this paradigm, Peer-to-Peer (P2P) energy trading frameworks have emerged as a revolutionary model, empowering prosumers to engage directly in energy transactions and fostering the development of collaborative energy communities [1,2]. However, as energy systems become more decentralized, they also grow in complexity [3]. The integration of heterogeneous energy resources introduces new challenges, including the variability of RES, the need for real-time coordination across spatially dispersed assets, and the importance of ensuring secure and transparent energy transactions. Multi-Agent Systems (MAS) have been widely recognized as a powerful tool for addressing these challenges, enabling decentralized coordination, scalability, and flexibility in managing energy resources. At the same time, advancements in blockchain technology and smart contracts promise enhanced security and automation for P2P trading, while electric machines are essential in supporting renewable integration and grid flexibility [4,5,6]. Despite significant progress in these areas, the intricate interplay between optimization frameworks, robust control mechanisms, and the role of electric machines in decentralized energy systems remain underexplored. The dynamic nature of P2P energy trading, coupled with the operational and technical constraints of modern DES, underscores the necessity for a comprehensive synthesis of existing knowledge. This study is motivated by the need to consolidate advancements in hierarchical optimization, MAS-based control, blockchain integration, and the application of electric machines. By providing a systematic review of these inter-related topics, this work seeks to advance the development of resilient, scalable, and efficient energy management solutions that align with the evolving demands of decentralized energy systems.

Distributed energy systems have shown significant potential in mobile applications, particularly in the integration of RES, energy storage technologies, and EVs. Recent studies highlight the advancements in energy management strategies for hybrid electric vehicles, where the optimization of fuel cell health and battery thermal management plays a critical role. For instance, ref. [7] proposes a novel cost-minimization Energy Management Strategy (EMS) for hydrogen Fuel Cell Hybrid Electric Vehicles (FCHEVs), integrating thermal safety and degradation awareness for on-board battery systems, which reduces battery aging by 34.8% and total operating costs by 12.3%. Similarly, ref. [8] introduces a multi-agent approach with verifiable and data-sovereign information flows, enabling decentralized redispatch in DES while ensuring privacy and scalability through digital Self-Sovereign Identities (SSIs) and Zero-Knowledge-Proofs (ZKPs). This approach addresses the challenges of integrating small-scale assets, such as photovoltaic systems and EVs, into grid stabilization processes. The optimization of EV charging scheduling has also been a focal point in enhancing the efficiency of regional clean energy power supply networks. Ref. [9] demonstrates how intelligent algorithms and machine learning-based demand forecasting can reduce the peak grid load by 15%, increase clean energy consumption by 23%, and lower total electricity consumption by 10%. Furthermore, ref. [10] presents an integrative lifecycle design approach for renewable–battery–consumer energy systems, optimizing the photovoltaic capacity and battery storage to achieve a higher net present value and lower carbon intensity, thereby promoting sustainable development and carbon neutrality.

In the context of distributed energy management systems, ref. [11] leverages edge computing and machine learning to improve energy allocation, reduce energy waste by 18%, and enhance the system response speed by 30%. This approach highlights the importance of the real-time data processing and dynamic scheduling of distributed energy resources. Additionally, ref. [12] proposes a low-carbon planning method for flexible distribution networks, incorporating EV orderly charging and discharging strategies to minimize comprehensive operating costs and improve grid stability. Similarly, ref. [13] addresses power management challenges in Hybrid Microgrids (HMGs) by introducing a 3-D droop-based control scheme for bidirectional DC-DC converters, ensuring accurate power sharing and the efficient operation of EVs within HMGs.

The development of DES has steadily evolved over the past decade, driven by the imperative to create sustainable, efficient, and resilient energy networks. Early foundational work in 2014 laid the groundwork for integrating RES into distributed networks, focusing on optimizing energy flows to minimize losses and address the inherent variability of RES. These studies emphasized hierarchical control frameworks and robust optimization methods to manage uncertainties in distributed energy environments, marking a significant step toward modern energy management systems [1]. This early emphasis on robustness and efficiency provided a blueprint for future innovations. In subsequent years, the concept of MAS emerged as a transformative approach to managing distributed resources. Research in 2017 demonstrated how MAS frameworks could enable decentralized coordination and dynamic energy trading in P2P systems. This work highlighted MAS’s ability to balance supply and demand autonomously while addressing scalability challenges in increasingly complex energy networks [14]. Around the same time, blockchain technology began to be explored as a solution to enhance the security and transparency of energy transactions. A notable study in 2017 introduced blockchain-based systems for P2P energy trading, emphasizing the elimination of intermediaries and the secure automation of transactions, thus addressing critical concerns of trust and efficiency in decentralized markets [15]. By 2019, optimization models had evolved to address the variability and unpredictability inherent in RES. One prominent study employed a hybrid robust Conditional Value at Risk (CVaR) framework to optimize energy allocation and reduce operational costs under uncertain conditions. This approach demonstrated the potential of stochastic optimization in balancing financial and operational risks, paving the way for more sophisticated energy management models [16]. Simultaneously, MAS applications were refined to improve their scalability and adaptability. Clustering techniques were introduced, allowing MAS frameworks to coordinate large-scale networks of prosumers more efficiently by grouping similar agents and reducing computational complexity [2].

The integration of hierarchical scheduling frameworks into decentralized energy systems gained momentum in 2020. These systems provided innovative solutions for aggregating demand-side flexibility and optimizing energy trading among interconnected microgrids. A landmark study that year demonstrated how hierarchical scheduling could manage Thermostatically Controlled Loads (TCLs) as virtual batteries, significantly improving energy efficiency and enabling privacy-preserving trading mechanisms [17]. Concurrently, the role of EVs within DES became a focal point. Research highlighted the dual functionality of EVs as mobile energy storage units and contributors to Vehicle-to-Grid (V2G) operations, showcasing their potential to enhance grid stability and integrate renewable energy more effectively [18]. From 2021 onward, the integration of advanced Artificial Intelligence (AI) techniques with MAS frameworks marked a turning point in the field. Deep Reinforcement Learning (DRL) and federated learning frameworks were introduced, enabling decentralized systems to make real-time decisions under highly dynamic conditions. For example, a 2022 study employed multi-agent DRL to optimize P2P energy trading strategies, achieving substantial cost reductions and improved resource utilization compared to traditional optimization techniques [5]. At the same time, blockchain technology evolved to include smart contracts, automating energy transactions and ensuring compliance with regulatory frameworks. These advancements highlighted the potential of combining AI with blockchain to address the dual challenges of scalability and transparency in decentralized energy systems [19]. The role of electric machines within DES also gained significant attention during that period. Some research has explored the integration of advanced control algorithms with synchronous and reluctance generators to stabilize microgrid operations. A study in 2022 demonstrated how advanced droop-controlled inverters interfaced with LC filters could maintain the voltage and frequency stability during P2P energy trading, showcasing the critical role of electric machines in ensuring operational reliability [20]. Recent advancements in 2023 and 2024 reflect the convergence of MAS, blockchain, and robust optimization frameworks. A 2023 study introduced a hierarchical blockchain model integrated with MAS to enhance the scalability and security of microgrid operations. This framework employed innovative trust-based consensus protocols to address privacy concerns and ensure robust energy management under uncertain conditions [6]. Simultaneously, a 2024 study examined the application of multi-cluster DRL to coordinate P2P energy trading in large-scale networks. This approach demonstrated a superior performance in handling the heterogeneity of prosumers and optimizing energy allocation [6]. In addition, the integration of EVs with Vehicle-to-Everything (V2X) frameworks has emerged as a promising area of research. A recent study demonstrated how V2X systems could optimize EV battery usage for dynamic energy exchanges, enhancing the flexibility and resilience of distributed networks while ensuring safe operational limits for EVs [21].

While research on decentralized energy systems and P2P energy trading has advanced considerably, significant gaps remain within the identified thematic areas, highlighting opportunities for further investigation. In the domain of optimization and modeling, most existing studies rely on static or deterministic approaches that often fail to account for the dynamic and stochastic nature of RES and demand fluctuations in distributed networks. Although robust models have been proposed to manage uncertainties, such as those leveraging CVaR to optimize energy allocation [16], they are typically constrained by idealized assumptions and lack the real-time adaptability required for practical deployment in P2P trading environments. The integration of co-optimization strategies that simultaneously address economic, environmental, and operational objectives remains an underexplored area. In the realm of MAS and distributed control, substantial progress has been made in enabling decentralized coordination and flexibility in energy management. However, existing MAS frameworks often face scalability limitations when applied to large, heterogeneous networks. Studies have demonstrated the effectiveness of MAS in small-scale applications [14], but the absence of advanced coordination mechanisms capable of balancing the computational efficiency with the complexity of managing prosumer communities hampers their effectiveness in real-world applications. Furthermore, the integration of MAS with emerging technologies, such as blockchain and AI, has been limited. Few studies, such as [5], explore how hybrid approaches combining these technologies could enhance the scalability, security, and adaptability of P2P energy trading frameworks.

Simulations, case studies, and real-world applications have provided valuable insights into validating theoretical models and exploring practical challenges in decentralized energy systems. However, a critical evaluation of the existing literature reveals significant gaps that hinder the scalability and practical deployment of P2P energy trading frameworks. While many studies have demonstrated the feasibility of P2P systems in small-scale or controlled environments, such as isolated microgrids [17], they often fail to address the socio-economic and regulatory complexities that arise in larger, more diverse deployments. This limitation underscores the need for research that considers the broader context of P2P systems, including the integration of heterogeneous energy resources and the impact of varying market conditions.

Blockchain technology and smart contracts have emerged as promising tools for enabling secure and transparent energy transactions. However, their adoption in resource-constrained environments, such as microgrids, is hampered by high computational demands and energy-intensive consensus mechanisms. Although hybrid blockchain models have been proposed to mitigate these challenges [6], their practical implementation remains in its infancy, with the limited exploration of optimization strategies tailored to low-power, distributed networks. Similarly, while advanced communication tools and AI have shown potential for predictive optimization in P2P energy markets [19], their integration into real-time energy management systems is still underdeveloped, leaving a gap in leveraging these technologies for dynamic decision making.

Frameworks for robust and decentralized energy management have made notable progress in recent years, particularly in addressing resilience and adaptability under uncertainty. However, many of these frameworks, such as those presented in [2], remain largely theoretical, focusing on idealized scenarios rather than the practical challenges of dynamic energy resource allocation in highly distributed systems. For instance, while EVs have been recognized for their dual functionality as energy storage units and grid stabilizers in V2G and V2X systems [18], their broader integration into energy management frameworks is still underexplored. This highlights a critical gap in the literature, as the coordination of diverse resources, such as EVs, is essential for achieving scalable and resilient decentralized energy systems. Furthermore, the role of electric machines in DES is a relatively nascent area of research. While studies such as [20] have emphasized their importance in stabilizing microgrid operations and enhancing system flexibility, the integration of electric machines into P2P energy trading and decentralized markets remains poorly understood. This represents a significant research gap, as electric machines play a crucial role in energy generation, storage, and conversion, particularly in systems dominated by renewable energy sources.

To address these critical gaps, this paper provides a systematic review of advancements in decentralized energy systems, with a specific focus on P2P energy trading frameworks. The main challenge in this field lies in the lack of a holistic approach that integrates the diverse components of decentralized energy systems—such as optimization models, multi-agent systems, blockchain technologies, and electric machines—into a cohesive framework capable of addressing real-world complexities. Previous works have often focused on isolated aspects of these systems, such as optimizing energy allocation in small-scale microgrids or developing theoretical frameworks for blockchain-based P2P trading. However, they have largely failed to address the interdependencies between these components, particularly in large-scale, dynamic environments where socio-economic, regulatory, and technical factors play a significant role.

For instance, while robust optimization models have been proposed to manage uncertainties in energy allocation [9], they often rely on idealized assumptions and lack the real-time adaptability required for practical deployment in P2P trading environments. Similarly, although blockchain and smart contracts have shown promise for secure and transparent energy transactions, their computational demands and energy-intensive consensus mechanisms limit their applicability in resource-constrained settings like microgrids [6]. Furthermore, the integration of electric machines—critical for renewable energy integration and grid stability—has been largely overlooked in the context of P2P energy trading, despite their potential to enhance operational flexibility and resilience [13].

By consolidating findings from 94 high-quality studies across six thematic areas—optimization and modeling in energy systems, MAS and distributed control, simulations and real-world applications, blockchain and emerging technologies, frameworks for robust energy management, and the role of electric machines—this review not only identifies key trends and challenges, but also highlights the underexplored synergies between these areas. Specifically, it emphasizes the integration of electric machines into decentralized systems, a critical yet neglected aspect that can significantly enhance renewable energy integration, grid stability, and operational flexibility. By bridging these knowledge gaps, this work lays the foundation for developing scalable, resilient, and efficient energy management solutions tailored to modern decentralized networks, addressing the limitations of previous studies and paving the way for future research and practical implementations.

This paper is organized as follows: Section 2 details the review methodology, Section 3 presents results, Section 4 presents the discussion of the paper, and Section 5 concludes with a summary and future research directions.

2. Methodology for Selection of Studies

2.1. Methodology: PRISMA Framework

The PRISMA methodology (Preferred Reporting Items for Systematic Reviews and Meta-Analyses), updated in 2021 [22], is used for conducting the literature review documented in this manuscript. PRISMA ensures transparency, reproducibility, and a high standard of academic rigor by guiding researchers through four distinct phases:

• Identification: An exhaustive search is conducted across multiple databases to compile a comprehensive pool of studies potentially relevant to the research objectives. This step involves designing detailed search strategies tailored to the scope of the review and employing bibliographic tools to manage and organize retrieved records.
• Screening: Duplicate records are identified and removed, and studies are evaluated based on predefined inclusion criteria through title and abstract reviews. This phase narrows the dataset to studies most likely to meet the review’s objectives while maintaining transparency and consistency in the selection process.
• Eligibility and Inclusion: A detailed full-text evaluation of the remaining studies is performed. Each study is assessed using specific criteria for relevance, methodological rigor, innovation, data quality, and impact. Studies meeting the defined threshold for quality and alignment are included in the synthesis. This phase ensures that only high-quality, pertinent studies are retained for the review.
• Synthesis: The selected studies undergo comprehensive analysis to integrate their findings, methodologies, and contributions. This phase involves summarizing, comparing, and synthesizing evidence to address the research questions and provide a cohesive understanding of the topic.

Figure 1 illustrates the roadmap for implementing the PRISMA methodology in the systematic literature review process. A detailed flowchart outlining the entire procedure for narrowing down articles from the initially identified items in prestigious digital databases is presented in Appendix A, Figure A1.

2.2. Identification of the Studies

The identification phase constitutes the cornerstone of the systematic review process, ensuring that the literature aligns comprehensively with the research scope. For this study, the timeframe was set to encompass publications from the last decade (2014–2024), including peer-reviewed journal articles and conference proceedings available in English and with full-text access. Publications such as editorials, review articles, letters, short communications, and policy briefs were excluded as they did not meet the requirements for addressing the study’s specific research questions.

The literature search utilized two prominent bibliographic databases: Scopus and Web of Science (WoS). Both platforms are renowned for their extensive and curated repositories of peer-reviewed literature:

• Scopus: Offers advanced search functionalities, comprehensive indexing, and analytical tools for tracking trends and citations. Its content spans multiple disciplines and includes high-impact publishers like Elsevier, Springer, Wiley, and MDPI. The search string used for Scopus was as follows: TITLE-ABS-KEY (“Energy management”) AND (“Hierarchical” OR “Multi-agent” OR “Robust”) AND (“peer to peer”)) AND PUBYEAR > 2013 AND PUBYEAR < 2025 AND (LIMIT-TO (DOCTYPE, “ar”) OR LIMIT-TO (DOCTYPE, “cp”)) AND (LIMIT-TO (LANGUAGE, “English”)).
• Web of Science: Known for its rigorous curation, WoS provides a multidisciplinary collection of high-quality publications, with robust citation tracking and analysis capabilities. For Web of Science, the following query was applied: ALL = (Energy management) AND (ALL = (Hierarchical) OR ALL = (Multi-agent) OR ALL = (Robust)) AND (ALL = (peer to peer)).

The search queries were designed meticulously to align with the review’s objectives. Refinements were applied to both queries to limit results by publication year (2014–2024), document type (articles and conference proceedings), and language (English). The initial search yielded 82 records from Scopus and 215 records from Web of Science.

Given the substantial overlap between Scopus and WoS, duplicate entries were anticipated. To address this, all retrieved records were imported into a bibliographic management tool for deduplication. This process identified and removed 53 duplicate records, resulting in a final dataset of 244 unique entries.

This refined dataset provides a robust foundation for the subsequent Screening, Eligibility, and Inclusion phases.

2.3. Screening of the Identified Studies

The screening phase served to refine the initial dataset obtained during the identification stage, ensuring that only studies meeting the strict inclusion criteria were retained for further evaluation. This process involved a systematic review of titles and abstracts to assess each record’s relevance and alignment with the research objectives. The inclusion criteria applied during this stage were as follows:

• Publication Year: Studies published between 2014 and 2024 were included, capturing advancements over the past decade while maintaining a focus on recent developments. This time frame provided a balanced view of contemporary research trends and innovations in the field.
• Document Type: Only original research articles published in peer-reviewed journals or presented at conferences were considered. Works such as review articles, editorials, letters, opinion pieces, policy briefs, and other non-empirical contributions were excluded. This criterion ensured the inclusion of studies offering novel insights, experimental findings, or robust analytical frameworks directly relevant to the research objectives.
• Language: Articles written in English were selected, reflecting the predominant language of international scientific discourse. This approach enhanced the comparability of studies and ensured a broad representation of globally recognized research.
• Full-Text Availability: Only studies with accessible full texts, either through institutional subscriptions or open-access platforms, were included. This criterion allowed for an in-depth analysis of methodologies, findings, and conclusions, which was critical for maintaining the rigor of the systematic review.
• Thematic Relevance: Each study was evaluated for alignment with the review’s scope and objectives. Only research explicitly addressing energy management within hierarchical, multi-agent, or robust systems—particularly in the context of P2P energy trading—was retained. This ensured that the review focused on the literature directly contributing to the understanding and resolution of the research problem.

Records failing to meet all of the above criteria were excluded. To ensure consistency and reduce potential biases, the screening process was conducted independently by two reviewers. Any discrepancies or disagreements were resolved through discussion and consensus.

During the screening phase, each study underwent a binary evaluation process, focusing on the review of titles and abstracts to ensure compliance with the predefined inclusion criteria. Of the initial 244 records identified, 182 studies (74.9%) satisfied all criteria and were deemed relevant to the research topic. These criteria included the publication year, document type, language, full-text availability, and thematic relevance, ensuring that the selection of high-quality studies aligned with the research objectives. Figure 2 provides a statistical overview of the results obtained during the Screening phase.

According to Figure 2, the 182 selected studies were distributed across two sources: 74 records originated from Scopus, while 108 were retrieved from WoS. This distribution emphasizes the complementary coverage of these databases and their collective value in ensuring a comprehensive and diverse dataset. The temporal analysis of the selected studies revealed a clear upward trend in the research output over the years. Between 2014 and 2018, the number of publications was relatively modest, with incremental growth: 1 study in 2014, 3 in 2015, 8 in 2017, and 4 in 2018. Beginning in 2019, a noticeable increase emerged, with 10 studies published that year. This growth accelerated further in subsequent years, with 14 publications in 2020, 24 in 2021, and 35 in 2022. The trend peaked in 2023 and 2024, which recorded 34 and 49 studies, respectively. This trajectory reflects the rising academic interest and recognition of the significance of this research domain in recent years.

These 182 studies will proceed to the next phase, Eligibility and Inclusion, where a more detailed evaluation will confirm their suitability for final analysis and synthesis. This additional scrutiny aims to refine the dataset further, ensuring that only the most relevant and methodologically robust studies contribute to the systematic review.

2.4. Eligibility and Inclusion of Screened Manuscripts

The Eligibility and Inclusion phase is vital in refining the selection process to ensure that only the most relevant and high-quality studies were included in this systematic review. Each of the 182 articles was carefully assessed through an in-depth full-text evaluation, utilizing a detailed three-level Likert scale based on the following criteria specifically tailored to the objectives of this investigation:

• Eligibility Criterion 1—Alignment with Research Objectives: This criterion evaluates how well this study aligns with the research scope, particularly its focus on hierarchical, multi-agent, or robust optimization approaches in energy management within P2P frameworks. Studies addressing the role of electric machines in DES or their integration with RES are prioritized (Scoring: 1—Peripheral, 2—Related, 3—Highly Relevant).
• Eligibility Criterion 2—Methodological Rigor: This criterion assesses the robustness and appropriateness of this study’s methodology, including experimental setups, simulation models, validation frameworks, and the overall clarity of the research process. The methods must provide sufficient evidence to support the study’s findings (Scoring: 1—Needs Improvement, 2—Acceptable, 3—Strong).
• Eligibility Criterion 3—Originality and Innovation: This criterion evaluates the novelty of the proposed strategies or frameworks in this study. This includes innovative approaches to multi-agent optimization, robust control mechanisms, or the role of electric machines (e.g., SiC- or GaN-based technologies) in enhancing energy efficiency and reliability in P2P systems (Scoring: 1—Minor, 2—Moderate, 3—Major).
• Eligibility Criterion 4—Data Quality and Analysis: This criterion measures the reliability, reproducibility, and depth of the data analysis presented in this study. High-scoring studies provide detailed, well-documented datasets and analyses, including energy flow modeling, the machine performance, or system-level optimization outcomes in P2P environments (Scoring: 1—Satisfactory, 2—Good, 3—Excellent).
• Eligibility Criterion 5—Impact and Contribution to the Field: This criterion evaluates this study’s potential contribution to advancing knowledge in the domain of distributed energy management. High-impact studies are those that propose transformative insights, address significant challenges, or establish new benchmarks in energy management and electric machine applications measured by the number of citations (Scoring: 1—Low impact, 2—Moderate impact, 3—High impact).

The Eligibility and Inclusion phase marked a critical step in refining the dataset, ensuring that only the most relevant, methodologically robust, and innovative studies were selected for synthesis. Two independent researchers conducted a thorough full-text evaluation of the 182 articles identified during the screening phase. To minimize bias, discrepancies in the evaluations were resolved through discussion and consensus. The threshold for inclusion was set at 12 out of 16 points, striking a balance between selectivity and inclusivity. This threshold was carefully chosen to ensure that selected studies met the high standards of relevance, rigor, and innovation required for the systematic review, while remaining inclusive enough to capture valuable contributions. Figure 3 presents an excerpt from the verification matrix completed by reviewers to evaluate each article fully.

As can be evidenced in Figure 3, out of the 182 articles evaluated, 94 studies (51.6%) met or exceeded the threshold of 12 points and were included in the final dataset for synthesis. These articles demonstrated a strong alignment with the review’s objectives, provided robust methodological approaches, and contributed innovative and impactful findings to the field. Studies that did not meet the threshold were not dismissed as inherently poor or lacking merit. Instead, their exclusion reflects the need to focus on works most pertinent to the specific objectives of this systematic review.

The 94 selected studies will proceed to the synthesis phase, where their findings, methodologies, and contributions will be analyzed and integrated to provide a comprehensive understanding of hierarchical, multi-agent, and robust energy management in P2P frameworks.

2.5. Synthesis of the Selected Studies

The 94 studies selected for synthesis form a robust foundation for addressing the research objectives. Figure 4 provides a detailed visual representation of the bibliometric analysis, including trends in the publication volume, journal/conference distributions, and other relevant metrics. The cumulative H-index of 30 for the selected studies reflects their significant influence in advancing the understanding of energy management systems within P2P frameworks. The high citation rates observed for many works within this dataset suggest that these studies are widely recognized and can serve as foundational references for subsequent research in this area. The temporal distribution of the selected studies (see Figure 4) reveals clear trends in the research activity over the past decade. This analysis highlights a steady increase in academic interest, particularly from 2021 onwards, coinciding with the growing emphasis on renewable energy integration, DES, and P2P energy trading. The peak in 2022, with 26 studies, means a heightened focus on these topics, potentially driven by advancements in multi-agent optimization and robust control techniques.

The selected studies were published in a diverse range of reputable journals and conferences, showing the multidisciplinary nature of the research domain. Figure 4 also summarizes the distribution of studies by publication venue.

This distribution indicates that the field is well-represented in high-impact journals and conferences, showcasing its multidisciplinary nature and broad applicability. Key contributions come from venues such as IEEE Transactions on Smart Grid, which has led 11 studies and emphasizes its role as a premier platform for cutting-edge research in smart grid technologies, hierarchical optimization, and distributed energy management. Applied Energy follows with 10 contributions, reflecting its strong focus on sustainability and energy efficiency, particularly through practical and innovative approaches to energy system optimization. IEEE Access, with nine studies, highlights emerging technologies such as blockchain, smart contracts, and MAS, benefiting from its open-access model to enhance visibility and interdisciplinary collaboration. The International Journal of Electrical Power and Energy Systems, featuring five studies, emphasizes advanced optimization models and technical–economic perspectives on power systems, particularly in decentralized contexts. IEEE Transactions on Power Systems, contributing four studies, reveals its focus on power system operations and planning, addressing challenges in distributed generation, energy storage, and robust control frameworks. Energies and IEEE Transactions on Industrial Informatics, each contributing three studies, demonstrate complementary focuses, with Energies bridging theoretical advancements and practical implementations in renewable energy systems. At the same time, IEEE Transactions on Industrial Informatics delve into the intersection of industrial applications and informatics, particularly in smart grid technologies and distributed control. Collectively, these seven journals account for a substantial portion of the selected studies, reflecting the prominence and diversity of hierarchical, multi-agent, and robust optimization systems in modern energy research. The presence of a significant number of studies in diverse venues highlights the multidisciplinary scope and broad applicability of research in hierarchical, multi-agent, and robust optimization systems for P2P energy management.

Figure 5 provides a word cloud map constructed from the authors’ keywords across the 94 selected studies by using an online free tool: www.freewordcloudgenerator.com (accessed on 25 January 2025). This word cloud highlights the frequency of recurring terms, offering valuable insights into the central themes and research trends within the dataset. The size of each term in the cloud corresponds to its frequency of appearance, reflecting the predominant topics in the reviewed studies.

To guide the synthesis of the findings and uncover the structure of current research in the field, the selected studies were classified into six inter-related thematic areas. These categories emerged through a combination of word cloud analysis and conceptual alignment with the research questions of this review. Each theme addresses a key dimension of distributed energy systems and, collectively, they form an integrated narrative that spans from modeling and control to implementation and supporting technologies. Optimization and modeling provide the analytical foundation for system operation; multi-agent systems enable the decentralized execution of these models; simulations and case studies validate their practical relevance; emerging technologies such as blockchain and AI facilitate secure and automated system interactions; robust frameworks ensure reliability under uncertainty; and electric machines embody the physical components driving energy conversion and exchange. Together, these themes capture the multi-layered complexity of peer-to-peer energy management in microgrids and offer a coherent lens through which the reviewed literature is interpreted.

The analysis of the word cloud enables the classification of the 94 studies into the following six broad thematic areas, each capturing a critical aspect of the research landscape:

• Optimization and Modeling in Energy Systems: This theme encompasses studies focused on enhancing the efficiency and sustainability of energy systems through mathematical models, advanced algorithms, and optimization techniques. Research within this category addresses co-optimization strategies for distributed energy resources, energy planning methodologies, and robust analyses to manage uncertainties in distributed generation, microgrids, and P2P energy trading.
• Multi-Agent Systems and Distributed Control: Studies in this category explore the use of MAS and distributed control approaches for energy management. Key aspects include scalability and flexibility for coordinating energy trading in complex networks, such as prosumer communities and microgrids, and decentralized cooperation to enhance resilience in energy systems.
• Simulations, Case Studies, and Real-World Applications: This theme brings together studies that validate energy management strategies and trading models through simulations to explore theoretical frameworks, case studies, and practical implementations addressing real-world challenges, experimental evidence, and insights into applying P2P systems and microgrid technologies.
• Blockchain, Smart Contracts, and Emerging Technologies: Research in this category examines how emerging technologies enhance the functionality of energy systems. Key technologies include blockchain and smart contracts for secure, transparent, and automated transactions, AI and advanced communication tools for system optimization, applications in P2P energy markets, and prosumer networks.
• Frameworks for Robust and Decentralized Energy Management: This category highlights the development of conceptual and methodological frameworks to improve the stability and flexibility of DES under uncertainty. Studies emphasize the decentralized coordination of energy resources and robust strategies to adapt to dynamic changes in the system operation, enhancing resilience against disruptions.
• Electric Machines and Their Role in Distributed Energy Systems: Unlike the previous themes derived directly from the word cloud analysis, this category emerges from the specific focus of this systematic review. Based on the eight studies addressing this topic, the role of electric machines in DES is emphasized. Key contributions include their integration in energy generation, storage, and conversion; applications in microgrids, nanogrids, and decentralized energy markets; and the exploration of renewable energy integration, system stability, and operational flexibility enabled by electric machines, including wind generators, synchronous machines, and reluctance generators.

Finally,

Table A1. Summaries of the main findings of the 94 studies selected through the systematic review process.

N°	ID	Ref.	Title	Author	Year	Findings
1	S-035	[1]	Day-ahead scheduling of 100% isolated communities under uncertainties through a novel stochastic–robust model	Tostado-Véliz, M. et al.	2022	Developing of a three-stage day-ahead scheduling strategy for isolated 100% energy communities, involving P2P transactions among prosumers.
2	S-073	[6]	Scalable coordinated management of peer-to-peer energy trading: A multi-cluster deep reinforcement learning approach	Qiu, D. et al.	2021	Design of a hierarchical peer-to-peer energy trading mechanism with coordinated management and distributed optimization.
3	WoS-071	[29]	Optimal participation of prosumers in energy communities through a novel stochastic–robust day-ahead scheduling model	Tostado-Véliz, M. et al.	2023	Optimization of prosumer participation in local energy communities under uncertainty using robust multi-objective strategies.
4	S-012	[65]	Design and management of a distributed hybrid energy system through smart contract and blockchain	Li, Y. et al.	2019	Implementation of a distributed hybrid energy system with intelligent agents to enhance resiliency and operational flexibility.
5	S-019	[50]	An electric power trading framework for smart residential community in smart cities	Hanumantha, R.	2019	Development of a blockchain-enabled framework for electricity trading in smart grids using auction-based mechanisms.
6	S-034	[42]	Federated reinforcement learning for smart building joint peer-to-peer energy and carbon allowance trading	Qiu, D. et al.	2023	Application of federated reinforcement learning to optimize energy consumption and trading decisions in smart buildings.
7	S-041	[66]	Towards sustainable smart cities: A secure and scalable trading system for residential homes using blockchain and artificial intelligence	Samuel, O. et al.	2022	Proposal of a secure and decentralized P2P energy trading platform leveraging blockchain for smart city applications.
8	S-055	[62]	A Scalable Privacy-Preserving Multi-Agent Deep Reinforcement Learning Approach for Large-Scale Peer-to-Peer Transactive Energy Trading	Ye, Y. et al.	2021	Proposal of a novel P2P transactive trading scheme based on the MAAC algorithm, which addresses some typical technical challenges.
9	S-063	[84]	Distributed peer-to-peer energy trading for residential fuel cell combined heat and power systems	Nguyen, D.	2021	Design of a real-time, distributed P2P energy trading platform tailored for remote microgrids with renewable integration.
10	WoS-043	[53]	An optimal scheduling strategy for peer-to-peer trading in interconnected microgrids based on RO and Nash bargaining	Wei, C. et al.	2021	Formulation of an optimal scheduling model for peer-to-peer electricity trading using mixed-integer programming and game theory.
11	WoS-058	[77]	A Novel Locality Algorithm and Peer-to-Peer Communication Infrastructure for Optimizing Network Performance in Smart Microgrids	Marzal, S. et al.	2017	Development of a locality-based P2P communication model for enhanced performance in smart grid energy transactions.
12	WoS-096	[5]	Multi-Agent Deep Reinforcement Learning for Coordinated Energy Trading and Flexibility Services Provision in Local Electricity Markets	Ye, Y. et al.	2023	Use of multi-agent deep reinforcement learning for optimal P2P energy trading coordination in local electricity markets.
13	WoS-112	[17]	Hierarchical Scheduling of Aggregated TCL Flexibility for Transactive Energy in Power Systems	Song, M. et al.	2020	Design of a hierarchical scheduling method for aggregating TCL flexibility to enable efficient P2P market participation.
14	WoS-140	[67]	Blockchain-Based Decentralized Virtual Power Plants of Small Prosumers	Cioara, T. et al.	2021	Proposal of a decentralized virtual power plant system using blockchain and smart contracts for secure energy trading.
15	WoS-168	[64]	A Novel Discounted Min-Consensus Algorithm for Optimal Electrical Power Trading in Grid-Connected DC Microgrids	Xu, Y.	2019	Introduction of a novel min-consensus algorithm for optimal matching of supply and demand in P2P electricity trading.
16	WoS-105	[56]	Secondary Control for Storage Power Converters in Isolated Nanogrids to Allow Peer-to-Peer Power Sharing	González-Romera, E. et al.	2020	Design of a secondary control strategy for storage converters in DC microgrids participating in P2P energy markets.
17	S-001	[68]	Blockchain-based smart contract for energy demand management	Wang, X. et al.	2019	Implementation of a blockchain-based smart contract model for efficient demand-side energy management in P2P networks.
18	S-007	[52]	Hierarchical Energy Management in Islanded Networked Microgrids	Hong, Y.	2022	Development of a hierarchical energy management strategy for islanded networks involving distributed energy trading.
19	S-008	[54]	Customized decentralized autonomous organization based optimal energy management for smart buildings	Ding, Y. et al.	2024	Introduction of a customized decentralized autonomous organization (DAO) structure for managing P2P energy markets.
20	S-013	[55]	Coordination of commercial prosumers with distributed demand-side flexibility in energy sharing and management system	Zheng, S. et al.	2022	Proposal of a distributed coordination strategy for commercial prosumers using Nash equilibrium in energy trading.
21	S-015	[26]	Aggregator-Network Coordinated Peer-to-Peer Multi-Energy Trading via Adaptive Robust Stochastic Optimization	Zou, Y.	2024	Proposal of a coordinated P2P trading mechanism using aggregator–network interaction to enhance market efficiency.
22	S-017	[81]	A hierarchical and decentralized energy management system for peer-to-peer energy trading	Elkazaz, M.	2021	Development of a hierarchical and decentralized energy management system enabling P2P trading in AC/DC hybrid microgrids.
23	S-033	[61]	Distribution loss allocation in peer-to-peer energy trading in a network of microgrids	Bai, L.	2020	Design of a distribution loss allocation method tailored for P2P energy markets to ensure fair cost distribution.
24	S-045	[2]	Multi-Agent Based Optimal Scheduling and Trading for Multi-Microgrids Integrated with Urban Transportation Networks	Liu, Y. et al.	2021	Formulation of a multi-agent-based scheduling and trading model for optimal energy management in community microgrids.
25	S-046	[51]	Mean-Field Multi-Agent Reinforcement Learning for Peer-to-Peer Multi-Energy Trading	Qiu, D. et al.	2023	Implementation of a mean-field multi-agent reinforcement learning model for scalable optimization in P2P electricity trading.
26	S-048	[25]	A distributed robust ADMM-based model for the energy management in local energy communities	Khojasteh, M.	2023	Design of a robust distributed optimization model using ADMM for dynamic P2P market clearing under uncertainty.
27	S-051	[96]	Multi-Agent Reinforcement Learning for Automated Peer-to-Peer Energy Trading in Double-Side Auction Market	Qiu, D. et al.	2021	Application of multi-agent reinforcement learning to automate P2P electricity trading and improve decision-making efficiency.
28	S-052	[85]	Coordinated management of aggregated electric vehicles and thermostatically controlled loads in hierarchical energy systems	Liu, G. et al.	2021	Proposal of a coordinated charging/discharging scheme for aggregated EV fleets participating in P2P energy trading.
29	S-058	[72]	Towards the detailed modeling of deregulated electricity markets comprising Smart prosumers and peer to peer energy trading	Symiakakis, M.	2023	Development of a modeling framework for analyzing deregulated electricity markets including P2P trading dynamics.
30	S-059	[97]	Data-Driven Distributionally Robust Co-Optimization of P2P Energy Trading and Network Operation for Interconnected Microgrids	Li, J. et al.	2021	Introduction of a data-driven co-optimization model using distributionally robust techniques for integrated P2P trading and reserve allocation.
31	S-064	[47]	Multi-agent energy management of smart islands using primal-dual method of multipliers	Mohamed, M.	2020	Development of a multi-agent based energy management framework for smart islands promoting sustainability and resiliency.
32	S-071	[98]	Scalable multi-agent reinforcement learning for distributed control of residential energy flexibility	Charbonnier, F.	2022	Proposal of a scalable multi-agent reinforcement learning strategy to manage distributed energy resources in smart grids.
33	S-077	[14]	Distributed multi-agent based coordinated power management and control strategy for microgrids with distributed energy resources	Rahman, M.	2017	Design of a distributed coordinated power flow control system using multi-agent systems in P2P microgrids.
34	WoS-003	[45]	Multi-Agent Microgrid Management System for Single-Board Computers: A Case Study on Peer-to-Peer Energy Trading	Comes, L.	2020	Implementation of a multi-agent microgrid system for intelligent management of single and interconnected communities.
35	WoS-028	[99]	A Two-Tier Distributed Market Clearing Scheme for Peer-to-Peer Energy Sharing in Smart Grid	Ullah, M.	2022	Design of a two-tier distributed market clearing scheme supporting P2P energy trading in interconnected microgrids.
36	WoS-037	[16]	Impact assessment of shared storage and peer-to-peer trading on industrial buildings in the presence of electric vehicle parking lots: A hybrid robust-CVaR analysis	Qiu, W.	2024	Assessment of the impact of shared energy storage systems on efficiency and fairness in P2P energy trading.
37	WoS-048	[91]	Deep reinforcement learning for energy trading and load scheduling in residential peer-to-peer energy trading market	Wang, J.	2023	Application of deep reinforcement learning for efficient energy trading among multiple agents in smart grid environments.
38	WoS-049	[49]	Real-Time Multi-Agent Based Power Management of Virtually Integrated Microgrids Comprising Prosumers of Plug-in Electric Vehicles and Renewable Energy Sources	Sifakis, N.	2024	Real-time power management of community microgrids using multi-agent systems for adaptive and autonomous operation.
39	WoS-061	[19]	Hierarchical Blockchain Design for Distributed Control and Energy Trading Within Microgrids	Yang, J. et al.	2022	Design of a hierarchical blockchain framework for secure and scalable P2P energy trading in distributed systems.
40	WoS-066	[71]	Energy trading and scheduling in networked microgrids using fuzzy bargaining game theory and distributionally robust optimization	Mohseni, S.	2023	Proposal of a hierarchical trading and scheduling model for networked microgrids using dual decomposition techniques.
41	WoS-080	[32]	A Hierarchical Deep Reinforcement Learning-Based Community Energy Trading Scheme for a Neighborhood of Smart Households	Yan, L. et al.	2022	Development of a hierarchical deep reinforcement learning model for cooperative energy trading and scheduling in smart grids.
42	WoS-082	[57]	TrustyFeer: A Subjective Logic Trust Model for Smart City Peer-to-Peer Federated Clouds	Kurdi, H. et al.	2018	Introduction of a trust model (TrustyFeer) based on subjective logic to enhance security in P2P energy trading systems.
43	WoS-086	[59]	Power Loss Minimization of Parallel-Connected Distributed Energy Resources in DC Microgrids Using a Distributed Gradient Algorithm-Based Hierarchical Control	Jiang, Y. et al.	2022	Optimization of power loss minimization using a control strategy for parallel inverters in P2P microgrid environments.
44	WoS-091	[58]	Research on Blockchain-Enabled Smart Grid for Anti-Theft Electricity Securing Peer-to-Peer Transactions in Modern Grids	Din, J. et al.	2024	Analysis of blockchain-based transactive energy trading systems and their applicability to smart grid infrastructures.
45	WoS-098	[82]	A Fully Decentralized Hierarchical Transactive Energy Framework for Charging EVs With Local DERs in Power Distribution Systems	Yang, J. et al.	2022	Development of a fully decentralized transactive energy system using hierarchical agent-based coordination.
46	WoS-107	[33]	Smart Grid Ecosystem Modeling Using a Novel Framework for Heterogenous Agent Communities	Pereira, H. et al.	2022	Proposal of a novel ecosystem modeling framework for simulating smart grid operation with P2P and distributed control.
47	WoS-111	[30]	Bilateral energy-trading model with hierarchical personalized pricing in a prosumer community	Huang, T. et al.	2022	Design of a bilateral energy trading model with hierarchical structures to enable efficient peer negotiation and control.
48	WoS-113	[15]	A Decentralized Approach for Frequency Control and Economic Dispatch in Smart Grids	Lü, P. et al.	2017	Presentation of a decentralized control approach for frequency stability in distributed and P2P energy networks.
49	WoS-117	[93]	A novel hierarchical fault management framework for wireless sensor networks: HFMF	Moridi, E. et al.	2022	Development of a hierarchical framework for fault management in decentralized smart grids using multi-agent systems.
50	WoS-124	[87]	A stochastic hierarchical optimization and revenue allocation approach for multi-regional integrated energy systems based on cooperative games	Han, F. et al.	2023	Design of a stochastic optimization and revenue mechanism for hierarchical P2P electricity markets in smart grids.
51	WoS-131	[92]	A hierarchical blockchain-based electricity market framework for energy transactions in a security-constrained cluster of microgrids	Esfahani, M.	2022	Development of a hierarchical blockchain-based electricity market model for enhancing trust and decentralization in P2P trading.
52	WoS-143	[34]	Distributed mixed-integer fuzzy hierarchical programming for municipal solid waste management. Part I: System identification and methodology development	Cheng, G. et al.	2017	Design of a fuzzy hierarchical optimization model for distributed and uncertain peer-to-peer energy management.
53	WoS-144	[70]	An Energy Sharing Mechanism Achieving the Same Flexibility as Centralized Dispatch	Chen, Y. et al.	2021	Proposal of an energy sharing mechanism ensuring equal economic benefits across prosumers in P2P microgrids.
54	WoS-152	[23]	A trust model for recommender agent systems	Majd, E.	2017	Development of a trust-based recommender agent system to improve decision making in decentralized energy environments.
55	WoS-153	[60]	Distributed Dynamic Resource Management and Pricing in the IoT Systems With Blockchain-as-a-Service and UAV-Enabled Mobile Edge Computing	Asheralieva, A.	2020	Design of a distributed dynamic resource management system to ensure secure and privacy-preserving P2P energy trading.
56	WoS-160	[35]	Multi-microgrid low-carbon economy operation strategy considering both source and load uncertainty: A Nash bargaining approach	Xu, J.	2023	Optimization of multi-microgrid operations under low-carbon constraints using cooperative peer-based mechanisms.
57	WoS-184	[100]	Tracking down coupled innovations supporting agroecological vegetable crop protection to foster sustainability transition of agrifood systems	Boulestreau, Y. et al.	2022	Analysis of coupled innovations supporting P2P energy initiatives and citizen engagement in energy transitions.
58	WoS-197	[69]	Neural-Blockchain-Based Ultrareliable Caching for Edge-Enabled UAV Networks	Sharma, V. et al.	2019	Integration of neural networks and blockchain for ultrareliable energy trading and caching in edge networks.
59	WoS-209	[101]	Making the Most of Mealtimes (M3): protocol of a multi-centre cross-sectional study of food intake and its determinants in older adults living in long term care homes	Keller, H. et al.	2017	Protocol design for optimizing mealtime routines in communal settings with potential smart grid synergies.
60	WoS-210	[48]	Enabling machine learning-ready HPC ensembles with Merlin	Peterson, J. et al.	2022	Implementation of HPC-ready ensembles with machine learning capabilities for advanced simulation of P2P energy systems.
61	S-002	[63]	Decentralized energy trading in microgrids: a blockchain-integrated model for efficient power flow with social welfare optimization	Umar, A.	2024	Comprehensive review of decentralized energy trading mechanisms in microgrids and their practical challenges.
62	S-011	[36]	Distributed robust operation strategy of multi-microgrid based on peer-to-peer multi-energy trading	Gao, J. et al.	2023	Proposal of a robust distributed operation strategy for P2P energy cooperation in multi-microgrid frameworks.
63	S-018	[4]	Peer-to-peer energy arbitrage in prosumer-based smart residential distribution system	Ullah, M.	2019	Design of an energy arbitrage model using P2P trading among prosumers to reduce electricity costs.
64	S-022	[80]	Peer-to-Peer Energy Cooperation in Building Community over A Lossy Network	Lyu, C.	2021	Development of a cooperation model for P2P energy exchange in building communities using game-theoretic approaches.
65	S-024	[86]	A peer-to-peer energy trading model for community microgrids with energy management	Ravivarma, K.	2024	Community-oriented P2P trading model supporting energy sharing, optimized pricing, and user participation.
66	S-025	[37]	Peer-to-Peer Energy Trading Among Networked Microgrids Considering the Complementary Nature of Wind and PV Solar Energy	Michon, D.	2023	Implementation of a communication-driven multi-microgrid P2P energy trading strategy with real-time adaptability.
67	S-027	[38]	Peer-to-Peer Energy Trading Among Prosumers in Energy Communities Based on Preferences Considering Holacracy Structure	Afzali, P. et al.	2024	P2P energy exchange optimization among prosumers in virtual power plants using distributed decision models.
68	S-032	[46]	A Fully Distributed Privacy-Preserving Energy Management System for Networked Microgrid Cluster Based on Homomorphic Encryption	Yuan, Z. et al.	2024	Privacy-preserving distributed architecture for secure P2P energy trading among agents in microgrids.
69	S-037	[102]	Robust Energy Management of Multi-microgrids System Considering Incentive-Based Demand Response Using Price Elasticity	Datta, J.	2022	Robust management system for interconnected microgrids ensuring energy balance and stability in P2P settings.
70	S-038	[27]	Designing Fairness in Autonomous Peer-to-peer Energy Trading	Behrunani, V. et al.	2023	Proposal of a fairness-oriented framework for autonomous P2P trading balancing efficiency and equity.
71	S-062	[83]	Peer-To-peer Energy Transaction Incorporating Prosumers’ Tendency with load and PV uncertainty consideration	Moradi, M. et al.	2024	Proposal of a novel P2P energy trading system integrating distributed energy storage and bi-directional energy flows.
72	S-066	[18]	V2GNet: Robust Blockchain-Based Energy Trading Method and Implementation in Vehicle-To-Grid Network	Liang, Y.	2022	Development of a robust blockchain-based energy trading platform (V2GNet) ensuring integrity and scalability in P2P transactions.
73	S-068	[95]	Multi-agent system architecture for enhanced resiliency in autonomous microgrids	Lakshminarayanan, V.	2018	Design of a hierarchical multi-agent system architecture to improve resilience and flexibility in P2P energy trading.
74	S-070	[24]	Optimal Scheduling of Hierarchical Energy Systems with Controllable Load Demand Response	Xiaoguang, Z. et al.	2022	Optimal scheduling strategy for hierarchical energy systems enabling P2P trading and improved demand–supply matching.
75	S-074	[89]	A Robust Decentralized Peer-to-Peer Energy Trading in Community of Flexible Microgrids	Saatloo, A.	2023	Robust decentralized framework for P2P energy transactions ensuring operational resilience against faults and uncertainties.
76	S-076	[75]	Design of a Multiagent-Based Voltage Control System in Peer-to-Peer Networks for Smart Grids	Rohbogner, G. et al.	2014	Multi-agent-based voltage control system for P2P networks ensuring stable operation of distributed energy systems.
77	S-078	[39]	Decentralized Active Power Management in Multi-Agent Distribution Systems Considering Congestion Issue	Tofighi-Milani, M. et al.	2022	Proposal of a decentralized power management system for multi-agent P2P networks under partial observability.
78	S-079	[44]	Distributed Topology Optimization for Agent-based Peer-to-Peer Energy Market	Kilthau, M.	2023	Formulation of a distributed topology optimization model for agent-based control in smart grid energy trading.
79	WoS-011	[40]	A Byzantine-Resilient Distributed Peer-to-Peer Energy Management Approach	Chang, X. et al.	2023	Development of a byzantine-resilient distributed framework for secure and trustworthy P2P energy trading.
80	WoS-019	[90]	Designing a Robust Decentralized Energy Transactions Framework for Active Prosumers in Peer-to-Peer Local Electricity Markets	Mehdinejad, M. et al.	2022	Design of a robust decentralized energy trading mechanism focusing on fault tolerance and peer integrity.
81	WoS-022	[31]	Two-Stage Credit Management for Peer-to-Peer Electricity Trading in Consortium Blockchain	Zhou, K.	2024	Two-stage credit management system to enhance trust and financial robustness in P2P energy exchanges.
82	WoS-023	[73]	Modelling and analysis of a two-level incentive mechanism based peer-to-peer energy sharing community	Wang, Y. et al.	2021	Modeling of a two-level incentive system for motivating participants in P2P energy communities.
83	WoS-024	[74]	Peer-to-Peer energy trading considering the output uncertainty of distributed energy resources	Xia, Y. et al.	2022	Optimization model for P2P trading that incorporates the impact of uncertainty in distributed generation.
84	WoS-031	[41]	Multi-agent based energy community cost optimization considering high electric vehicles penetration	Faia, R. et al.	2023	Multi-agent-based optimization framework for cost-efficient energy exchange in community energy systems.
85	WoS-047	[43]	A Cross-Layer Trust-Based Consensus Protocol for Peer-to-Peer Energy Trading Using Fuzzy Logic	Chowdhury, M. et al.	2022	Design of a trust-based consensus protocol to secure P2P transactions and mitigate dishonest behavior.
86	WoS-051	[21]	A Multi-Agent Framework for P2P Energy Trading With EV Aggregators Supporting V2X Services	Singh, A. et al.	2024	Development of a multi-agent framework for EV-based P2P energy trading with charging station coordination.
87	WoS-053	[88]	A Novel Distributed Paradigm for Energy Scheduling of Islanded Multiagent Microgrids	Tofighi-Milani, M. et al.	2022	Proposal of a distributed scheduling paradigm for decentralized energy management in transactive energy systems.
88	WoS-054	[76]	Loss Allocation in Joint Transmission and Distribution Peer-to-Peer Markets	Moret, F. et al.	2021	Design of a novel loss allocation methodology combining transmission and distribution levels in P2P markets.
89	S-010	[79]	Double-Consensus-Based Distributed Energy Management in a Virtual Power Plant	Naina, P.	2022	Implementation of a double-consensus distributed energy management strategy for microgrids enabling secure P2P coordination.
90	S-039	[28]	An optimal energy management system for real-time operation of multiagent-based microgrids using a T-cell algorithm	Harmouch, F. et al.	2019	Design of an optimal real-time home energy management system supporting dynamic pricing and P2P energy flows.
91	WoS-059	[20]	Power-Flow-Based Secondary Control for Autonomous Droop-Controlled AC Nanogrids With Peer-to-Peer Energy Trading	Roncero-Clemente, C. et al.	2021	Development of a secondary control model based on power flow estimation for autonomous P2P microgrids.
92	WoS-104	[94]	Privacy-Preserving Distributed Learning for Renewable Energy Forecasting	Gonçalves, C.	2021	Proposal of a privacy-preserving distributed learning scheme for renewable energy forecasting in P2P systems.
93	S-006	[78]	A Robust-Based Home Energy Management Model for Optimal Participation of Prosumers in Competitive P2P Platforms	Zetawi, A. et al.	2024	Design of a robust home energy management model integrating renewable energy and electric vehicles for P2P trading.
94	S-026	[103]	Intent Profile Strategy for Virtual Power Plant Participation in Simultaneous Energy Markets With Dynamic Storage Management	Aguilar, J. et al.	2022	Introduction of an intent-based strategy for optimizing prosumer interactions within virtual power plant structures.

summarizes the main findings from the reviewed studies, which will be discussed in detail in the following section.

3. Results

This section presents the findings of the systematic review, examining 94 selected studies across six core themes that emerged from the eligibility and synthesis processes. Each subsection addresses a critical aspect of DES and P2P energy management. First, Optimization and Modeling in Energy Systems investigates robust and hybrid techniques for handling uncertainties in DES, including variable renewable generation and market fluctuations. Next, Multi-Agent Systems and Distributed Control explore decentralized frameworks and reinforcement learning methods that enhance scalability and resilience in complex energy networks. Simulations, Case Studies, and Real-World Applications highlight how theoretical models and frameworks translate into practical solutions, providing evidence of cost savings and environmental benefits under diverse conditions. Blockchain, Smart Contracts, and Emerging Technologies discuss the use of secure, transparent platforms to automate P2P transactions, integrate AI-driven consensus protocols, and address scalability challenges. Frameworks for Robust and Decentralized Energy Management focus on hierarchical and multi-layered approaches that handle fault tolerance, market disruptions, and evolving resource configurations. Finally, Electric Machines and their role in DES underscores the design and operational enhancements of electric machines—especially their integration into V2G and decentralized microgrids—to bolster flexibility, efficiency, and renewable energy penetration.

3.1. Optimization and Modeling in Energy Systems

3.1.1. Main Trends

Recent advancements in the optimization and modeling of energy systems have been marked by the development of algorithms and models tailored to address the complexity and uncertainties inherent in DES. A prominent trend involves the application of stochastic and robust optimization techniques, which effectively handle uncertainties related to market prices, renewable generation, and variable loads [23]. For instance, a stochastic–robust model designed for 100% renewable energy communities integrates P2P transactions, collective assets like wind turbines and batteries, and coordinated scheduling among prosumers. This approach has demonstrated significant reductions in unserved energy and operational costs while maximizing resource utilization under uncertain conditions [1]. The adoption of hybrid approaches, combining robust optimization with risk management techniques such as CVaR, represents another key trend. These methods effectively model financial and operational risks associated with uncertain parameters, including limited historical data or unpredictable solar irradiation. A study employing this methodology illustrated that combining shared storage and P2P trading could lower operational costs by 21.49%, emphasizing the effectiveness of hybrid frameworks in addressing high-variability scenarios [16].

Multi-objective optimization models have gained traction by simultaneously addressing economic, environmental, and technical goals. For example, a hierarchical energy management framework incorporates algorithms such as NSGA-II to optimize costs, carbon emissions, and wind energy spillage in DES. This approach aligns with modern demands for sustainability while enhancing decision making for resource allocation [24]. Moreover, decentralized models for energy communities have emerged as a solution to the challenges of resource coordination in complex systems. Techniques such as the Alternating Direction Method of Multipliers (ADMM) enable efficient energy management across multiple actors, reducing the computational time and improving operational flexibility under uncertain conditions [25,26,27].

3.1.2. Originality

The originality in this domain lies in how recent models effectively integrate uncertainty and transition from theoretical frameworks to practical applications. For example, the stochastic–robust model for fully renewable communities also introduces practical solutions for day-ahead scheduling, incorporating P2P transactions among prosumers [2,28,29]. This represents a notable departure from traditional, deterministic approaches. Hybrid methodologies, such as those combining robust optimization with CVaR, demonstrate a novel ability to address financial and operational risks simultaneously. These frameworks provide actionable insights even in scenarios with limited data, addressing challenges previously considered insurmountable [4,30].

Another innovative aspect is the emphasis on privacy-preserving decentralized systems. Distributed platforms employing blockchain and algorithmic solutions enable participants to share minimal information while collectively optimizing operations. This dual focus on efficiency and security responds to critical challenges in the energy market [18,31]. Additionally, recent models increasingly leverage emerging technologies like AI and deep learning. For instance, a hybrid model utilizing Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks has shown remarkable accuracy in predicting controllable loads, facilitating more effective resource management in distributed systems [24,32,33]. Such applications signify a shift towards practical, technology-driven solutions that address real-world complexities.

3.1.3. Comparative Analysis

Co-optimization in distributed networks represents a substantial improvement over centralized systems, offering enhanced flexibility, resilience, and adaptability [34]. One key distinction lies in the superior capacity of distributed systems to handle uncertainties through advanced methodologies such as stochastic and robust optimization. Unlike centralized systems, which often rely on deterministic models and historical data, distributed frameworks like ADMM-based models dynamically address uncertainties in real-time, significantly enhancing the operational reliability [25]. Another advantage of distributed approaches is their ability to integrate multiple objectives and diverse stakeholders. For instance, Nash Bargaining-based models also ensure equitable benefit distribution among participants—an aspect often overlooked in centralized systems [35].

Decentralization also brings improved data privacy and security. Blockchain-enabled platforms eliminate the need for intermediaries, creating a more transparent and secure environment for P2P energy transactions [18]. This contrasts sharply with centralized systems, which are more vulnerable to cybersecurity risks and often lack transparency in operations. Then, distributed systems are better positioned to adopt emerging technologies such as AI and evolutionary algorithms. These tools enable dynamic resource optimization in decentralized networks, overcoming the rigidity and scalability issues inherent to centralized structures [24]. The inherent adaptability of distributed systems positions them as a practical solution for the dynamic and multifaceted challenges of modern energy systems.

3.2. Multi-Agent Systems and Distributed Control

3.2.1. Main Trends

The adoption of MAS in energy management has transformed how DERs are coordinated and operated [36,37,38]. These systems enable decentralization in prosumer communities and provide robust frameworks for managing energy trading and distribution in increasingly complex networks [39,40,41]. Among the key advancements is the use of MAS to facilitate P2P energy trading in Local Electricity Markets (LEMs). These platforms enable prosumers to trade surplus energy directly, offering economic benefits and enhancing the system efficiency. For example, a decentralized framework proposed for LEMs incorporates advanced agents that act collaboratively via local computations and P2P communication, achieving effective frequency regulation and economic dispatch [15].

Another significant trend is the application of DRL within MAS frameworks. DRL-based methods have been shown to enhance the scalability and performance in P2P trading systems. One notable example utilizes Multi-Agent Deep-Reinforcement Learning (MADRL) to manage local energy markets, optimizing both trading and flexibility services. This approach integrates techniques like a multi-actor attention critic and prioritized experience replay to achieve superior results compared to traditional methods, improving system reliability and economic outcomes [5]. Similarly, Federated Reinforcement Learning (FRL) has emerged as a promising approach to maintain privacy while optimizing energy trading and carbon allowance management in building-level Multi-Energy Systems (MES). By employing FRL, agents can balance economic and environmental goals, achieving notable reductions in energy costs and carbon emissions [42]. Clustering techniques have also enhanced the scalability of MAS in energy systems. For instance, a multi-cluster DRL framework addresses the heterogeneity of prosumers by grouping them based on their DER portfolios. This method employs shared policies within clusters to accelerate training and reduce the computational complexity. It has demonstrated strong generalization capabilities and an improved performance in large-scale scenarios, such as managing the energy trading of 300 residential households [6].

In addition to trading, MAS frameworks have been integrated with other emerging technologies like V2X systems. These frameworks optimize the use of EV batteries for various operations, such as V2G and vehicle-to-vehicle (V2V) energy exchanges, while ensuring safe operational limits. For instance, a MAS framework that incorporates V2X has been shown to enhance microgrid stability and facilitate dynamic energy exchanges, achieving efficient resource utilization under real-world constraints [21]. Furthermore, agent-based systems have been applied to control strategies in microgrids with inverter-interfaced DERs. These systems leverage graph theory to enable effective communication between agents, resulting in the proportional sharing of real and reactive power, even under volatile conditions [14]. In the same case, trust management within MAS has received significant attention due to its critical role in decentralized systems. A cross-layer consensus protocol based on fuzzy logic introduces human-centric trust mechanisms to address vagueness and ensure transparency in P2P energy trading platforms. This protocol also enhances resilience against advanced attacks, such as Sybil and selfish mining, further strengthening the security of decentralized energy trading frameworks [43].

3.2.2. Originality

Recent advancements in MAS have introduced innovative approaches to improve coordination among agents and optimize energy management [44,45]. One of the most notable contributions is the integration of DRL and FRL techniques to balance privacy, efficiency, and scalability in decentralized systems [38,46,47,48]. For instance, the use of federated learning within MAS allows for decentralized training without compromising data privacy. By utilizing abstract critic networks and deep deterministic policy gradient methods, FRL-based systems achieve robust coordination in MES, providing economic and environmental benefits [42]. Another innovative approach involves the application of MADRL, which incorporates advanced techniques such as prioritized experience replay and multi-actor attention critics. These enhancements enable prosumer agents to make informed decisions in dynamic and uncertain environments, addressing the limitations of model-based approaches. The method has been validated in large-scale real-world settings, demonstrating significant improvements in the trading efficiency and operational stability [5].

Scalability has also been addressed through clustering strategies, as demonstrated by the multi-cluster DRL framework. This approach combines parameter sharing within clusters with diversity across clusters, allowing agents to adapt to heterogeneous prosumer characteristics while accelerating the training process. Such innovations are critical for managing the increasing complexity of DER coordination in large-scale systems [6]. The integration of V2X systems into MAS frameworks is another area of originality [49]. By developing safe operational envelopes for EV batteries and incorporating model predictive control architectures, these systems optimize the interaction between EVs and microgrids. This allows for efficient energy trading and enhances grid stability without violating operational constraints [21]. On the other hand, the introduction of human-centric trust mechanisms in P2P energy trading platforms represents a novel contribution. A fuzzy logic-based cross-layer protocol addresses the inherent vagueness of trust values while ensuring transparency and security in decentralized trading environments [50].

3.2.3. Research Gaps

Despite their advancements, MAS faces several challenges that limit its full potential. One significant issue is scalability, especially in large-scale systems where the heterogeneity of prosumers and the complexity of their interactions can strain computational resources. While clustering techniques and parameter-sharing methods offer partial solutions, further research is needed to develop more efficient algorithms capable of managing thousands of agents without compromising the performance [6,51]. Another critical gap is data security and privacy in decentralized systems. While blockchain-based platforms have introduced robust mechanisms for transparency and immutability, the integration of human-centric trust protocols remains underexplored. Existing solutions often fail to address the complexities of human-in-the-loop scenarios or adequately prevent advanced cyberattacks, such as 51% or Sybil attacks [43]. The dynamic nature of energy markets also poses challenges for MAS frameworks. Many current methods struggle to adapt to rapidly changing conditions, such as fluctuating renewable energy generation and unpredictable demand patterns. Advanced techniques like MADRL and FRL require further refinement to ensure their reliability in highly variable environments [5,42]. In this context, the integration of emerging technologies like V2X into MAS frameworks necessitates further investigation. Although promising results have been achieved in optimizing EV participation in energy trading, the interaction between V2X operations and other DERs remains a complex challenge. Enhanced coordination strategies are needed to fully exploit the potential of EVs in decentralized energy systems [14,21].

3.3. Simulations, Case Studies, and Real-World Applications

3.3.1. Main Trends

The validation of energy management strategies through simulations and case studies has become a cornerstone in the transition from theoretical frameworks to practical, scalable solutions. Hierarchical control frameworks, for instance, have been extensively studied and validated in various contexts. A prominent example is the scheduling of TCLs aggregated into a virtual battery model. This hierarchical system coordinates energy trading among aggregators while preserving user privacy. By addressing the scalability challenges of centralized systems, this approach has demonstrated its ability to optimize distributed energy trading with significant efficiency improvements in simulated scenarios [17]. Networked microgrids have also been a focus of simulation studies, particularly in isolated and constrained environments. A hierarchical energy management system (EMS) for islanded microgrids incorporates blockchain for secure energy transactions and synchronization algorithms for microgrid interconnection. The EMS also includes a real-time controller for energy storage systems, addressing the unique challenges posed by floating photovoltaic platforms. Hardware-in-the-Loop (HIL) simulations confirm the efficacy of the EMS in maintaining stability and coordinating energy flows under dynamic conditions, underscoring its practicality in real-world applications [52].

The rise of P2P energy trading frameworks has introduced innovative mechanisms for decentralized energy management. A scheduling strategy for interconnected microgrids, incorporating robust optimization (RO) and Nash bargaining principles, ensures fair profit sharing while handling the uncertainties of wind power. Decentralized solutions using ADMM algorithms have shown through simulations that such systems can achieve cost reductions and maintain participant privacy, validating their application in DES [53]. Similarly, Decentralized Autonomous Organizations (DAO) for smart buildings utilize day-ahead scheduling mechanisms combined with P2P energy sharing. These frameworks enhance the operational efficiency, reducing costs by 33.7% while maintaining resilience against communication failures [54]. Further advancing energy management, studies on commercial prosumers and neighborhood-level flexibility systems have highlighted their potential for achieving net-zero energy communities. For instance, a hierarchical system in subtropical Hong Kong integrates demand-side flexibility with P2P trading, reducing operational costs by 24.6% and CO2 emissions by 7.1%. Such studies underscore the economic and environmental viability of coordinated energy sharing in diverse urban contexts [55].

3.3.2. Originality

Simulations and case studies have revealed the significant impact of geographic, technological, and contextual factors on the implementation of energy systems. In islanded and remote microgrids, hierarchical EMS solutions address the unique challenges of intermittent RES and limited connectivity. The use of specialized real-time controllers for energy storage in floating PV systems exemplifies how tailored solutions can address site-specific constraints while ensuring energy reliability [52]. In urban settings, DAO-based energy management frameworks have demonstrated their ability to meet the complex demands of densely populated areas. By incorporating autonomous scheduling and energy-sharing mechanisms, these systems optimize resource use while addressing privacy concerns and potential communication disruptions. This approach highlights the critical role of advanced algorithms in achieving decentralized, resilient energy networks [54]. In isolated environments, P2P trading frameworks tailored for low-voltage nanogrids have demonstrated innovative strategies for managing energy exchanges between prosumers.

A secondary control method ensures the stability of power-sharing arrangements, leveraging the droop control and adaptive voltage regulation to maintain reliable operation in resource-constrained systems [56]. Similarly, the integration of V2X technologies into P2P trading frameworks showcases the potential of EVs to support dynamic energy management in urban microgrids. These frameworks balance grid constraints with EV battery safety, enabling efficient trading without compromising operational limits [21]. Geographic and political factors further influence the implementation of energy systems. In Hong Kong, for example, demand-side flexibility strategies have been integrated with P2P trading to address regional energy challenges. These systems align with local environmental goals and regulatory frameworks, demonstrating how localized solutions can enhance sustainability while maintaining compatibility with the grid infrastructure [55].

3.3.3. Critical Analysis

Despite significant progress in simulations and case studies, gaps persist between theoretical models and real-world applications [57]. Many simulations, such as those addressing hierarchical TCL scheduling in transactive energy systems, rely on simplified assumptions about user behavior and market dynamics. These idealized scenarios often fail to account for the complexities of real-world environments, limiting the scalability and adaptability of the proposed solutions [17]. Challenges related to data reliability and system resilience also emerge when transitioning from simulations to practical implementations. Blockchain-enabled frameworks for smart grids, for instance, offer robust solutions for enhancing security and transparency.

However, these systems are susceptible to vulnerabilities such as cyberattacks and high computational demands, particularly in environments with advanced decentralized consensus mechanisms [58]. Similarly, decentralized algorithms for managing DERs in DC microgrids may underperform in the presence of communication failures or hardware limitations, which are often underestimated in simulated scenarios [59,60]. The integration of emerging technologies and diverse stakeholders into energy management frameworks presents further challenges. While V2X-enabled P2P trading frameworks have shown potential in simulations, real-world applications must address issues such as EV battery longevity and the variability of vehicle arrivals and departures. These dynamic factors add complexity to system coordination, requiring robust strategies to fully realize the benefits of V2X technologies [21]. Likewise, achieving a balance between economic and environmental goals in urban smart grids requires more comprehensive strategies that consider regional variations in the energy demand and policy constraints [55].

3.4. Blockchain, Smart Contracts, and Emerging Technologies

3.4.1. Main Trends

The adoption of blockchain and smart contract technologies has revolutionized the management of DERs, particularly in enabling secure, transparent, and efficient energy transactions. These technologies have been widely applied in P2P energy trading systems, which decentralize energy markets by eliminating the need for intermediaries and automating transactions [61,62]. Blockchain’s inherent features—immutability, transparency, and decentralization—align perfectly with the goals of modern energy systems [63]. A hierarchical blockchain framework, for instance, has been utilized within microgrids to integrate distributed control with energy trading [64]. By combining public and private blockchain systems, this approach ensures high throughput for trading while maintaining system stability and security against false data injection attacks [19].

One prominent trend is the use of blockchain-based smart contracts to automate energy transactions and enforce rules defined by market participants. For example, a blockchain-enabled system in Singapore incorporated smart contracts to manage energy exchanges among residential, commercial, and industrial users. By modeling inter-sectorial interactions with non-cooperative game theory and considering renewable generation uncertainties through receding horizon optimization, the system flattened electricity procurement schedules and improved the overall market efficiency [65]. Artificial intelligence has also been integrated with blockchain to enhance the system performance and security. A notable example is a secure energy trading platform for residential homes that combines blockchain with an Improved Sparse Neural Network (ISNN). This system introduced a novel Proof-of-Computational-Closeness (PoCC) consensus protocol, which optimized computational costs and provided robust defenses against Sybil attacks. The ISNN further accelerated error convergence rates and minimized data transmission times, proving scalable and efficient for large-scale implementations [66].

Another innovative application is the use of blockchain to manage Virtual Power Plants (VPPs) composed of small-scale prosumers. These decentralized systems aggregate energy resources through public blockchains and self-enforcing smart contracts, which automate energy contributions, delivery tracking, and remuneration processes. The experimental results demonstrated that such systems operate effectively on public blockchains with manageable computational overheads, offering a scalable solution for energy aggregation [67]. Blockchain’s role in secure grid management extends beyond trading. A hierarchical blockchain model demonstrated its capability to safeguard distributed control systems within microgrids. By employing Proof of Authority (PoA) and smart contracts to regulate control parameters like active power and frequency, this model ensured a robust performance under adverse conditions while improving the social welfare of microgrid participants through double-auction pricing mechanisms [19].

3.4.2. Originality

Blockchain technology, despite its potential, faces several inherent challenges, particularly regarding the energy consumption and scalability. Traditional consensus mechanisms like Proof of Work (PoW) are resource-intensive, making them unsuitable for energy-efficient applications. To address this, innovative lightweight blockchain solutions have emerged. For example, a hierarchical blockchain structure incorporating PoA for energy trading and private blockchains for distributed control minimizes computational overhead while maintaining high security and reliability. This hybrid approach enables the practical application of blockchain in energy-constrained systems like microgrids [19]. Privacy concerns in decentralized energy systems have also been addressed through blockchain innovations. A multi-pseudonym mechanism integrated into a blockchain-based trading platform preserves user anonymity while ensuring transaction transparency [68]. This system leverages advanced optimization algorithms to reduce the computational complexity and enhance scalability, proving effective in managing large-scale energy trading environments [66]. Similarly, a neural blockchain model designed for ultra-reliable caching in edge networks combines blockchain’s immutability with machine learning to ensure data reliability, connectivity, and reduced energy consumption, highlighting its potential for communication-intensive applications [69].

The development of trust mechanisms has introduced a novel dimension to blockchain applications. A cross-layer trust-based consensus protocol uses fuzzy logic to quantify trust levels in human-interpretable terms, addressing vagueness in traditional trust models. By incorporating human-in-the-loop processes, this protocol enhances the security against attacks like selfish mining and 51% attacks, ensuring robust and reliable operation in decentralized energy trading systems [43]. Blockchain-based frameworks have also innovated energy demand management. In distributed hybrid energy systems, smart contracts automate energy trading while incorporating renewable generation uncertainties [70]. The results of such implementations demonstrate significant improvements in renewable energy utilization and demand-side flexibility, making them an attractive option for future energy systems [65].

3.4.3. Key Contributions

Blockchain and smart contracts have made substantial contributions to addressing long-standing challenges in decentralized energy systems, including regulatory compliance, privacy protection, and operational inefficiencies. A critical contribution lies in automating regulatory adherence through smart contracts. For example, a decentralized energy trading platform encoded legal and market constraints into smart contracts, ensuring seamless compliance and fair profit distribution among participants even under uncertain conditions [71]. Blockchain technologies have significantly enhanced privacy protection. Mechanisms such as multi-pseudonym protocols and decentralized identifiers ensure that sensitive user data remain secure during energy transactions, fostering trust and transparency in P2P energy markets [72,73,74]. These solutions maintain the dual goals of data protection and system accountability, addressing one of the most pressing concerns in decentralized systems [66].

Blockchain’s ability to enable decentralized governance has transformed the energy landscape. Virtual aggregation models, for instance, use blockchain to coordinate energy services for small-scale prosumers, facilitating efficient energy delivery and remuneration processes. These systems not only provide scalability, but also break down market barriers for more minor participants, allowing them to benefit from larger-scale energy systems [19,67]. The integration of blockchain into distributed control systems further ensures grid stability and resilience. By combining blockchain with advanced control algorithms, these systems protect against cyber threats and improve the robustness of energy operations. A notable example is the use of hierarchical blockchain models in microgrids, which prevent data manipulation while maximizing participant benefits through innovative pricing schemes [19]. Blockchain and smart contracts are also instrumental in fostering equitable energy trading practices [75]. Decentralized P2P trading systems employing robust optimization models ensure fair profit allocation, incentivizing active participation from all stakeholders [76,77]. These models enhance market efficiency while maintaining the integrity of trading processes, highlighting blockchain’s role in enabling sustainable energy systems [71].

3.5. Frameworks for Robust and Decentralized Energy Management

3.5.1. Main Trends

The increasing complexity of DES has driven the development of robust and decentralized energy management frameworks. These frameworks address uncertainties, enhance resilience, and maintain stability in multi-layered energy networks [78,79]. A prominent approach involves hierarchical structures, where tasks are divided into distinct layers to optimize the performance and scalability. For example, a hierarchical energy management system for P2P trading incorporates three levels: individual household optimization, P2P energy sharing, and joint optimization among selected prosumers [20,80]. This system demonstrated up to 8.96% annual cost savings for households participating in energy communities, emphasizing its potential to enhance economic efficiency while decentralizing decision making [81].

Decentralized frameworks have also leveraged the flexibility of aggregated energy resources, such as EVs and TCLs. A multi-layered strategy integrates these resources into battery energy storage systems and virtual energy storage systems to balance supply and demand [82,83,84]. The coordination spans the transmission, distribution, and behind-the-meter layers, using CNN-LSTM algorithms to predict local loads and distributed optimization techniques to safeguard privacy. This structure provides economic and environmental benefits [85,86]. Robust energy management frameworks have been instrumental in autonomous microgrids and nanogrids. By adopting secondary control mechanisms, such as power-flow-based algorithms in droop-controlled AC nanogrids, these systems achieve reliable P2P trading while maintaining voltage and frequency stability. Experimentally validated models show that such frameworks can ensure energy-sharing efficiency and operational consistency in complex distributed networks [20]. Furthermore, frameworks combining robust optimization and cooperative game theory have been developed to manage interconnected multi-microgrid systems. These approaches address the operational risks associated with renewable energy intermittency and market fluctuations. For instance, a stochastic hierarchical optimization model utilizes the Nash bargaining game to allocate revenues fairly among regional energy systems, improving renewable energy utilization and reducing carbon emissions in interconnected grids [87].

3.5.2. Originality

Recent innovations in decentralized energy management frameworks highlight their capacity to balance robustness with sustainability. For instance, a novel framework for islanded microgrids introduces P2P management with scenario-based stochastic optimization and CVaR analysis. This combination ensures reliable energy scheduling under high uncertainty, enabling microgrid operators to optimize local resources effectively while accounting for operational risks [88]. Another original contribution comes from frameworks designed to handle community-level energy sharing. A robust P2P energy trading platform for flexible microgrids incorporates emerging technologies like energy storage systems, demand response programs, and EVs. By using cooperative bargaining games and decentralized optimization methods, the platform enhances self-sufficiency and economic benefits for participants while accommodating uncertainties in electricity markets [89,90,91].

Hierarchical blockchain-based frameworks have also emerged as innovative solutions for energy transaction security and efficiency. These systems integrate ledger summarization and reliability constraints into the transaction verification process, ensuring secure and optimized energy trading within clusters of microgrids. By addressing congestion issues in transmission corridors and providing scalable solutions for market clearing, these frameworks contribute to the robustness of decentralized energy systems [92]. Fault management frameworks have further advanced decentralized systems. A hierarchical model for wireless sensor networks, for example, uses clustering algorithms to enhance fault detection and recovery in inaccessible areas. By incorporating backup mechanisms and self-detection methods, the framework minimizes energy consumption and maintains optimal activity despite hardware limitations, demonstrating the potential for robust control in energy systems with high fault tolerance requirements [93].

3.5.3. Future Perspectives

The future of robust and decentralized energy management lies in integrating advanced predictive control and AI to enhance system adaptability. Predictive models, such as those combining CNN with LSTM algorithms, have already shown promise in forecasting load demands and coordinating distributed resources [94]. Expanding these methods to include real-time data integration and adaptive learning could significantly improve the dynamic response capabilities of decentralized systems [85]. AI-driven optimization techniques, such as hybrid grey-wolf and whale optimization algorithms, have demonstrated potential in multi-microgrid systems by addressing renewable energy intermittencies and demand-side variability. Incorporating such methods into decentralized frameworks could further enhance their scalability and efficiency in managing complex, interconnected grids [16].

Additionally, emerging technologies like blockchain and smart contracts are expected to play a crucial role in ensuring transparency and security in decentralized markets. By integrating trust-based consensus protocols and human-centric metrics, future frameworks could address challenges related to privacy and trust in P2P energy trading systems, enabling broader adoption and acceptance [92]. On the other hand, decentralized energy management frameworks must increasingly focus on resilience to disruptions in dynamic markets. Adaptive control mechanisms that leverage real-time analytics and scenario-based planning will be essential for mitigating the impacts of market volatility and renewable energy fluctuations. These advancements will ensure that decentralized systems remain reliable and efficient even under rapidly changing conditions, paving the way for sustainable and robust energy networks.

3.6. Electric Machines and Their Role in Distributed Energy Systems

3.6.1. Current Developments

The design and operation of electric machines have undergone significant advancements to address the evolving needs of DES. Modern designs increasingly prioritize flexibility, efficiency, and compatibility with distributed generation sources, such as solar and wind power. Synchronous machines and reluctance generators, among other types, are being optimized for integration into microgrids and nanogrids. For instance, advanced droop-controlled inverters interfaced with LC or LCL filters have been employed in autonomous nanogrids to maintain voltage and frequency stability during P2P energy trading. These systems ensure a robust control performance under variable generation conditions, enabling effective energy sharing and grid support [20].

Reluctance generators, known for their simplicity and robustness, are being re-engineered to enhance their efficiency in renewable energy applications. Their adaptability to variable speed operations, particularly in wind energy systems, makes them well-suited for dynamic environments. Coupled with advanced power electronics and secondary control mechanisms, these machines can deliver a consistent performance while supporting energy trading in distributed networks [2]. Electric vehicles are also playing an important role in DES as mobile energy storage units. Integrating EV charging stations into microgrids has opened new opportunities for utilizing electric motors in V2G and V2V operations. This dual functionality of electric machines—both as energy converters and storage units—has significantly enhanced the flexibility and stability of distributed systems, particularly in urban transportation networks [95].

3.6.2. Innovative Approaches

The intersection of the electric machine design and DES has led to novel applications and innovative control strategies. A notable example is the integration of power converters with hierarchical control frameworks to improve the operational resilience. In autonomous microgrids, MAS-based controllers with self-healing capabilities ensure robust fault detection and recovery [16]. By combining sentinel agents with workload distribution techniques, these systems maintain a continuous energy supply even during power generation failures. Such architectures highlight the adaptability of electric machines to emerging control paradigms in distributed systems [95]. Another original application involves power-flow-based secondary control in droop-regulated nanogrids. Here, electric machines are paired with advanced power-sharing algorithms to manage real and reactive power dynamically. Laboratory-scale experiments have demonstrated that this approach enables effective energy sharing and enhances the overall stability of autonomous microgrids. This integration of a traditional machine design with modern control techniques reflects a significant leap forward in distributed energy management [20].

Electric machines also play a critical role in blockchain-enabled VPPs. By aggregating small-scale prosumers, these systems utilize electric machines in decentralized frameworks to track and validate energy contributions. Smart contracts ensure that energy delivery and remuneration processes are executed seamlessly, leveraging the inherent capabilities of machines to interface with digital and physical systems. This innovative application showcases the growing importance of electric machines in facilitating decentralized energy markets [67].

3.6.3. Technology Comparison

Traditional electric machines, while reliable and durable, were initially designed for centralized systems with predictable load and generation patterns [92]. In DES, however, the demands on these machines have changed significantly. Advanced systems now require machines capable of operating under variable conditions, such as intermittent RES and fluctuating loads. For instance, synchronous machines have been adapted to work seamlessly with modern inverters and advanced control strategies, ensuring their compatibility with microgrid operations [20]. In contrast, newer designs, such as reluctance generators and direct-drive machines, have been specifically developed to address the unique challenges of DES. These machines exhibit high efficiency and reliability in environments with variable speeds, such as wind farms, and are equipped with enhanced cooling systems and optimized electromagnetic designs to reduce energy losses. Additionally, their integration with smart grid technologies enables real-time monitoring and control, providing advantages over traditional systems [2]. The dual role of electric machines in EVs—functioning as propulsion systems and energy storage units—further distinguishes advanced designs from their traditional counterparts. The ability to discharge energy back to the grid during peak demand periods or store surplus energy during low-demand hours has made EV-based energy systems a cornerstone of modern DES. These capabilities, combined with V2G and V2V operations, exemplify how advanced electric machines are shaping the future of distributed energy networks [95]. Finally,

Table 1. Comprehensive analysis of trends, challenges, and future directions in distributed energy systems.

Theme	Ref.	Challenges	Future Directions
1. Optimization and Modeling in Energy Systems	[1,2,3,4,5]	Addressing multi-objective optimization with conflicting goals; limited scalability of models for large-scale distributed systems; incorporation of dynamic market uncertainties.	Development of integrated AI and robust optimization frameworks; deployment of real-time optimization techniques to enhance responsiveness in dynamic environments.
2. MAS and Distributed Control	[5,12,14,15,17,20]	Scalability in large and complex systems; ensuring data privacy in multi-agent environments; effective coordination in systems with heterogeneous agents and varied energy profiles.	Implementation of federated learning within MAS for privacy-preserving operations; adoption of advanced clustering methods to manage scalability; increased focus on cybersecurity mechanisms in distributed networks.
3. Simulations, Case Studies, and Real-World Applications	[21,22,23,24,25]	Bridging the gap between simulated and real-world performance; addressing the variability of real-world factors such as renewable intermittency, market volatility, and consumer behavior.	Use of hardware-in-the-loop simulations for enhanced validation; development of simulation models incorporating real-world data to improve accuracy and reliability.
4. Blockchain, Smart Contracts, and Emerging Technologies	[30,31,32,33,34,35,36]	High computational costs of traditional blockchain systems; challenges in ensuring scalability and efficiency; addressing regulatory compliance and privacy concerns.	Development of lightweight blockchain systems; integration of trust-based consensus protocols; leveraging hybrid blockchain models combining public and private elements for efficiency and security.
5. Frameworks for Robust and Decentralized Energy Management	[39,40,41,42,43,45]	Balancing robustness with sustainability; handling uncertainties in decentralized systems; integrating fault tolerance in real-time operations.	Application of predictive control and AI to enhance adaptability; designing fault-tolerant mechanisms for decentralized systems; improved coordination among energy resources in interconnected microgrids.
6. Electric Machines and Their Role in DES	[42,44,46,48]	Adapting traditional machines for variable renewable integration; managing efficiency losses in high-demand scenarios; ensuring compatibility with advanced control systems.	Enhancing designs of reluctance generators and synchronous machines for renewable applications; integrating real-time monitoring and adaptive controls for energy conversion systems; advancing V2G technology for microgrid applications.

summarizes the key trends, current challenges, and future directions across the six central themes of the analysis, linking them to essential references. This overview highlights how each theme addresses specific issues in DES and identifies innovative solutions to overcome these limitations.

3.7. Critical Synthesis and Future Directions

The six analyzed themes reveal interconnected challenges in DES. On the one hand, stochastic–robust optimization models (Section 3.1) have been shown to reduce operational costs by up to 21.49% [16], but their reliance on centralized computations clashes with the decentralized nature of multi-agent systems (Section 3.2). This disconnect suggests the need for hybrid approaches, such as federated ADMM algorithms [25,42], which could bridge the gap between global optimization and distributed control. On the other hand, while blockchain (Section 3.4) offers transparency in P2P trading, its high energy consumption in protocols like PoW [66] limits its adoption in microgrids with practical constraints (Section 3.3). Emerging solutions, such as lightweight PoA variants [19] or sharding techniques [69], could balance this tension between security and efficiency.

A key finding is the disparity between theoretical studies and real-world implementations. For example, hierarchical frameworks for autonomous microgrids (Section 3.5) show annual savings of 8.96% in simulations [81], but often overlook human behavior variability or communication failures in real environments [17]. Similarly, traditional electric machines (Section 3.6), while robust, face unprecedented challenges in dynamic scenarios like V2G integration, where their efficiency declines under fluctuating loads [95]. This underscores the urgency for adaptive designs, such as CNN-LSTM-based predictive control systems [24,85], capable of anticipating real-time fluctuations.

4. Discussion

The findings of this systematic review demonstrate significant progress across all six thematic areas of DES, while simultaneously revealing critical interdependencies and persistent challenges that must be addressed to realize their potential fully. A central insight emerging from our analysis is the inherent tension between theoretical advancements and practical implementation, particularly evident in three key dimensions: the scalability–decentralization paradox, the security-efficiency trade-off, and the simulation-reality gap. The optimization and modeling of energy systems have made notable strides in addressing uncertainties through advanced algorithms and hybrid methodologies. Stochastic–robust optimization and CVaR frameworks have proven particularly effective in managing financial and operational risks, as evidenced by their ability to reduce costs by up to 21.49% in certain configurations [16]. However, these approaches often rely on centralized computational architectures that fundamentally conflict with the decentralized philosophy underlying MAS and distributed control frameworks. This paradox becomes even more pronounced when considering the energy demands of blockchain implementations, where the transparency and security benefits of protocols like PoW [66] come at the cost of significant computational overhead—a trade-off that currently limits their practical adoption in energy—constrained microgrid environments [19].

Multi-agent systems and distributed control have emerged as transformative solutions, offering enhanced flexibility and adaptability through innovations like DRL and FRL [5,42]. These approaches excel at preserving privacy while managing complexity in decentralized networks yet face substantial challenges when scaling to large, heterogeneous systems. While clustering techniques provide partial solutions [6], the development of more robust coordination mechanisms remains crucial, particularly for applications requiring real-time responsiveness in dynamic market conditions. The integration of trust mechanisms adds an important layer of security [43], but their implementation in rapidly evolving environments requires further investigation to balance security needs with the system performance.

The gap between the simulated performance and real-world implementation emerges as a recurring theme across all six analytical dimensions. Hierarchical energy management systems and blockchain-enabled frameworks demonstrate considerable potential in controlled environments, with some models showing annual savings of 8.96% in the simulation [81]. However, these results often rely on idealized assumptions that fail to account for the full complexity of human behavior, communication failures, or unpredictable market fluctuations [17]. This discrepancy underscores the need for more sophisticated validation approaches, such as hardware-in-the-loop simulations incorporating real-world data streams [52], to bridge the divide between theoretical models and practical applications. The evolution of the electric machine design illustrates both the progress and remaining challenges in DES implementation. While advanced synchronous and reluctance generators have adapted well to variable renewable sources and emerging paradigms like V2G and V2V trading [95], their performance in highly dynamic environments requires further optimization. The integration of these machines with MAS and blockchain frameworks shows promise for enhancing system resilience, but additional work is needed to fully realize these synergies, particularly in edge computing applications where real-time responsiveness is critical.

Moving forward, the DES field must prioritize three key areas of development. First, the creation of co-design frameworks that systematically integrate optimization models with distributed control architectures, potentially through quantum-assisted computing approaches, could help resolve the current scalability–decentralization impasse. Second, the development of adaptive policy mechanisms, particularly for blockchain implementations, is essential to ensure regulatory compliance while maintaining system efficiency. Finally, the establishment of standardized benchmarking metrics across domains will be crucial for objectively comparing decentralized solutions and guiding future research investments. These advancements, combined with the continued refinement of component technologies and validation methodologies, will be essential for transforming DES from a promising theoretical framework into a practical, scalable solution for modern energy challenges. The interconnected nature of these challenges suggests that future progress will require unprecedented levels of interdisciplinary collaboration, combining insights from energy systems engineering, computer science, economics, and behavioral science. Only through such integrated approaches can we hope to develop DES solutions that are simultaneously robust, efficient, scalable, and adaptable to the complex realities of real-world energy systems. The findings presented here not only map the current state of DES research, but also chart a course for future investigation, highlighting both the most promising avenues for advancement and the most critical barriers that must be overcome.

5. Conclusions

This systematic review has consolidated the most relevant advancements in optimization strategies, multi-agent control architectures, blockchain-based platforms, robust energy management frameworks, and the integration of electric machines within P2P microgrid environments. The analysis covered 94 high-quality studies identified through a rigorous PRISMA-guided process, which allowed for a comprehensive and transparent synthesis of current research trends.

The findings confirm that stochastic–robust optimization and multi-agent reinforcement learning are among the most effective strategies for handling uncertainty and enabling scalability in decentralized systems. These approaches facilitate real-time coordination among distributed energy resources and prosumers, particularly under dynamic and volatile conditions. While blockchain technologies offer enhanced security, transparency, and automation in P2P trading environments, their current energy consumption, latency, and computational demands present limitations for low-scale or resource-constrained microgrids.

Simulation-based validation, testbed experiments, and field trials are increasingly bridging the gap between theoretical modeling and real-world deployment. However, persistent challenges remain, especially in integrating regulatory frameworks, modeling user behavior, and ensuring interoperability among heterogeneous devices and agents. Furthermore, although electric machines play a fundamental role in energy conversion and storage—particularly in V2G and V2X applications—their integration into intelligent control frameworks remains limited in scope and depth across current studies.

Future research should focus on the development of unified and scalable architectures that combine adaptive control, predictive optimization, efficient consensus protocols, and machine-integrated coordination schemes. Particular attention should be given to enhancing the modularity and interoperability of decentralized frameworks, improving trust management in P2P transactions, and ensuring the robustness of control systems under high variability in supply and demand. Cross-disciplinary approaches that incorporate behavioral economics, regulatory foresight, and hardware-in-the-loop validation will be critical for transitioning from theoretical designs to operational distributed energy ecosystems.

Ultimately, this review underscores that no single technological strategy is sufficient on its own. Progress in the field depends on the convergence of algorithmic intelligence, secure digital infrastructures, resilient control schemes, and physical assets—such as electric machines—that enable energy flow. Advancing toward sustainable, self-organizing, and intelligent microgrids will require hybrid solutions that respond flexibly and reliably to the increasing complexity of distributed energy systems.