Assessing Agri-Food Waste Valorization Challenges and Solutions Considering Smart Technologies: An Integrated Fermatean Fuzzy Multi-Criteria Decision-Making Approach

Abstract

With the growth of the worldwide population and depletion of natural resources, the sustainable development of food systems cannot be ignored. The demand for agri-food waste valorization practices like high-value compounds production has received widespread attention; however, numerous challenges still exist. The present study aims to identify those challenges of agri-food waste valorization and propose effective solutions based on smart technologies. Based on a systematic review of the literature, the study combs existing challenges of agri-food waste valorization and constructs a six-dimension conceptual model of agri-food waste valorization challenges. Moreover, the study integrates a Fermatean fuzzy set (FFS) with multi-criteria decision-making (MCDM) methods including stepwise weight assessment ratio analysis (SWARA), decision-making trial and evaluation laboratory-interpretative structural modeling method (DEMATEL-ISM), and quality function deployment (QFD) to evaluate the weights of each dimension, find causal interrelationships among the challenges and fundamental ones, and rank the potential smart solutions. Finally, the results indicate that the “Government” dimension is the severest challenge and point out five primary challenges in agri-food waste valorization. The most potential smart solution is the “Facilitating connectivity and information sharing between supply chain members (S8)”, which may help government and related practitioners manage agri-food waste efficiently and also facilitate circular economy.

1. Introduction

In the face of a mounting global food crisis, where millions of people grapple with severe food insecurity, the issue of agri-food waste stands as a stark contradiction, exacerbating the problem. The 2024 Global Report on Food Crises by the Food and Agriculture Organization (FAO) highlights that despite the dire need, approximately 281 million individuals worldwide suffer from inadequate access to nutritious food, while an astonishing 1.3 billion tons of food waste is generated annually, accounting for 13.8% of global food production [1,2]. Of particular concern is the agri-food system, which not only contributes significantly to this waste but also bears the brunt of its environmental consequences [3,4]. The prevalent practice of landfilling agri-food waste not only squanders valuable resources but also emits greenhouse gases and pollutes groundwater, posing a threat to both ecological balance and human health. Indeed, different from other waste, agri-food waste, rich in complex carbohydrates and bioactive compounds, presents a treasure trove of untapped potential for the production of value-added products [5]. To be specific, an abundance of biochemicals are plant-derived, with a lesser amount derived from animals including pomace, peels, leaves, meat by-products, and so on [6]. Those bioactive compounds constitute a broad spectrum of molecules with unique structures and properties. They can be utilized in the manufacture of bio-fertilizers, fuel, compost, cosmetics, and functional foods [7]. Thus, the valorization of agri-food waste (AFW) has emerged as a possible way for the transformation and sustainable development of the global agri-food system to be realized [8]. It is also considered to have a substantial impact on the United Nations’ Sustainable Development Goals, particularly SDG 2 (zero hunger) and SDG 12 (responsible consumption and production) [9]. By transforming agri-food waste into value-added products like bio-fertilizers, fuel, compost, cosmetics, and functional foods, the valorization process reduces food waste and contributes to food security. This directly aligns with SDG 2, which aims to end hunger and achieve food security and improved nutrition. Moreover, valorizing agri-food waste promotes circular economy practices, reducing waste generation and encouraging the use of resources more efficiently. This not only mitigates the environmental impact of waste disposal but also fosters sustainable production and consumption patterns. Overall, successful agri-food waste resource utilization can bring significant economic, environmental, and social benefits (

Table 1. Benefits of successful agri-food waste valorization.

Economic	Environmental	Social
Increasing revenue sources: High value-added products can create new sources of income.	Reducing greenhouse gas emissions: Resource utilization can reduce emissions from incineration.	Enhancing public awareness: Successful resource utilization cases can promote public participation in waste management.
Saving costs: By optimizing waste management processes, waste disposal cost can be reduced.	Conserving water: Extracting valuable compounds from waste can reduce dependence on natural resources, especially water resources.	Creating employment opportunities: Developing waste resource utilization industry requires human resource support, creating new employment opportunities.
Bringing investment return: The investment of the government and enterprises can bring long-term economic returns.	Protecting ecological environment: Resource utilization can reduce waste pollution to soil, water and air, and protect the diversity of ecosystems.	Increasing social trust: Transparent and traceable waste management processes can enhance consumer trust in product.

).

In fact, the full utilization of agri-food waste for production of value-added materials remains largely untapped, although its considerable potential has been recognized [10]. On one hand, the valorization of agri-food waste is challenged by its intrinsic complexity, which is marked by heterogeneous composition, short lifespan, distribution pattern [11], and environmental sensitivity. Taking heterogeneous composition for instance, agri-food waste comprises a wide variety of materials, including pomace, peels, leaves, meat by-products, etc. Each component has unique biochemical properties, requiring tailored processing methods. On the other hand, the operations of agri-food waste valorization encompasses a multifaceted procedure including gathering, transportation, storage, treatment, and final disposal [12]. During these processes, a range of challenges, including environmental, social, and economic issues, are likely to emerge [13,14]. Hence, it is imperative to sort out and analyze the barriers that hinder the execution of agri-food waste valorization.

The aforementioned dual barriers to valorizing agri-food waste have persisted as enduring challenges, proving recalcitrant to resolution via conventional technological means. Recent advancements have seen a notable rise in innovative waste management strategies that harness smart technologies aligned with Industry 4.0 principles for enhanced efficiency and effectiveness. For instance, smart technologies facilitate the transition from conventional waste management systems to novel frameworks incorporating smart sensors, enabling real-time monitoring and fostering a sophisticated management infrastructure. Pertaining to the agri-food sector, a myriad of smart technologies, particularly big data analytics (BDA), blockchain, artificial intelligence (AI), Internet of Things (IoT), digital twins, smart sensors and robotics, and Information and Communication Technology (ICT), could revolutionize traditional practices, enhance efficiency, and promote sustainability [15]. The integration of AI, BDA, and IoT technologies can facilitate the application of automation and robotics in waste management, which reduces labor costs and human errors. Through IoT technology, combined with blockchain, RFID tags, and GIS, the entire agri-food supply chain can be tracked, which helps to promptly identify and respond to potential risks, ensuring the safe disposal and reuse of waste. Artificial intelligence (AI), which involves programming computers to mimic human behaviors like machine learning, artificial neural networks, and deep learning, offers immense potential for data-driven science within agri-food supply chains, especially synergized with high-performance computing technologies [16]. Moreover, some business modes comprehensively leverage big data analytics (BDA) to extract value from agri-food waste, thereby optimizing the existing linear supply chain [17]. Likewise, through the integration of smart technology and e-commerce, a digital platform could address the inefficiencies in agri-food waste management by aggregating and analyzing waste data, and identifying potential business collaborators who may repurpose agri-food waste into commercially valuable products [18]. Nevertheless, the introduction of these technologies in organizations without meticulous planning and scientific analysis is doomed to be unproductive and may even incur substantial financial burdens for the organizations [12]. As a result, it is crucial to contemplate mitigation strategies combining smart technologies, particularly in the context of specific challenges associated with the valorization of agri-food waste.

This study aims to tackle several critical research questions:

Identify and prioritize challenges: Uncover the primary obstacles that hinder the socialization and standardization of agri-food waste valorization;

Assess challenges priority: Determine the challenges that should be addressed first, considering their priority and limited resources.

Systematize interrelationships: Systematically map the intricate causal and hierarchical relationships among these challenges.

Explore smart technological solutions: Identify and evaluate the most effective smart technological solutions considering all factors comprehensively.

Guided by these research questions, the objectives of this paper are as follows:

Objective 1: Identify and prioritize the challenges associated with agri-food waste valorization.

Objective 2: Clarify the underlying causal and hierarchical relationships among these challenges.

Objective 3: Determine the smart technological solutions and validate the most feasible options for addressing challenges in the valorization of agricultural food waste.

To achieve these objectives, this study scrutinizes existing challenges in agri-food waste valorization and formulates a six-dimension conceptual model. Then, the Fermatean fuzzy stepwise weight assessment ratio analysis (FF-SWRAR) is applied to evaluate the significance of each dimension and challenge. Following this, the Fermatean fuzzy decision-making trial and evaluation laboratory-interpretative structural modeling method (FF-DEMATEL-ISM) is employed to discern cause-and-effect dynamics among the identified challenges and core challenges. Finally, the Fermatean fuzzy quality function deployment (FF-QFD) aids in ranking prospective smart technological solutions.

To the best of authors’ knowledge, prior study has not thoroughly investigated smart technological solutions for agri-food waste valorization. While a study has touched upon the integration of smart technology with biowaste valorization, it has been confined to a narrow perspective, such as the application of AI [19]. Additionally, our work firstly introduces an integrated framework that combines FFS, SWRAR, DEMATEL-ISM, and QFD methodologies, a combination that has never before been utilized to comprehensively assess both the challenges and potential solutions in agri-food waste valorization. Therefore, the study may provide valuable insights for related policymakers to devise strategies aimed at enhancing the valorization of agri-food waste, thereby contributing to the circular economy. Similarly, the agri-food waste valorization industry is expected to benefit from the study’s findings, which will guide the formulation of more effective and sustainable decisions.

2. Literature Review

This section discusses the concepts as challenges of agri-food waste valorization, smart technologies in agri-food system, and MCDM methods in waste management. At the end of the section, the research gap is highlighted.

2.1. Challenges of Agri-Food Waste Valorization

Currently, the majority of studies concerning the challenges of agri-food waste valorization predominantly concentrate on the exploration of a particular type of agri-food waste or a distinct valorization methodology from the viewpoints of biology and chemistry. For instance, considering the valorization of agri-food waste derived from olive oil and wine production, Tapia-Quirós et al. have advocated for the recovery of phenolic compounds as an effective way and elucidated techniques available for the analysis, extraction, and refinement of polyphenols from the olive mill and winery by-products [20]. Also, Mannaa et al. have proposed the integration of insects with organic waste in the bioconversion processes and accentuated the prospective efficacy of these biorefinery systems in surmounting the prevailing challenges associated with agri-food waste [21].

There are few works in the literature that study the challenges of agri-food waste valorization from a holistic perspective. Berenguer et al. have discussed some pivotal challenges in the valorization of agri-food wastes based on several perspective applications [6], so the scope of challenges identified are limited and the study lacks quantitative research and fails to probe into the significance and intrinsic interrelations of these challenges.

2.2. Smart Technologies in the Agri-Food Sector

The technological prowess of corporations is crucial in driving their innovative endeavors, which is viewed as one of the most significant dynamic competencies required to maintain enduring competitiveness [22]. In agri-food sector, the application of smart technologies provides the sustainable solutions to different agricultural problems [23]. Therefore, multiple studies have investigated the application status and emerging trends of smart technologies in the agri-food sector [24,25,26]. In terms of different regions, developed countries tend to exhibit a greater engagement with smart technologies [27]. Furthermore, among various smart technologies, the application of blockchain in agri-food supply chain has received more attention [28,29]. Similarly, regarding stakeholders within the agri-food supply chain, downstream companies are more willing to embrace smart technologies to cope with the uncertainty of the supply chain [30].

It is worth noting that although there is research in the literature introducing smart technologies for waste prevention and reduction in the agricultural food industry [31], it remains essential to thoroughly analyze the specific challenges encountered during the valorization of agri-food waste to determine the smart technological solutions that can be effectively employed and their priority in addressing these challenges.

2.3. MCDM Methods in Waste Management

Multi-criteria decision-making (MCDM) is a valuable approach for tackling complex decision-making scenarios where multiple criteria need to be taken into account [32]. It provides a structured framework to ensure a more informed and rational choice. In this context, it is evident that MCDM techniques are advantageous, as they enable a systematic comparison of challenges and strategies. Furthermore, it is common for decision-makers to articulate their subjective judgments through linguistic expressions in reality. This practice poses challenges when attempting to precisely model such information using crisp values. Consequently, to accommodate this imprecision, fuzzy set theory has been widely utilized in various cases [33].

In previous research related to waste management, fuzzy MCDM techniques have been commonly employed to assess challenges and formulate effective strategies. For example, Çelik et al. apply intuitionistic fuzzy multi-criteria decision-making (IFMCDM) methods to identify the most effective hospital for medical waste management in Erzurum, Turkey [34]. Komal integrates intuitionistic fuzzy sets (IFSs) with the weighted aggregated sum product assessment (WASPAS) method to assess health-care waste disposal methods [35]. Kabirifar et al. design a hybrid fuzzy MCDM approach to analyze nineteen factors influencing the management of construction and demolition waste [36].

The research methodology of this paper is developed based on a study conducted by Karuppiah [37]. The researcher combines Fermatean fuzzy set (FFS) with AHP, DEMATEL, and TOPSIS to explore e-waste mitigation strategies. In order to enhance the operability and pertinence of problem analysis, this study introduces another integrated Fermatean fuzzy multi-criteria decision-making approach (i.e., FF-SWRAR, DEMATEL-ISM, and QFD). In contrast to AHP, SWARA necessitates fewer pairwise comparisons for ascertaining weights, thereby rendering it a user-friendly approach for decision-makers [38]. Additionally, QFD is more oriented towards tackling specific issues, while TOPSIS primarily concentrates on the relative gaps between solutions [39].

2.4. Research Gap

Based on the above review and analysis, it is evident that while certain studies have explored agri-food waste management from specific perspectives, there remains a dearth of comprehensive examinations regarding the global challenges associated with agri-food waste valorization. Furthermore, the majority of existing research introduces the application of smart technology in the agri-food sector, yet lacks a quantitative analysis. This study endeavors to bridge the gaps by introducing a holistic evaluation framework for agri-food waste valorization challenges and solutions within uncertain environments that not only conduct an exhaustive investigation of diverse factors but also probe into their intricate relationships.

3. Methods

This section is comprised of the two following subsections: preliminaries and the research framework. In first subsection, the definition of FFS and related operation rules will be introduced in detail. In the second subsection, the overall research framework, including three major stages and integrated four-part methodology (i.e., FFS-SWRAR-DEMATEL-QFD), is described thoroughly, as shown in Figure 1.

3.1. Preliminaries

3.1.1. Definition of Fermatean Fuzzy Set

Definition 1. 

Assuming that $\mathrm{X}$ is a universe of discourse, a Fermatean fuzzy set $F$ on $\mathrm{X}$ is defined by Senapati and Yager as a function that applied to $\chi$ [40]:

$$F=\left\{〈\chi ,{\mu }_F(\chi ),{\nu }_F(\chi )〉|\chi \in \mathrm{X}\right\}$$

(1)

where ${\mu }_F(\chi )\in \left[0,1\right],{\nu }_F(\chi )\in \left[0,1\right]$ denote the degree of membership and non-membership of element $\chi \in \left[0,1\right]$, respectively, satisfying $0\le {\mu }_F{(\chi )}^3+{\nu }_F{(\chi )}^3\le 1$. For any FFS, the degree of indeterminacy of $\chi \in {\rm X}$ to $F$ is defined as:

$${\pi }_F(\chi )=\sqrt[3]{1-{\mu }_F{(\chi )}^3-{\nu }_F{(\chi )}^3}$$

(2)

In addition, $F=({\mu }_F,{\nu }_F)$ is called a Fermatean fuzzy number (FFN).

It is worth noting that FFS, an extension to IFS and PFS, has enlarged the domain of membership and non-membership, which is shown in Figure 2. Therefore, compared to IFS and PFS, FFS is more efficient in solving multi-criteria decision-making problems under uncertainty.

3.1.2. Related Operations for Fermatean Fuzzy Set

Definition 2. 

Let $F_1=({\mu }_{F1},{\nu }_{F1})$ and $F_2=({\mu }_{F2},{\nu }_{F2})$ be two FFNs, $\lambda >0$, defined as [40]:

• $F_1\oplus F_2=(\sqrt[3]{{\mu }_{F1}^3+{\mu }_{F2}^3-{\mu }_{F1}^3{\mu }_{F2}^3},{\nu }_{F1}{\nu }_{F2})$;
• $F_1\otimes F_2=({\mu }_{F1}{\mu }_{F2},\sqrt[3]{{\nu }_{F1}^3+{\nu }_{F2}^3-{\nu }_{F1}^3{\nu }_{F2}^3})$;
• $\lambda F_1=(\sqrt[3]{1-{(1-{\mu }_{F1}^3)}^{\lambda }},{\nu }_{F1}^{\lambda })$;
• $F_1^{\lambda }=({\mu }_{F1}^{\lambda },\sqrt[3]{1-{(1-{\nu }_{F1}^3)}^{\lambda }})$;

Definition 3. 

Let $F=({\mu }_F,{\nu }_F)$ be a FFN, the score function is defined as [33]:

$$score(F)={\mu }_F^3-{\nu }_F^3$$

(3)

For any FFN, $score(F)\in [-1,1]$.

• The accuracy function is defined as [33]:

$$accuracy(F)={\mu }_F^3+{\nu }_F^3$$

(4)

For any FFN, $accuracy(F)\in [0,1]$.

According to score and accuracy values, the comparison between any two FFNs $F_1=({\mu }_{F1},{\nu }_{F1})$ and $F_2=({\mu }_{F2},{\nu }_{F2})$ is determined:

If $score(F_1)<score(F_2)$, then $F_1<F_2$;

If $score(F_1)>score(F_2)$, then $F_1>F_2$;

If $score(F_1)=score(F_2)$, then

 If $accuracy(F_1)<accuracy(F_2)$, then $F_1<F_2$;

 If $accuracy(F_1)>accuracy(F_2)$, then $F_1>F_2$;

 If $accuracy(F_1)=accuracy(F_2)$, then $F_1=F_2$.

Definition 4. 

Let $F_i=({\mu }_{Fi},{\nu }_{Fi}) (i=1,2,\dots ,n)$ be a set of FFNs, then a Fermatean fuzzy weighted average (FFWA) is calculated [37]:

$$FFWA(F_1,F_2,\dots ,F_n)=({\displaystyle \sum _{i=1}^n{\omega }_i}{\mu }_{Fi},{\displaystyle \sum _{i=1}^n{\omega }_i}{\nu }_{Fi})$$

(5)

where ${\omega }_i\in [0,1]$ is the weight of $F_i$ with ${\displaystyle \sum _{i=1}^n{\omega }_i}=1$.

3.2. Research Framework

Stage 1. Identifying and modeling the challenges associated with agri-food waste valorization.

Step 1. Identifying corresponding challenges through a systematic review of the literature.

The following framework is adopted to collect articles relevant to agri-food waste valorization [41].

1. Identification: Searching articles considering five aspects in the following order: (1) source type, (2) source quality and relevance, (3) search engine, (4) search period, and (5) search keyword.

2. Screening: Excluding articles returned from the search that do not completely satisfy search criteria and some duplicate copies.

3. Eligibility: Assessing full text to make sure content relevance.

4. Inclusion: Performing a countercheck and a content analysis on the curated articles.

Step 2. Constructing conceptual model of agri-food waste valorization challenges.

5. Classifying and Modeling: Subsequent to the initial step of investigating and analyzing publications, the challenges of agri-food waste valorization are divided into a six-dimensional conceptual model by experts.

Stage 2. Evaluating the weights of challenges and elucidating relationships between them.

This stage predominantly uses FF-SWRAR to calculate initial weights of indicators. Then, the FF-DEMATEL-ISM method is applied to figure out causal relationship and influence degree among the identified indicators.

Step 3. Estimating the indicators’ initial weights using FF-SWRAR.

6. Evaluating the expertise level of decision makers [42]: The expertise of each DM is appraised through linguistic expressions delineated in

Table 2. Linguistic terms of decision makers’ expertise level.

Linguistic Terms	Absolute Expertise (AE)	High Expertise (HE)	Moderate Expertise (ME)	Less Expertise (LE)	No Expertise (NE)
$\mu$	0.95	0.75	0.55	0.3	0.1
$\nu$	0.1	0.3	0.55	0.75	0.95

along with corresponding FFS equivalents.

Let $M$ represent the count of DMs within the collective. The expertise level of a given DM $m$, symbolized as $E_m=({\mu }_{\mathrm{m}},{\nu }_m)$, dictates the influence of the DM’s assessment in the decision procedure. The crisp number reflecting a DM’s assessment influence among all can be computed:

$${\eta }_m={\displaystyle \frac{1+{\mu }_m^3-{\nu }_m^3}{{\displaystyle \sum _{m=1}^M(1+{\mu }_m^3-{\nu }_m^3)}}}$$

(6)

7. Constructing a linguistic decision matrix for the evaluation of indicators: The linguistic terms infer the linguistic assessment rating of an indicator and further turn into FFN (

Table 3. Linguistic terms of indicators.

Linguistic Terms	$\mathit{\mu }$	$\mathit{\nu }$
Absolutely Important (AI)/Absolutely High Related (AHR)	0.99	0.10
Very Strong Important (VSI)/Very High Related (VHR)	0.90	0.20
Strong Important (SI)/High Related (HR)	0.80	0.30
Important (I)/Medium High Related (MHR)	0.65	0.40
Equally Important (EI)/Exactly Equal Related (EER)	0.50	0.50
Unimportant (U)/Medium Low Related (MLR)	0.35	0.70
Strong Unimportant (SU)/Low Related (LR)	0.20	0.80
Very Strong Unimportant (VSU)/Very Low Related (VLR)	0.10	0.90
Absolutely Unimportant (AU)/Absolutely Low Related (ALR)	0.01	0.99

) [43]. Consider a FF evaluation matrix $Q=[q_{im}]$ provided by experts, where each element $q_{im}=({\mu }_{im},{\nu }_{im})$ denotes the corresponding FFN for the linguistic evaluation of DM $m$ for indicator $i$.

8. Combining decision makers’ judgments: Let $N$ represent the cardinality of indicator set where $n=1,2,\dots ,N$. Considering expertise weights, the judgments of all DMs on an indicator are aggregated as follows:

$$I_i=({\displaystyle \sum _{m=1}^M{\eta }_m}{\mu }_{im},{\displaystyle \sum _{m=1}^M{\eta }_m}{\nu }_{im})$$

(7)

9. Calculating the comparative significance of each indicator: Firstly, the positive score of each indicator, symbolized as $P{S}_i$, is determined as: $P{S}_i=1+score(I_i)$.

Then, rank the indicators in descending order according to the values of $P{S}_i$.

Based on the order, the comparative significance $C{S}_i$ of each indicator is calculated as:

$$C{S}_i=\left\{\begin{array}{ll}0& i=1\\ P{S}_i-P{S}_{i-1}& i>1\end{array}\right.$$

(8)

10. Computing the indicator weights [44]: Firstly, the comparative coefficient $C{C}_i$ is estimated as:

$$C{C}_i=\left\{\begin{array}{ll}1& i=1\\ C{S}_i+1& i>1\end{array}\right.$$

(9)

Then, the recalculated weight $q_i$ of each indicator is determined as:

$$q_i=\left\{\begin{array}{ll}1& i=1\\ {\displaystyle \frac{q_{i-1}}{C{C}_i}}& i>1\end{array}\right.$$

(10)

Finally, the initial weight of each indicator is calculated as:

$$w_i={\displaystyle \frac{q_i}{{\displaystyle \sum _{i=1}^N{q}_i}}} i>1$$

(11)

Step 4. Specifying the relationships between the indicators and adjusting weights of challenges using FF-DEMATEL-ISM.

11. Establishing the FF direct relationship matrix: DMs make pairwise comparisons of indicators to obtain mutual influence strength using

Table 4. Linguistic terms of influence score.

Linguistic Terms	Influence Score	FFN
Very High (VH)	4	(0.9,0.1)
High (H)	3	(0.7,0.2)
Low (L)	2	(0.4,0.5)
Very Low (VL)	1	(0.1,0.75)
No influence (NO)	0	(0,1)

[45], where influence data among the indicators are expressed by FFN.

12. Constructing aggregate FF direct relationship matrix: Use FFWA operator to aggregate the judgments of multiple DMs as follows:

$$A=\left[\begin{array}{c}\left({\mu }_{F11},{\nu }_{F11}\right) \left({\mu }_{F12},{\nu }_{F12}\right)\dots \left({\mu }_{F1n},{\nu }_{F1n}\right)\\ \left({\mu }_{F21},{\nu }_{F21}\right) \left({\mu }_{F22},{\nu }_{F22}\right)\dots \left({\mu }_{F2n},{\nu }_{F2n}\right)\\ \dots \\ \left({\mu }_{Fn1},{\nu }_{Fn1}\right) \left({\mu }_{Fn2},{\nu }_{Fn2}\right)\dots \left({\mu }_{Fnn},{\nu }_{Fnn}\right)\end{array}\right]$$

(12)

13. Defuzzification [46]: The FF defuzzification function $\phi$ is employed to turn the FFN matrix $A$ into crisp number matrix $X$ as follows:

$${\phi }_{ij}=1+score\left({\mu }_{Fij},{\nu }_{Fij}\right)$$

(13)

$$X=\left[\begin{array}{c}{\phi }_{11} {\phi }_{12}\dots {\phi }_{1n}\\ {\phi }_{21} {\phi }_{22}\dots {\phi }_{2n}\\ \dots \\ {\phi }_{\mathrm{n}1} {\phi }_{n2}\dots {\phi }_{nn}\end{array}\right]$$

(14)

14. Normalization: The new aggregate direct relationship matrix $X$ is normalized using following equations:

$$G=s^{-1}X$$

(15)

where $s=\max (\underset{1\le i\le n}{\max }{\displaystyle \sum _{j=1}^n{x}_{ij}},\underset{1\le j\le n}{\max }{\displaystyle \sum _{i=1}^n{x}_{ij}})$.

15. Constructing total relationship matrix $T$:

$$G=s^{-1}X$$

(16)

where $I$ is the identity matrix.

16. Classifying indicators into cause and effect groups as follows:

$$D={\left({\displaystyle \sum _{j=1}^n{t}_{ij}}\right)}_{1\times n}={(t_i)}_{1\times n}$$

(17)

$$C={\left({\displaystyle \sum _{i=1}^n{t}_{ij}}\right)}_{n\times 1}={(t_i)}_{n\times 1}$$

(18)

The value of $C+D$ represents centrality, while the value of $C-D$ represents causality.

17. Adjusting the weights of challenges: Combine centrality and initial weights calculated by SWRAR to obtain final weights of challenge using weighted average method. The specific weights are determined by relevant experts.

18. Obtaining initial reachability matrix (IRM): According to the following formula, total relationship matrix $T$ is converted to the initial reachability matrix $R$. The threshold $\lambda$ can be set based on the sum of mean and standard deviation in statistical distribution, effectively reducing subjective influence [47].

$$R=\left\{\begin{array}{ll}1& t_{ij}\ge \lambda \\ 0& t_{ij}\le \lambda \end{array}\right.$$

(19)

19. Constructing final reachability matrix (FRM): To obtain the FRM, the transitivity of the IRM is examined. According to the transitivity rule, if factor $i$ has an impact on factor $j$, and if factor $j$ affects factor $k$, then factor $i$ also impacts factor $k$ [48].

20. Partitioning level: A level partitioning operation was performed to acquire the reachability, antecedent, and intersection set.

Stage 3. Ranking the potential smart solutions to agri-food waste valorization.

In the stage, some solutions considering smart technologies are proposed to promote valorization of agri-food waste. Then, the FF-QFD method is utilized to prioritize them.

Step 5. Identifying smart agri-food waste valorization solutions.

21. Identifying strategies in the perspective of smart technologies: Based on the relevant literature and experts’ suggestions in the field, some potential strategies are provided.

Step 6. Prioritizing smart agri-food waste valorization solutions using FF-QFD.

The steps of FF-QFD are explained as follows:

22. Specifying the indicators: The indicators (i.e., challenges and solutions) have been decided in step 1 and 4.

23. Obtaining the importance weights of challenges: Each challenge has been evaluated based on FF-SWRAR.

24. Defining relationships between challenges and solutions: DMs use the scale as in Table 2 to define the relationship matrix $R_{ij}\left(i=1,2,\dots ,n,j=1,2,\dots ,k\right)$. If there is no relationship between the challenge and solution, the cell is left blank.

25. Calculating the relative importance of solutions: The relative importance $R{I}_j\left(j=1,2,\dots ,k\right)$ of solution $j$ is determined using FFWA operator as:

$$R{I}_j={\displaystyle \sum _{i=1}^n{w}_i{R}_{ij}}=({\displaystyle \sum _{i=1}^n{w}_i}{\mu }_{Fi},{\displaystyle \sum _{i=1}^n{w}_i}{\nu }_{Fi})\left(j=1,2,\dots ,k\right)$$

(20)

26. Creating correlation matrix: The correlations $S_{j{j}^{\prime }}\left(j\ne j^{\prime }\right)$ between solutions are created using the scale as in Table 2. There are three states that described an interrelationship: positive (+), negative (−), or non-existent (designated by a blank box).

27. Calculating score value for positive and negative correlations: Aggregate DMs’ judgments of correlation matrix by FFWA operator and calculate final score value.

28. Finding absolute importance for each solution [49]: The absolute importance $A{I}_j\left(j=1,2,\dots ,k\right)$ for solution $j$ can be computed as:

$$A{I}_j=R{I}_j\oplus {\displaystyle \sum _{j^{\prime }=1}^k{S}_{j{j}^{\prime }}\otimes R{I}_{j^{\prime }}}(j=1,2,\dots ,k;j^{\prime }\ne j)$$

(21)

29. Obtaining final score value of solutions and ranking them: Use FF defuzzification function to obtain crisp value of each solution and prioritize these solutions.

The integrated Fermatean fuzzy MCDM approach, comprising FFS-SWRAR, FFS-DEMATEL-ISM, and FFS-QFD, significantly contributes to the comprehensive and structured analysis of agri-food waste valorization challenges and solutions. By addressing uncertainty through FFS, simplifying weight calculation with SWRAR, unraveling causal dynamics with DEMATEL-ISM, and prioritizing solutions with QFD, this approach ensures a more informed and rational decision-making process for policymakers and practitioners.

4. Results and Discussion

In the preceding section, the general outline of a complete study has been established. Detailed calculations and corresponding results of the FF-MCDM studies are described in the following subsections.

4.1. Results

According to the three main stages of research framework, the applied procedure based on FF-SWRAR, FF-DEMATEL-ISM, and FF-QFD is summarized as follows:

Stage 1. Identifying and modeling the challenges associated with agri-food waste valorization.

Step 1. Identifying corresponding challenges through a systematic review of the literature.

The Web of Science databases were utilized to search for the following topics: “agri-food waste management”, “agro-food waste management”, “agri-food waste valorization”, and “agro-food waste valorization”. This search yielded 573 publications spanning from 2019 to April 2024. Following a series of screening procedures, a refined compilation of 43 articles was selected for further analysis. Then, the challenges of agri-food waste valorization were identified.

Step 2. Constructing conceptual model of agri-food waste valorization challenges.

Based on the literature [50], experts categorized the challenges into six distinct dimensions: organization, environment, technology, economy, government, and society. In this study, the classification framework could be constructed from a systems theory perspective. In this framework, the government, organization, and consumer are stakeholders of the agri-food waste valorization system, which constitute the internal core components, while the environment, economy, and technology are supporting and influencing factors, which constitute the external conditions for the operation of the system (

Table 5. Conceptual model of agri-food waste valorization challenges.

Dimensions	Codes	Factors	Codes	References
Organization	C1	Poor logistical and infrastructural systems	C11	[51,52,53,54]
		Less standardized operational practices	C12
		The absence of intermediary companies/departments collecting and directing wastes to specific points for processing	C13
		Rare cooperation between supply chain members in the process of agri-food waste valorization	C14
		Lack of instructions about approaches of agricultural waste valorization	C15
		No safety assessment of biotechnologically materials	C16
Environment (including biochemical property)	C2	The region-dependent and seasonal availability of a waste stream	C21	[55,56,57,58]
		Variable quality of the waste stream due to deterioration	C22
		New product safety issues like contamination of heavy metals	C23
		High sensitivity of microorganisms to operating conditions	C24
		High standard on properties of the raw materials like element proportion, moisture content	C25
		Production of environmental footprint in extraction processes	C26
Technology	C3	Lack of the most efficient and cost-effective extraction method for specific waste streams	C31	[54,59,60,61,62]
		Limited technological capabilities available for sorting, safe storing, and distribution of food waste	C32
		No full understanding of emerging technologies	C33
		Loss of biocompounds caused by conventional extraction technology	C34
		High energy consumption of technology	C35
Economy	C4	High transport costs due to collection and processing of biomasses	C41	[51,54,63,64,65]
		High expenses related to the techniques utilized	C42
		The shortage of investment in technologies/solutions	C43
Government	C5	Lack of robust and detailed legal and regulatory foundation	C51	[54,66,67,68]
		The absence of agri-waste management digital platforms	C52
		Lack of relevant incentive systems	C53
Customer	C6	Less trust of consumers in safety of new products based on agricultural by-products	C61	[51,67,69]
		Little public awareness about agri-food waste valorization	C62
		Obscure consumer acceptance due to changes in sensory quality	C63

).

Stage 2. Evaluating the weights of challenges and elucidating relationships between them.

Step 3. Estimating the indicators’ initial weights using FF-SWRAR.

In this step, the FF-SWRAR methodology was employed to determine initial weights of each challenge through assessment of three experts.

Table 6. Expertise levels of decision makers.

DM	Degree of Expertise	Influence of Assessment
$E_1$	HE	0.411
$E_2$	ME	0.295
$E_3$	ME	0.295

outlines the respective expertise levels of these three experts.

Furthermore,

Table 7. Weights of dimensions and challenges.

Dimensions	Factors	Local Weights of Challenges	Overall Weights of Challenges
C1 0.246	C11	0.23	0.05
	C12	0.21	0.02
	C13	0.20	0.06
	C14	0.16	0.05
	C15	0.11	0.04
	C16	0.09	0.03
C2 0.123	C21	0.24	0.02
	C22	0.20	0.02
	C23	0.18	0.02
	C24	0.16	0.01
	C25	0.14	0.03
	C26	0.09	0.02
C3 0.186	C31	0.31	0.05
	C32	0.26	0.06
	C33	0.20	0.02
	C34	0.13	0.02
	C35	0.10	0.04
C4 0.102	C41	0.39	0.03
	C42	0.35	0.04
	C43	0.26	0.04
C5 0.252	C51	0.33	0.08
	C52	0.36	0.09
	C53	0.31	0.08
C6 0.091	C61	0.44	0.04
	C62	0.31	0.03
	C63	0.25	0.02

provides a comprehensive overview of the local weights and overall weights assigned to each challenge.

Clearly, “Government (C5)” emerges as the paramount dimension among the challenges to agri-food waste valorization, closely followed by the “Organization (C1)”dimension. Moreover, the initial pivotal challenges are “The absence of agri-waste management digital platforms (C52)”, “Lack of relevant incentive systems (C53)”, “Lack of robust and detailed legal and regulatory foundation (C51)”, “Limited technological capabilities available for sorting, safe storing, and distribution of food waste (C32)”, and “The absence of intermediary companies/departments collecting and directing wastes to specific points for processing (C13)”.

Step 4. Specifying the relationships between the indicators and adjusting weights using FF-DEMATEL-ISM.

In this step, the FF-DEMATEL-ISM was used to clarify interrelationships among challenges.

Table 8. Causal relationships of challenges.

Factors	C	D	C + D	Rank	D − C	Category
C11	1.28	0.01	1.286	2	−1.267	effect
C12	0.60	0.09	0.692	8	−0.514	effect
C13	0.77	0.11	0.885	6	−0.657	effect
C14	0.32	0.36	0.676	10	0.036	cause
C15	0.20	0.36	0.556	13	0.157	cause
C16	0.22	0.23	0.457	19	0.008	cause
C21	0.00	0.24	0.239	25	0.239	cause
C22	0.38	0.11	0.491	16	−0.264	effect
C23	0.41	0.12	0.532	15	−0.291	effect
C24	0.00	0.28	0.281	24	0.281	cause
C25	0.09	0.00	0.093	26	−0.093	effect
C26	0.35	0.00	0.346	23	−0.346	effect
C31	0.39	0.74	1.138	3	0.351	cause
C32	0.29	0.47	0.752	7	0.180	cause
C33	0.29	0.39	0.688	9	0.100	cause
C34	0.42	0.00	0.420	20	−0.420	effect
C35	0.17	0.18	0.354	22	0.008	cause
C41	0.37	0.23	0.593	12	−0.141	effect
C42	0.39	0.16	0.551	14	−0.237	effect
C43	0.44	0.66	1.098	5	0.226	cause
C51	0.17	0.31	0.477	17	0.140	cause
C52	0.07	1.04	1.113	4	0.971	cause
C53	0.18	1.34	1.524	1	1.160	cause
C61	0.41	0.06	0.471	18	−0.343	effect
C62	0.00	0.60	0.599	11	0.599	cause
C63	0.00	0.40	0.404	21	0.404	cause

presents the identified causal relationships.

Table 9. Hierarchical structure of challenges.

Level	Factors
1	C25
2	C11, C61
3	C13, C23, C26, C34
4	C12, C22, C35, C41, C42
5	C14, C16, C31, C32
6	C15, C33, C43, C51
7	C62, C63
8	C21, C24, C52, C53

illustrates the derived hierarchical structure.

Given the values of C−D in Table 8, the challenges have been categorized into cause-and-effect groups, as depicted in Figure 3. In the cause group, the most important challenges are “Lack of relevant incentive systems (C53)” and “The absence of agri-waste management digital platforms (C52)”. In the effect group, the most important challenges are “The absence of intermediary companies/departments collecting and directing wastes to specific points for processing (C13)” and “Poor logistical and infrastructural systems (C11)”. Based on C and D values, the prominence of the critical factors have been evaluated. The top five ranked challenges are “Lack of relevant incentive systems (C53)”, “Poor logistical and infrastructural systems (C11)”, “Lack of the most efficient and cost-effective extraction method for specific waste streams (C31)”, “Lack of robust and detailed legal and regulatory foundation (C51)”, and “The shortage of investment in technologies/solutions (C43)”.

After adjusting weights, the final most important challenges are “Lack of relevant incentive systems (C53)”, followed by “The absence of agri-waste management digital platforms (C52)”, “Poor logistical and infrastructural systems (C11)”, “Lack of the most efficient and cost-effective extraction method for specific waste streams (C31)”, and “The absence of intermediary companies/departments collecting and directing wastes to specific points for processing (C13)”.

According to the analysis results of ISM (Table 9), the challenges of agri-food waste valorization can be divided into eight levels. The essential causal factors at the bottom level are “The region-dependent and seasonal availability of a waste stream (C21)”, “High sensitivity of microorganisms to operating conditions (C24)”, “Lack of relevant incentive systems (C53)”, and “The absence of agri-waste management digital platforms (C52)”.

Stage 3. Ranking the potential smart solutions to agri-food waste valorization.

Step 4. Identifying smart agri-food waste valorization solutions.

An in-depth investigation was conducted in the Web of Science databases, encompassing all existing publications about the utilization of smart technologies in agri-food waste valorization. Additionally, insights from professional experts were solicited. Consequently, a total of 18 innovative smart solutions were identified (

Table 10. Smart solutions to agri-food waste valorization.

Codes	Solutions	References
S1	Employing AI to predict and classify the properties or characteristics of biowaste	[17,25,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84]
S2	Utilizing AI to predict the volatile organic compounds (VOCs) and supply for waste materials
S3	Improving transparency and safety of agri-food supply chains through contamination tracing and efficient food production system, e.g., IoT, blockchain, big data, RFID tags, GIS
S4	Obtaining real-time and up-to-date digital information on crop growth, safety, and nutrition by UAVs, cloud computing, GIS
S5	Using digital devices and platforms in rural agriculture as early warning system by using ICT, RFID tags, remote sensors
S6	Cooperating between technology providers and adopters to advance sustainable agri-food supply chain management using remote sensors, weather forecasting systems, bio-stimulants
S7	Integrating innovative agricultural technologies with farmers’ traditional knowledge and constructing a knowledge-sharing platform
S8	Facilitating connectivity and information sharing between supply chain members
S9	Designing agri-food waste apps to link manufacturers, supermarkets, restaurants, and individual households
S10	Searching and analyzing current databases to guide the selection of suitable agri-food waste valorization approach through AI
S11	Identifying the exact parameters in the operational process based on BDA together with the sensors
S12	Automatically identifying consumer needs to inform manufacturers and retailers utilizing text mining and information sharing platform
S13	Applying IoT to monitor environmental parameters like temperature, dissolved oxygen and pH in the production process
S14	Using intelligent algorithms for site selection and transportation path planning
S15	Minimizing the carbon footprint of the entire supply chain by cloud computing
S16	Implementing autonomous robots to reduce costs and improve operational professionalism
S17	Increasing awareness of cybersecurity at all stages of the supply chain
S18	Adopting digital twins to evaluate agricultural food waste quality and tailor supply chains to reduce losses

).

Step 5. Prioritizing smart agri-food waste valorization solutions using FF-QFD.

As depicted in

Table 11. Importance of smart solutions.

Codes	Absolute Importance	Rank
S1	0.016	10
S2	0.142	3
S3	0.154	2
S4	0.069	5
S5	0.042	6
S6	0.008	12
S7	0.020	9
S8	0.365	1
S9	0.099	4
S10	0.014	11
S11	0.002	17
S12	0.021	8
S13	0.003	14
S14	0.003	15
S15	0.004	13
S16	0.002	16
S17	0.002	18
S18	0.034	7

, the results of FF-QFD analysis reveal that in addressing current challenges, the most highly prioritized solutions are “Facilitating connectivity and information sharing between supply chain members (S8)”, “Improving transparency and safety of agri-food supply chains through contamination tracing and efficient food production system e.g., IoT, Blockchain, Big Data, RFID tags, GIS (S3)”, and “Utilizing artificial intelligence (AI) to predict the volatile organic compounds (VOCs) and supply for waste materials (S2)”.

4.2. Discussion

This section focuses on the in-depth analysis of the aforementioned results. In terms of different dimensions of challenges, “Government (C5)” ranks the highest. Typically, the local government assumes a guiding role in a project, with its primary responsibility being to facilitate the participation of enterprises and the public. Particularly in the case of agri-food waste value-added initiatives, which necessitate substantial initial investments and yield returns over an extended duration, the role of governmental guidance and backing is imperative. In an empirical study, Xiang and Gao prove that government support exerts a remarkably positive influence on the sustainable development of the agricultural sector [85]. Notably, agricultural extension services and ecological subsidies, as key constituents of government support, contribute significantly to agricultural sustainability. Furthermore, through evolutionary games, some scholars demonstrate that it is crucial to enhance government’s accountability and regulatory proficiency, robustly pursue technological advancements, and refine the incentive and disciplinary mechanisms to achieve both specialization and socialization of agricultural waste valorization [86]. The second important dimension is “Organization (C1)”. Related business organizations constitute a significant driving force in the generation of waste, as well as the innovation and utilization of Industry 4.0 technologies [87]. Therefore, organizations serve as the actual main participants responsible for the valorization of agri-food waste. Should there be a lack of active engagement, an absence of instructions on agri-food waste valorization methods, and infrequent collaboration with other supply chain members, they are prone to adopting unscientific and unsystematic practices in managing agri-food waste, overlooking potential flaws in the logistical and infrastructural systems. Taking Kampala city for example, to achieve environmental, economic, and technical goals within urban settings, related organizations should carefully choose suitable technology-driven systems for agri-food waste valorization [88].

According to the Pareto principle [89], also known as the “80/20” rule, a deeper analysis has been conducted on the top 5 challenges out of a total of 26 identified challenges. The foremost challenge lies in “Lack of relevant incentive systems (C53)”, which falls under the “cause” category. Actually, in Pakistan, Malaysia, and China, research finds that government incentives have a positive effect on the innovation of circular economy in small and medium enterprises [90]. Moreover, in Australia, lack of government incentive is a major barrier to developing a circular economy [91]. However, only a few countries, such as France, Italy, Austria, and Germany, have provided financial support in certain areas of agri-food waste valorization, but such financial support is only applicable to small-scale pilot projects and cannot be scaled up for large-scale promotion [92].

The following challenge is “The absence of agri-waste management digital platforms (C52)” belonging to the “cause” category. With regard to the governmental role, the traditional emphasis has predominantly centered on resources of financial wealth and administrative authority. However, other potential roles that governments could assume in fostering the development of agri-food waste valorization are often overlooked [93]. Specifically, there is a possibility for a government to leverage its central position within pivotal networks to gather advanced resources, thus creating a comprehensive digital platform to coordinate stakeholders and establish partnerships. Indeed, a key characteristic of the advancement of agri-food waste valorization lies in harnessing intricate networks of diverse actors, each possessing a range of requisite skills. In addition, the factor also highlights the necessity of using smart technology to address existing challenges.

The third important challenge is “Poor logistical and infrastructural systems (C11)”, within the “effect” category. The factor is significantly influenced by numerous other variables, especially “The absence of relevant incentive systems (C53)”, “The absence of agri-waste management digital platforms (C52)”, and “ The lack of the most efficient and cost-effective extraction method for specific waste streams (C31)”. These contributory factors largely constrain the effectiveness of logistical and infrastructural systems in managing agricultural waste. Due to factors C53 and C31, numerous agricultural enterprises bear elevated risks when confronted with substantial investments in technology, thereby deterring them from proactive upgrading of their current infrastructural facilities [94]. In addition, the valorization of agri-food waste is not feasible solely through the efforts of a single enterprise, but requires the collaboration across the entire industry chain and even societal engagement. Hence, the absence of a unified digital management platform (C52) poses a significant obstacle in achieving seamless and standardized logistics systems.

The next challenge is “The lack of the most efficient and cost-effective extraction method for specific waste streams (C31)” under the “cause” category. Extracting effective substances from agricultural food waste is a decisive step in the valorization of agricultural food waste. Taking the extraction of cellulose as an example, isolating cellulose from biomass poses a significant challenge due to the recalcitrant nature of biomass, which inherently limits the accessibility of cellulose for value-adding applications [95]. Furthermore, the diverse range of agri-food sources containing cellulose renders it exceedingly difficult to devise a standardized extraction method capable of efficiently recovering cellulose across all types of sources. It is recommended that the forthcoming five years should be dedicated to exploring the innovative thermal extraction technologies, with a comprehensive techno-economic analysis conducted to thoroughly assess the feasibility and effectiveness of implementing these technologies in the extraction process of agricultural byproducts [96].

The fifth significant challenge, classified under the “effect” category, pertains to “The absence of intermediary companies/departments collecting and directing wastes to specific points for processing (C13)”. In fact, as the waste bank is incapable of recycling the waste independently, the supply chain relies on a recycling factory to accomplish this task [97]. Besides the government dimension, the two most important influencing factors on the challenge are “Limited technological capabilities available for sorting, safe storing, and distribution of food waste (C32)” and “Rare cooperation between supply chain members in the process of agri-food waste valorization (C14)”. The former underscores the substantial resource allocation to streamline the procurement of agri-food waste, thereby guaranteeing consistency, microbial safety, and superior quality for processing of waste, which once again demonstrates the necessity of government and social support [54]. The latter reason is aligned with a finding that the conversion of food waste into valuable products necessitates a concerted effort spanning the entire value chain and adopting a comprehensive food system viewpoint, which entails a profound understanding of the boundaries stemming from the subject’s dynamic characteristics and interconnected dependencies [98].

The factors at the bottom level are fundamental factors. C21 and C24 are inherent attributes of the research subject. Specifically, the spatiotemporal distribution of agri-food waste and its high sensitivity to environment fundamentally impacts the cost and quality of biomass value-added processes. C53 and C52, in the “Government” dimension, play an external driving role in the valorization of agri-food waste, fully leveraging the aforementioned governmental prowess in resources and organization.

The subsequent discussion delves deeper into the top three solutions pertaining to smart technologies. Among these, the solution that emerges as the most effective is “Facilitating connectivity and information sharing between supply chain members by digital tools (S8)”. Enhanced visibility and transparency within the supply chain empower members to identify and mitigate risks in a more efficient manner, thereby reducing the likelihood of disruption, particularly considering region-dependent and seasonal availability of the waste stream. Additionally, through swift exchange of information, supply chain members respond promptly to changes in market conditions in regard to obscure consumer preference. The solution also contributes to the establishment of a comprehensive agri-food waste management platform on a large scale. Among the digital tools, big-data management appears to be the most suitable for achieving S8, given its capability to facilitate the collection and sharing of diverse data types among organizations, ultimately enhancing the accuracy of outcomes [70]. The second important solution is “Improving transparency and safety of agri-food supply chains to customers through contamination tracing and efficient food production system e.g., IoT, Blockchain, RFID tags (S3)”. Merely enhancing information exchange among enterprises within the supply chain is insufficient. It is essential to address the safety concerns of customers pertaining to new agri-food value-added products. Consequently, it becomes necessary to synchronize information derived from diverse production processes with customers to ensure their trust and satisfaction. In fact, the successful valorization of agri-food by-products heavily relies on robust traceability and rigorous quality monitoring in production and logistic system [29]. The third important solution lies in “Utilizing artificial intelligence (AI) to predict the volatile organic compounds (VOCs) (S2)”. In practice, the variability in feedstock derived from biowaste significantly hinders the widespread utilization of value-added products. To overcome the difficulties, the valorization of agri-food waste has embraced artificial intelligence (AI), a novel approach, as a potential solution. According to diverse components of biomass, the overall dataset for training and testing in AI learning and the application of AI algorithms is diverse [19].

5. Conclusions

This study advances the existing literature by proposing solutions to the challenges of agri-food waste valorization considering smart technologies in Industry 4.0. Through a comprehensive review of the literature and insights from agricultural experts, challenges have been identified and subsequently categorized into six distinct dimensions: organization, environment, technology, economy, government, and customer. Then, a novel integrated MCDM approach including FFS and SWRAR-DEMATEL-ISM-QFD is employed to evaluate the challenges and potential solutions in the light of expert insights. Based on the findings of the FF-SWRAR, the “Government” dimension emerges as the most crucial, with a significant weight of 0.252, indicating its importance in addressing the challenges of agri-food waste valorization. According to the final weights of challenges, the top five most pivotal challenges are C53, C52, C11, C31, and C13. Next, the FF-DEMATEL-ISM method divides these challenges into cause and effect groups with eight levels, identifying the fundamental factors. Finally, FF-QFD prioritizes smart technology solutions in accordance with the varying weight of current challenges. Among these, three solutions stand out as the most significant, as follows: S8, S3 and S2.

5.1. Theoretical Implications

This study categorizes agri-food waste valorization challenges into macro-dimensions, offering perspectives for cross-sector researchers to comprehend the issue comprehensively. Within the context of sustainability and digitization, it preliminarily explores smart tech-based solutions, inspiring agricultural managers to adopt scientific methods and foster tech advancements. Furthermore, it introduces a novel MCDM framework, uncommon in agri-food waste evaluation, which can be adapted across domains, bolstering result reliability.

5.2. Practical Implications

Drawing from the research outcomes, this study presents several managerial implications that are expected to benefit government agencies and other stakeholders engaged in the management of agri-food waste. For government, it requires more initiative or knowledge to foster the development of agri-food waste valorization. The government should establish reasonable incentive mechanisms to ensure the service quality of fiscal funds in the field of agri-food waste valorization. Therefore, the government should seize the opportunity of applying and promoting agri-food waste valorization to improve risk management and performance evaluation in the agricultural supply chain. Beyond financial investments, the government needs to engage more stakeholders and jointly construct a technology-supported ecosystem for agri-food waste management. The digital waste management platform is expected to be positioned as a more solution-oriented approach, leveraging the integration of smart technologies in a practical and innovative manner to address environmental and social issues, thereby assisting governments and enterprises in making scientific decisions. For supply chain members, they should also enhance information disclosure and technological innovation. The strategic integration of upstream and downstream enterprises in the supply chain is the first step. Cooperation with upstream enterprises with resource aggregation can greatly reduce the risks related to raw material supply, while cooperation with downstream enterprises with first-hand market information can reduce the risks of demand uncertainty. Secondly, as the immense operational pressures and high costs associated with adopting advanced technologies may hinder enterprises in technological innovation, a potential lightweight mitigation approach involves the training of current employees to collaborate with digital technology providers that offer modular solutions. For smart technology providers, it is recommended to adopt a platform-based business model rather than a product-centric one. By adhering to established data standards, it becomes feasible for data to traverse the entire waste management value chain with the waste stream, thereby facilitating end-to-end digitization.

5.3. Limitations and Future Research

There are some limitations of this study. Firstly, despite a diligent review of the literature, encompassing all existing research on agri-food waste valorization remains challenging, limiting the comprehensiveness of identified challenges. Future studies should expand on empirical surveys to fill this gap. Secondly, smart technology solutions’ practical implementation is complex, leading to limited detail in some proposed solutions. Further research should delve into precise smart technology applications for agri-food waste, with a more rigorous analysis. Lastly, while employing a Fermatean fuzzy framework, alternative uncertainty management methods warrant exploration and comparative analysis.