3.5.2. Agent Communication and Ontology

Agent communication in MAS can be accomplished in two ways: immediate communication among agents and interaction in a unitive environment [13]. MAS usually implements higher-level communication and supports reasoning abilities based on the Agent Communication Language (ACL) and a common vocabulary defined in an ontology [115]. The agent communication and ontology design in the selected literature is similar to the combined stages of ontology design, agent modeling, agent interaction, and specification of agent behaviors stages proposed in [110], and usually consist of standards for agent communication, interoperability, and ontology design.

• Standards for agent communication and interoperability

A standard for the communication between agents has been proposed by the Foundation for the Intelligent Physical Agent (FIPA [116]. The FIPA standards have been popularly used by MAS developers in the computer science community and FIPA was formally accepted as a standards committee of the IEEE Computer Society In 2005 [109]. Such standardization promotes open specifications for the interoperability between agents and MAS [117]. The FIPA standards include specifications for the agent communication language, communicative acts, content languages, and message transport protocols. It also includes a standard that proscribes the agents that a MAS must implement to be FIPA compliant

The FIPA-ACL specifies the syntax, the content of the message provides the semantics of the message including the content language and the ontology [118]. The messages built under the ACL structure allow the definition of various elements (e.g., performative, sender, receiver, content, language, and ontology, among others) and various communicative acts (e.g., agree, cancel, confirm, not-understood, etc.) [9]. Meanwhile, the correct interpretation of the meaning of the message is assured, the ambiguity is removed about the content [21]. The MASs in [3,47,48,52] all apply the FIPA-ACL. There are other ACL investigated in the literature as well [119,120], e.g., Open Agent Architecture (OAA) in the work of Praca et al. [42].

MASs developed by different platforms can interoperate with these FIPA standards, but it doesn't mean that useful information can be shared between agents if the MASs employ different ontologies [21,91]. It requires MASs share a common vocabulary, so the messages may be interpreted correctly among agents [47]. Therefore, ontologies are used to enabling the standardization of communications and interpretation of concepts between MASs [48].

The IEEE standards committee has identified the challenge of interoperable protocols, data formats and meaning and stated that open communication between smart devices using common protocols is crucial to interoperability [121]. Some standards in the power systems promote interoperability between devices within substations and open interfaces between energy management systems [91,109]. The most widely applied standard in the power system is the IEC 61970 Common Information Model (CIM), and its distribution management extension IEC 61968 [122].

IEC 61970 Standard is proposed by the International Electrotechnical Commission (IEC) to discuss and plan a variety of electrician and electron standards in order to procure international cooperation. IEC 61970 Standard defines the application program interface (API) of the energy management system is promulgated by IEC No.57 technical commission (Group 13) [13]. There are five main parts in the IEC 61970 standard: introduction and basic request, glossary, common information model (CIM), and two levels of component interface specification (CIS).

The CIM is a three-layer domain model, it defines a common vocabulary to describe the basic components used in electricity transportation and distribution [38], and CIM aims to facilitate power management processes (e.g., outage management, asset management, and customer information management) [50].

To achieve coherent and advantageous cooperation between different power systems, some reference models and frameworks are also popular used, e.g., SGAM (https://sgam-toolbox.org/) (the Smart Grid Architectural Model), USEF (https://www.usef.energy/) (the Universal Smart Energy Framework), and SEAS knowledge model (https://www.the-smart-energy.com/) (Smart Energy Aware Systems).

The Open Automated Demand Response (OpenADR) (https://www.openadr.org/) and energy@home (http://www.energy-home.it/SitePages/Home.aspx) models are also highly discussed in the literature. However, [50] states that 'none of these standards cover the whole semantics involved in a flexible urban energy network on its own, and they are not formally aligned with each other'. For example, the term 'equipment' could refer to transmission system equipment, or domestic appliance equipment [50].
