Multi-Agent Systems in ICT Enabled Smart Grid: A Status Update on Technology Framework and Applications

Multi-agent-based smart grid applications have gained much attention in recent times. At the same time, information and communication technology (ICT) has become a crucial part of the smart grid infrastructure. The key intention of this work is to present a comprehensive review of the literature and technological frameworks for the application of multi-agent system (MAS) and ICT infrastructure usages in smart grid implementations. In the smart grid, agents are defined as intelligent entities with the ability to take decisions and acting flexibly and autonomously according to their built-in intelligence utilizing previous experiences. Whereas, ICT enables conventional grid turned into the smart grid through data and information exchange. This paper summarizes the multi-agent concept of smart grid highlighting their applications through a detailed and extensive literature survey on the related topics. In addition to the above, a particular focus has been put on the ICT standards, including IEC 61850 incorporating ICT with MAS. Finally, a laboratory framework concepts have been added highlighting the implementation of IEC 61850.


I. INTRODUCTION
The word ''Smart grid'' has become a buzzword all over the world. A smart grid is defined by a bidirectional flow of electric power and data/information in order to establish a decentralized and intelligent electric power network among numerous interconnected elements (Generation-Transmission-Distribution) [1]. It is broadly accepted that ICT platform is playing one of the most vital roles in the establishment of smart grid concept. Due to the complex, distributive and highly unpredictable nature of smart grid, the control system has been evolving from centralized imposing architecture towards the decentralized, autonomous, flexible and robust structure. Multi-agent system (MAS) has emerged as an essential and useful tool to solve the implementation challenges of smart grid.
A multi-agent system is defined as a coalition of a number of intelligent agents cooperating in order to achieve The associate editor coordinating the review of this manuscript and approving it for publication was Francesco Piccialli. a specific objective, thus accomplishing the overall goal of the system. In contrast with Supervisory Control and Data Acquisition (SCADA), in a traditional power system, MAS based applications have introduced a new paradigm in developing decentralized systems including many features such as autonomy, sociality, reactivity, and pro-activity. In parallel with that, artificial intelligence has enabled MAS approaches to replace human operator in the power system especially in case of automated operation. In the literature, a wide range of works have been published on the topic of MAS which can be categorized into two-fold aspects from a broader point of view. The first one is the MAS concept, approach, design and challenges in its implementation [2]- [4]. The second one is the MAS based application in numerous technological categories such as grid control [5], protection [6], electricity market simulation [7], power system model [8], network operation [9], substation automation [10] and others. In addition to that, a number of review papers and surveys papers exist in the literature which summarizes MAS based application in the field of power VOLUME 7, 2019 This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see http://creativecommons.org/licenses/by/4.0/ engineering [11]- [14], MAS toolkits/software [15]- [17], MAS real-time applications [18], [19] and some other bibliographical analysis. However, little emphasis has been given on the ICT structure of the smart grid enabling an underlying MAS based operation. Simulations of potential case studies are also absent. It is then clear that MAS is becoming the norm in new developments for the smart grid. Since ICT is fundamental to its implementation, a new scope of MAS and ICT application in the smart grid is urgently needed. This work addresses the challenges for the integration of MAS and the ICT platform in the smart grid, additionally highlighting a protection mechanism following an IEC standard. So, the primary goal of this work is to introduce a brief overview of the modern decentralized smart grid employing multiple intelligent agents, its applications in different segments and the implementation of one of the communication standards-IEC 61850. The concept, architecture, benefits, and applications in the use of intelligent agents in the smart grid have been summarized following an extensive literature review. In addition to that, a detailed literature review of IEC 61850 has been performed in the context of substation automation, protection, control, and security. This work aims at the application of a MAS architecture enabled by a communication platform for the smart grid. This review is organized into seven sections. Section 2 presents the introduction of the modern decentralized smart grid and the concept behind it, Section 3 highlights the literature statistics of MAS research while the definition of agent concept, benefit, architecture, platforms is presented in section 4. Section 5 focuses on the introduction of communication platforms for the smart grid with a particular highlight on IEC 61850. The task of IEC 61850 for substation (Intelligent Electronic Devices) IED automation, control, and protection is briefly summarized with a detailed literature survey in this section. The applications of a multi-agent system (MAS) with an enabling ICT structure in the smart grid is presented in section 6 categorizing into different sections, and each section consists of a detailed literature survey of MAS. Finally, section 7 comprises a laboratory test framework employing the IEC 61850 standard for a smart grid protection system at Curtin University. Section 8 concludes the paper.

II. OVERVIEW OF ICT ENABLED SMART GRID
The Smart Grid idea has called for a revolutionary change in the concept of the conventional electric grid. The traditional way in which generation, transmission, distribution, control, and protection happens in the conventional electrical power grid is being completely transformed. The modern Smart Grid is idealized as an automated, resilient, intelligent structure. In other words, a Smart Grid is an intelligent electrical network combining all the parties involved in the grid -generators, consumers and prosumers to secure sustainable, efficient, reliable and economic electric power supply employing modern digital technologies, self-regulating monitor and control strategies and advanced information and communications technology (ICT). In the pathway to create the Smart Grid, modern and new edge technologies, innovative products and services facilitate the environmentally friendly distributed generation, the reliable and efficient transmission and distribution of energy, the resource-cheap and real-time monitoring, with unsupervised control, and unlimited protection of the grid, the efficient communication, the optimal demand response, the necessary energy storage facilities, and self-healing technologies. The whole power grid network and participating parties including distributed energy resources (DERs), distribution system operators (DSOs), transmission system operators (TSOs), consumers/prosumers and market mechanisms are continuously evolving around the smart grid to maintain its 'smart'' status quo. In this regard, digital information and communication technology (ICT) plays a very vital and core part in the evolution of this smart grid which has been depicted in Fig. 1. In the generation part of the smart grid, Distributed Energy Resources (DERs) brings the solution of environment-friendly and sustainable energy production in parallel with the conventional energy plants [20], [21]. In the transmission level, Synchrophasor based WAMS (Wide area monitoring system) technology has emerged as a game changer and transforms the grid into a smart one [22]. WAMS facilities real-time monitoring of complex power grid across larger areas employing PMUs (Phasor measurement units) and thus outshines the traditional Supervisory Control and data acquisition (SCADA) systems [23], [24]. The smart distribution system includes functionalities like self-healing of grid [25], smart energy management [26], intelligent monitor and control based on smart automation devices or intelligent nodes [27], integration with EVs (Electric Vehicles) [28], facilitating energy storage technology [29] and so on. The role of consumers in the smart grid has changed drastically and with the increased penetration of DERs, nowadays consumers are defined as prosumers and are contributing towards bidirectional power flow in the smart grid [30]. Additionally, consumers and prosumers have started to participate in demand-side energy management (DSEM) through the smart metering system and smart grid ICT market mechanism [31]. Different communication technologies have been employed via two major media: wired and wireless. Wired communication medium includes Ethernet, power line communication-PLC, optical fiber and wireless communication comprises satellite, WiMAX, cellular, z-wave [32]. A number of standards have been proposed in the evolution process of smart grid architecture such as IEC 61970/61968-CIM, IEC 61850, IEC 62351, IEC 62325 and so on [33]. A comprehensive summary of the available communication technologies for smart grid is presented in the Table 1 [34], [35].

III. MULTI-AGENT SYSTEM (MAS) TECHNOLOGY: LITERATURE STATISTICS
The Multi-agent system (MAS) technology enables the transformation of the centralized grid into the decentralized smart grid. It provides flexible distributed control for the grid and thus fulfils some essential criteria for the smart grid such as automatic fault identification, isolation, and restoration (FLISR) [36]. Extensive research works have been found in the field of MAS and Smart Grid. Fig. 2 presents year wise publication in the field of MAS technology-based smart grid (1992-2018) using a comprehensive SCOPUS database tool.

IV. AGENT DEFINITION AND CHARACTERISTICS
The objective of the multi-agent system is to segment a centralized system into several entities, commonly known as a decentralized system in order to solve a complex problem Intelligent agents must have the characteristics of autonomy, reactivity, pro-activeness and sociability [37]. Usually, an intelligent agent consists of four components input interface, output interface, decision mechanism and communication stricture [38].  avoid possible conflict of interest, the role of agents should be elaborated and defined in a decentralized power system. In this section, a detailed literature survey has been performed to identify different structures of MAS discussed in the smart grid. The first architecture is the centralized concept where the agents are working under a central agent [39,40]. The subordinating agents are called a reactive agent (lowlevel intelligence and limited communication) and a master agent is termed as cognitive agent (high-level intelligence and higher communication). A cognitive agent has the responsibility of taking the central decision, communicating reactive agents and managing the network whereas reactive agent only accesses the local data and performs actions requiring a faster response (tripping a circuit autonomously). Such concept of a centralized architecture of MAS has also been mention in [41]. On the other hand, distributed architecture can be formed as a number of agents communicating through a common interface. Agents have the individual knowledge of their parts and have the capabilities of performing the local tasks instead of having global knowledge. The communication among the agent is performed through a common interface-agent directory. Agents can see the activities of the other agents via this agent directory which is described as a yellow page.
Example of such distributed architecture of MAS has been discussed in [42]. Fig. 3 presents a centralized MAS structure in smart grid where the agents (agent 1, agent 2. . . agent 4) work as reactive agents, they are supervised by a cognitive agent and all critical decisions are taken by this supervisory agent.
Another approach in MAS structure in microgrid/smart grid is the hierarchical approach [43]. The majority of MAS based implementations in smart grid/microgrid found in literature have followed two level or three-level hierarchal architecture [44], [45]. In [46], a three-level hierarchy has been illustrated where the upper-level agents are related to grid operators or distribution system operators who take part in decision making, solving global problems, coordinating with mid or lower level agents, processing of data  and handling ICT protocols. Mid-level agents perform grid operation such as islanding operation or grid connection operation, communicate low-level agents for a specific task such as FLSR service. The bottom level agents are equipped with sensors and controllers (e.g., Breakers, DERS, Batteries and loads). In Fig. 4, a 3-level hierarchical MAS architecture is illustrated where the low level is comprised of load agent, generation agents 4, 6 and 8 (wind turbine, PV fuel cell) and energy storage agent 5 (Battery) located in two separate zones of a distribution network. One zone comprises agents 4, 5 and 6, another zone comprises agents 7 and 8. These two zones are monitored by two zone monitoring agents residing in the mid-level. In the top level, a grid supervisory agent acts as DSO taking all the central decision.
In [47] conventional agents such as DER agents, Load agent have been modelled whereas some works have presented service based agent such as electricity market planning agent, grid planning agent [48], [49].

V. MAS & ICT BASED SMART GRID: METHODOLOGY, STANDARDS AND DEVELOPMENT PLATFORMS
This section summarizes the methodology, standards and development Platform of MAS based smart grid in the literature. Fig. 5 describes a generalized framework for MAS integrated smart grid system consisting of three stages: methodology, available standards, and development platform respectively. The generalized methodology (Requirement analyse-Design-Implementation-Deployment) to create a MAS based system has been discussed in [50]. Additionally, other methodologies have been discussed in literature for MAS design, such as GAIA [51], SODA [52] and others. With the development of ICT and smart grid, there was a need for data interoperability among different devices as well as integrating with existing systems. Hence, several standards and specifications have been proposed by IEEE working groups and IEC such as Foundation for Intelligent Physical Agents (FIPA) [53] and IEC 61850 [54].
FIPA facilitates all the agents to communicate among themselves and thus ensuring data interoperability. It also provides two directories namely ''agent management service (AMS)'' containing registered agents' list and ''directory facilitator (DF)'' containing services provided by the registered agents [3]. FIPA employs KQML (Knowledge Query and Manipulation Language) [55] and ACL (Agent Communication Language) for standard communication language [56]. Ontologies also have an important part in the messaging structure defining concepts and relations among the agents [57]. The components according to FIPA standards are presented in Fig. 6 where two agents are communicating among themselves using ACL.
Numerous number of software and toolkits have been developed in order to implement MAS platform [50]. They range from their compatibility, operating system, application fields and distinctive features as well as help support from the developers. A comparison between different MAS development platforms is presented in Table 2 in order to have a  view on the development platform employed in smart grid available in the literature.
Having a highly cross-functional architecture, the smart grid has become immensely complex and rapidly developing concept integrating electric power layer and communication technology layer simultaneously. This information and communication technology platform consists of a number of communication structure to incorporate the electric power generation, transmission, and distribution and consumers/prosumers. IEEE 2030-2011 is one of the most widely adopted standards that provides smart grid's communication architectural overview [58]. Smart grid communications architecture has been categorized into three subgroups by IEEE 2030 standard -Home area network, Neighborhood area network and Wide area network [59]. A graphical representation of these subgroups including layers of the smart grid is shown in Fig. 7. Typically, Wide Area Network (distance > 10 km) covers generation and transmission part (power stations, PMUs, protection and control units, several NANs) while Neighborhood Area Network (distance 100 m-10 km) covers distribution side (DERs, protection and control unit, multiple HANs) and Home Area Network (distance < 100 mm) covers the consumers side (Smart meters, EVs, sensors). All the communication technologies are facilitated by both wired and wireless media with a data speed ranging from Kbps to Gbps and following a number of standards proposed by IEEE and IEC. However, this paper only focuses on extensive review on the standard IEC 61850-substation automation Systems.

A. IEC 61850: A QUICK LOOK
IEC 61850 is an Ethernet-based communication standard proposed by the International Electrotechnical Commission (IEC) aiming for the substation automation and protection [60]. Replacing DNP3 (Distributed Network Protocol, version 3) [61], IEC 61850 intends to cover all aspects of substation automation technology by reducing the complexity and maintenance cost, specifying protocols such as TCP/IP, UDP/IP and ensuring data interoperability among multivendor IEDs ( Intelligent Electronic devices) [62]. The timing requirements for information and data exchange in power substations is thoroughly described by IEC 61850 varying from (ranges from 3 ms to 500 ms) [63]. Fig. 8 presents a detailed architecture of IEC 61850 along with communication strategy. The IEC 61850 communication architecture consists of three levels: station, bay and process level respectively. Measurement devices such as CT/PT, I/O devices, sensors, and actuators correspond to process level whereas Bay level have IEDs and the station level comprises Human Machine Interface (HMI) and station controllers. Several works have been found in the literature regarding the architecture of IEC 61850 [64], [65]. Multi-agent based smart grid automation architecture employing IEC 61850/61499 intelligent, logical nodes have been focused in [66]. Reliability and availability VOLUME 7, 2019  analysis of IEC 61850 communication architectures have been performed in [67], [68].
Being an object-oriented protocol IEC 61850 breaks down a physical device into logical devices, which can be further broken down into Logical Nodes, Data Objects and Data Attributes [26] depicted in Fig 9. IEC 61850 allows mapping of abstract data model and service definition into a number of protocols. In [69], the mapping of IEC is mainly divided into two segments, client-server based communication (SCADA) and publisher, subscriber-based communication respectively. In the [70]  In other words, the data and the service model of ACSI are mapped to MMS (Manufacturing message specification) [71]. GOOSE and GSSE protocols, generally known as Generic Substation Events (GSEs), are employed to exchange critical data such as control signal and status information for example warning event in real time. The GOOSE protocol presents a fast horizontal communication service among different IEDs including interlocking and blocking and thus facilitates publisher-subscriber mechanism [72]. Analog data such as current and voltage measurement are exchanged through Sample Measured Value (SMV) protocol. Both GOOSE and SMV are directly stacked on Ethernet. The system clock is usually broadcast by the TS service to synchronize with the IEDs for the accuracy of measurement. The GPS and the IEEE 1588 Precision Time Protocol (PTP) are the most widely used tool for this purpose [73], [74]. The summary of IEC 61850 based communication protocol, message time, time requirement and application is presented in Table 3 [75]- [77]:

VI. MAS APPLICATION IN SMART GRID
This section reviews the application of MAS in the smart grid that have been discussed in the literature. It has been divided into sections spanning several categories based on the particular application: smart control, operation and management; smart grid protection; monitor; smart grid transactive energy management; smart grid security.

A. SMART GRID CONTROL, OPERATION AND MANAGEMENT
A number of works have been found and that utilizes the MAS strategy to control the grid Smart grid Control, Operation, and Management. They have been summarized in Table 4.

B. SMART GRID PROTECTION
Several approaches that utilizes the MAS strategy for the protection of the smart grid including different communication platform have been recently proposed, this subsection summarizes some of them. In [93] the authors focus on the implementation of MAS based auto-configurable protection scheme using an agent-based architecture with the IEC 61850/IEC 61499. Another MAS based autonomous microgrid protection employing DERs, IEDs and IEC 61850 is described in [94]. A simulation-based study was performed in [95] for an IEEE 39 bus test system for MAS based protection scheme using FIPA compliant JADE platform. Self-Healing strategy with MAS based system and system restoration (Resilient Microgrid) is focused in [96]- [101]. The mapping between a multi-agent based architecture for power system control and the IEC 61850 standard for utility automation is highlighted in [102], which depicts how a relay agent in a particular substation is mapped into different levels according to IEC 61850 including different LN interaction. Additionally, this relay agent at a particular substation can communicate with another relay agent located at another substation using ACL. In the control mechanism, the relay agent continuously monitors the network and gets triggered upon observing a current value higher than the threshold value, and breaker operates and thus ensuring multi-agent based protections scheme.

C. SMART GRID MONITOR AND DIAGNOSIS
Several works have focused on the use of the MAS approach for monitoring and diagnosis of the smart grid. Smart automation with MAS is presented in [103] using IEC 61850 and OPC UA. A number of projects regarding smart grid monitoring employing MAS and ICT have been presented in [104]. For example, a multi-agent system (COMMAS) for condition monitoring of transformers is presented in [105]. UK utility based called Protection Engineering Diagnostic Agent, known as PEDA, is described in [106].  These agents continuously monitor the network, and if any abnormal situation (fault) is identified, it is recorded, interpreted and validated through these agents and collation agent collaborate among all these agents whereas the assistant agent aids with network diagnosis task. Apart from PEDA, several automated agent systems were developed on distributed PLC and Microcontroller board for intelligent monitoring employing IEC 61499/IEC 61450 in literature [107]. Javabased JADE framework has been widely adopted among the researchers for testing of MAS based decentralized intelligent monitoring system [108].

D. SMART GRID TRANSACTIVE ENERGY MANAGEMENT
Transactive energy management is a new edge concept of a smart grid for demand response (DR) management in the electricity market including grid reliability requirements. VOLUME 7, 2019 It employs coordination (ex: price bidding) between several cognitive entities (ex: consumers, prosumers) in a particular environment thus enabling transactive control mechanism (TCM) [109]. At present, multi-agent technologies are dominating in demand response domain because of their flexibility, decentralized nature, robustness, autonomy and open architecture. There are a number of works found in the domain of MAS based TCM [110], [111]. In other words, the multi-agent platform have all the capabilities to satisfy all the attributes of TCM (flexibility, bidding, empowerment, autonomy, and connectivity). According to smart grid definition, active buyers and sellers take part in the Electricity market (EM) to trade electricity through supply and demand criteria. Many entities such as distribution system operators (DSOs), transmission system operators (TSOs), retailers, aggregators, regulators, consumers, prosumers, etc. are defined in this market with the ability to negotiate, bid, coordinate, and compete. ''Game Theory'' based decision-making process [112], [113], MAS based market (EM) modeling focusing strategic competition [114], bidding strategies [115], and market tariff pricing through market share optimization [116] are some of the highlighted works in the field of electricity market modeling. However, none of them have mentioned the required ICT structure for this market mechanism. To comply with smart grid ICT requirement, Transactive Energy Management and control utilizing peer-to-peer (P2P) energy trading and the required ICT framework have been focused in [117]. A detailed survey on residential demand-side management architecture was presented in [118]. A thorough review of agent-based modeling and simulation of the smart electricity grid is highlighted in [119].
A MAS-based solution including service-oriented architecture (SOA) and IEC 61850 structure is briefly mentioned in [120]. A number of tools that are utilized for agent-based demand response strategy is mentioned in [121] including GridLAB-D [122], OpenDSS [123], AMES [124] and MAT-LAB/SIMULINK incorporating JADE [125]. A significant work [126] that have summarized ICT enabled MAS based Electricity market, and demand response works is the Pow-erMatcher [127] which is depicted in Fig. 11. The Power-Matcher [128] is a distributed market control mechanism utilizing supply and demand matching (SDM) in the electricity market.
The implementation of Power Matcher [129] is an ideal example of MAS and ICT based smart grid transactive energy management where many intelligent agents exist (Device agent, Concentrator agent, Auctioneer agent, and unique Objective agent). All the physical devices are classified as device agent (DERS, household appliances), and they send bids and receive prices whereas the concentrator agent aggregates all bids from the sub-cluster of device agents and publishes a single bid towards the auctioneer agent. On top of the hierarchy, the bids from all agents are aggregated by the auctioneer agent and combined into a single bid. Finally, the objective agent played a vital role by providing purposes to the clusters and interfacing the agent clusters to the business logic based on specific market approach.

E. SMART GRID SECURITY
With the rapid development of the smart grid, it also introduces new security concerns. As electrical infrastructure is now integrating with ICT structure, the electrical grid has been transformed into a Cyber-Physical network where the physical layer consists of power generation, transmission and distribution units and the cyber layer comprises of communication protocols and data exchange for controlling the physical layer. Hence, the grid has always the possibility of security threats from hackers. There was a number of attacks on the EU electrical grid where hackers intended to disrupt the services [130].
The security threats for smart grids are categorized into two segments mainly technical (device failure, cyber threat) and non-technical threat (regulatory policy) respectively in [131]. A number of security threats are presented in Fig. 12 classifying into different categories such as global, local, data, control and device focusing substation automation system architecture. Similarly, in [132], all possible cyberattacks and challenges in the smart grid are summarized in brief.
A number of approaches have been mentioned in [133], [134] to ensure smart grid security. In order to protect the smart grid against cyber-attacks, a novel security agents based framework is described in [135]. Another multi-agent based approach focusing on SCADA network threat is presented in [136] including real-time testbed implementation. In [137], trust-based multi-agent scheme was employed to prevent false data injection into the smart grid. To analyze cyber security prospect in the smart grid arena, several cyber-physical system (CPS) testbed have been implemented. These test beds focus on three prospects of the smart grid security level of vulnerability, effect analysis; and attack-defense mechanism evaluation and validation respectively. Some cyber-physical test bed focusing on cyber security goals and that are found in the literature are summarized in the Table 5.

VII. MULTI-AGENT FRAMEWORK: CURRENT FORM AND NECESSITY
A number of approaches have been investigated in the literature to develop a multi-agent framework in the context of the smart grid. There are numerous reasons behind the necessity for developing such a framework. The scope is to create a suitable technological environment to implement a decentralized, automated, intelligent and resilient power grid according to the different communication and ICT standards proposed by IEEE and IEC. The technical aspects and implementation challenges of MAS have been addressed by researchers and industries together and have been continuing to be subject of research interest for years. Among them, a wide range of technical difficulties have been explored such as data interoperability, high speed data rate, linking power domain with communication domain, performance of different platforms, etc. Additionally, a hierarchy of controls awaits still to be designed for the successful implementation of the MAS integrated ICT enabled smart grid. The details of these potential designs have been discussed in detail in previous sections (i.e. MAS design approach, MAS platform, MAS standards, IEC 61850). A typical form of multi-agent enabled smart grid frame work studied in current literature [13], [145] is highlighted in Fig. 13.
Despite the identified advantages of this MAS enabled power grid framework, only few works report on how to perform real time simulations that incorporate ICT standards in the smart grid. The reason is that, it requires complex system design, fast computation algorithms as well as advanced network and telecommunication based technological approaches and protocols (i.e. Optical fiber, IEC 61850). In this regards, the aim of this paper is to propose a laboratory framework VOLUME 7, 2019 in ICT enabled smart grid employing multi-agent systems embedded in the IEC 61850 standard. Future work will focus on the hierarchical control system as well as smart strategies for the protection of grid. These works should also take into consideration some of the cyber physical aspects of the grid such as false tripping and bad data detection for available communication networks operating under some software architecture.

VIII. LABORATORY FRAMEWORK
In this section, the authors attempt to describe the preliminary work carried out for the development platform of a testbed for MAS application in ICT enabled smart grid at Curtin University IEC 61850 laboratory. This laboratory features smart grid testing facility with 23 IEDs, from commercial vendors like ABB, Siemens, Alstom, Schneider, and others. The laboratory's architecture, testing facility including the hardware configuration and software implementation, and some of its components will be highlighted in this section. The architecture of this test bed is classified into three stages and can be called a Cyber-Physical System (CPS) altogether altogether and illustrated in Fig. 14. This platform will be used to implement the hierarchical control of smart grid employing IEC 61850.
The first stage comprises physical power systems components based on Real Time Hardware in Loop (RT HIL) via OPAL-RT's RT-LAB simulator. This stage also consists of configuration of industry provided Intelligent Electronic Devices (IEDs) which is vital for Substation Automation System (SAS). The second stage enables the implementation of MAS control in JADE platform. The third stage resolves the interfacing challenge between different components utilizing industry recommended protocols. Additionally, it facilitates the simulation of the network topology, analysis of relevant service quality, and analysis of cyber security and threats through OPNET Modeler communication simulator.

A. POWER SYSTEM NETWORK
The platform features the simulation provision of typical power system including MATLAB/SIMULINK simulation platforms. In addition to that, for Real Time Hardware in Loop (RT HIL) testing, Real-time simulator (RTS) OPAL-RT OP4500 is utilized. Several protection features of IEDs from different commercial vendors are investigated in the real time employing RTLAB Opal-RT. Curtin Green Electric Energy Park (GEEP) facilities can be remotely accessed from IEC 61850 lab to implement hierarchical control. The power system simulation platform is connected to the Intelligent Electronic Devices (IEDs), through standard communication protocols IEC 61850, GOOSE and SV, and standard I/Os respectively.

B. MAS PLATFORM
In this testbed, MAS based platform is implemented in JADE software. This software platform features JADE utility agents and services such as Directory facilitator (DF) agent, Agent Management Services (AMS) agent, and Message Transport Service (MTS). The software platform is set up in compliance with FIPA (foundation for physical intelligent agents) standards to facilitate multi-agent features.

C. COMMUNICATION PLATFORM
Communication platform aims to connect all the components in the system as per different industry standards. While the grid is simulated in real-time employing the Sim Power Systems block set and the ARTEMiS blockset in OPAL-RT respectively, the Agent blocks and the simulated grid interact through TCP/IP network over OPC-RTLAB server. Additionally, OPC/UA server was utilized to connect physical IEDs and MAS platform. Ethernet-based IEC61850 communication protocol facilities real time communication between different IEDS and OPAL-RT OP4500.

IX. CONCLUSION
Due to the highly complicated nature and new energy technology, smart grid concept is evolving very rapidly. The wide spread deployment of smart grid requires developers to handle a large amount of data and information among many parties ensuring reliable, resilient and secured electric grid operation. In this regard, MAS application lends itself as one of the leading technologies because of its decentralized nature and operational robustness. On the other hand, information and communication technology (ICT) transforms conventional ''power grid'' into a ''smarter'' one and enables device integration, information exchange, data interoperability, and two-way communication. In this context, the standard communication protocol is an essential criterion for interfacing MAS based platform and the smart electric power grid. This paper attempts to provide an up to date technology status update on MAS applications with the required communication platform in the field of the smart grid. Initially focusing on different aspects of MAS, MAS definition, and MAS development platform moving onto standard communication protocols including ICT infrastructure for the smart grid are described. Along with MAS and ICT integration, a particular attention has been provided on the substation automation system (SAS) employing IEC 61850 including a laboratory conceptual framework. Finally, results of a detailed literature review has been presented highlighting MAS and smart grid ICT application in various areas. The authors hope that future research will benefit from this rigorous, comprehensive and up to date review both in terms of SWOT analysis (strengths, weaknesses, opportunities, and threats) of previously published papers and ready referencing in this rapidly developing field.