![Relations between hybrid multi-agent control model (HAM) and hierarchical control framework (HCF). [22]](https://www.researchgate.net/profile/Vitor-Coelho-3/publication/311505600/figure/fig2/AS:667868005949466@1536243549452/Relations-between-hybrid-multi-agent-control-model-HAM-and-hierarchical-control_Q320.jpg)
![FIPA compliant agent platform model. [57]](https://www.researchgate.net/profile/Vitor-Coelho-3/publication/311505600/figure/fig3/AS:667868005933063@1536243549830/FIPA-compliant-agent-platform-model-57_Q320.jpg)
Content uploaded by Vitor Nazário Coelho
Author content
All content in this area was uploaded by Vitor Nazário Coelho on Oct 17, 2017
Content may be subject to copyright.
Multi-agent systems applied for energy systems
integration: State-of-the-art applications and trends in
microgrids
Vitor N. Coelhoa,b,c,∗
, Miri Weiss Cohend, Frederico G. Guimar˜aese,∗
, Igor
M. Coelhof,a, Nian Liuh
aGrupo da Causa Humana, Ouro Preto, Brazil
bInstitute of Computer Science, Universidade Federal Fluminense, Niter´oi, Brazil
cGraduate Program in Electrical Engineering, Universidade Federal de Minas Gerais, Belo
Horizonte, Brazil
dDepartment of Software Engineering, Braude College of Engineering, Karmiel, Israel
eDepartment of Electrical Engineering, Universidade Federal de Minas Gerais
fDepartment of Computer Science, Universidade do Estado do Rio de Janeiro, Rio de
Janeiro, Brazil
gDepartment of Control and Automation Engineering, Universidade Federal de Ouro Preto,
Ouro Preto, Brazil
hSchool of Electrical and Electronic Engineering, North China Electric Power University,
102206 Beijing, China
Abstract
Mini/microgrids are a potential solution being studied for future systems relying
on distributed generation. Given the distributed topology of the emerging smart
grid systems, different solutions have been proposed for integrating the new
components ensuring communication between existing ones. The multi-agent
systems paradigm has been advocated as a useful and promising tool for a wide
range of applications. In this paper, the major issues and challenges in multi-
agent system and smart microgrids are discussed. We present a review of state-
of-the-art applications and trends. By discussing the possibilities considering
what has been done, future applications, with attention to renewable energy
resources integration in emerging scenarios, are placed on the agenda. It is
suggested that further studies keep growing in this direction, which will be able
to decentralize the high complex energy system, allowing users to participate in
the system more actively. This step may decentralize the infrastructure, giving
more weight to society wishes, as well as facilitating maintenance, reducing costs
and opening a the door for innovative ideas for low-cost based equipment. On
the other hand, letting several combinatorial optimization problems opened to
be improved and discussed along the next coming years.
Keywords: Multi-Agent Systems, Smart Grid, Microgrid, Smart-Microgrids,
∗Corresponding authors.
Email addresses: vncoelho@gmail.com (Vitor N. Coelho), fredericoguimaraes@ufmg.br
(Frederico G. Guimar˜aes)
Preprint submitted to Applied Energy – Review Paper December 28, 2016
Integrated Energy Systems
Contents
1 Introduction 2
2 MAS and Smart Grids overview 6
3 MAS and Smart-Microgrids 9
3.1 Microgrid operation and management ............... 10
3.1.1 Energy storage microgrid systems ............. 11
3.2 MG security, stability and protection ............... 12
3.3 Multiple microgrids ......................... 14
3.3.1 Microgrid market operations ................ 14
3.4 Stand-alone smart-microgrids .................... 15
3.5 Microgrid demand control ...................... 15
3.6 Service restoration .......................... 16
4 Future MAS in the context of Smart-Microgrid 17
4.1 Applications for allowing mini/microgrid autonomy ....... 17
4.2 System integration in emerging scenarios .............. 19
5 Conclusions 20
1. Introduction
Smart Grid (SG) is considered as the future of power grid able to manage1
production, transmission and electricity distribution. The task has been mainly2
done by using Information and Communication Technologies (ICT), Distributed3
Generation (DG) and Artificial Intelligence (AI). Due to the need of consistently4
adapting and integrating new tools to the current grid, SG has become a major5
challenge for developed and developing nations in both research and utilization6
aspects [1]. Investing in SG infrastructure is a key facilitator for public goods,7
such as decarbonisation and energy security [2]. SG are expected to play an8
important role in the resolution of many issues of current power grid systems9
[3,4]. As emphasized by Zhao et al. [5], power grids will become a mesh of10
networked Microgrids (MG) collaborating to deliver electricity to consumers11
and eventually, assisting stand-alone systems.12
Research and development has been demonstrating the technical and eco-13
nomical feasibility of greener generation technologies based on wind, solar, hy-14
drogen and hydro power. Integrating these technologies has become a priority15
in MG [6,7], not only because of introducing these Renewable Energy Resources16
(RER) but also because extra elements have been required, as pointed out by17
Farhangi [8]: sensor and metering network; network nodes with computation18
capabilities; switches or actuators which allow the grid setup to be changed and19
2
the capability of plug in or plug out new devices. Future MG may equip cus-20
tomers with distributed energy generation and storage systems that can change21
their overall demand behavior, promoting the development of several smart-22
microgrids. These tools will provide users the ability of taking profit of their23
generated energy as an important economic factor [9], helping them to turn24
into stand-alone systems and self-sustainable users. Providing autonomous as-25
sistance in order to aid complex decision making tasks will be required by an26
increasing number of MG users. Kyriakarakos et al. [10] defined concepts and27
discussed potential impacts of polygeneration microgrids, taking into account28
RER and sustainable Energy Storage Systems (ESS). The consideration of MAS29
was also suggested [11, 12], in order to design an optimal polygeneration mi-30
crogrid for a given amount of an investment and an autonomous collaborative31
system, respectively.32
Coordination and control of these new emerging grid components remain33
a great challenge [13]. Advanced networking, as well as ICT, have been mo-34
tivating the integration of the conventional power grid in smarter ways [14],35
inspiring the use of distributed Multi-Agent Systems (MAS). Autonomous con-36
trol of SG systems may allow placing additional DGs without reengineering the37
whole system, and using it in the peer-to-peer model eliminates the requirement38
of a complex central controller and associated telecommunication facilities [15].39
Logenthiran, Srinivasan & Wong underscored that MAS is one of the fastest40
growing domains in agent oriented technology which deals with autonomous de-41
cision modeling. Moreover, it has been showing to be crucial in SG operations42
[16]. MAS has spread to diverse SG applications in the field of power systems43
restoration, security and protection, control, monitoring, energy storage and44
maintenance scheduling, and electric power market simulation [17]. The need45
to integrate both fields of knowledge, MAS and SG, has increased extensively46
around the world in the recent years. Figure 1shows the number of publications47
relating MAS and SG in the Scopus database, performed on September 30th,48
2015. The red point, for 2016, indicates expected values after evaluation of all49
2016 submitted manuscripts.50
In particular, the MAS paradigm can be adapted to model, control, manage51
or test the operation of MG [18]. The latter had become a basic and fundamen-52
tal infrastructure in the SG environment and has been receiving attention in53
recent literature works (trends related to MG control were presented in Olivares54
et al. [6]), being envisioned as a possible future energy system archetype [19].55
As noticed by Jiayi, Chuanwen & Rong [20], MAS technology can be deployed56
over it in order to give support to the resolution of different operational prob-57
lems, such as: connecting small Distributed Energy Resources (DER) units,58
coordinating several local decisions; providing tools for MG in order to operate59
in a liberalized market of energy trade; promoting stability and high quality60
energy for its local environment [21]. With MAS, each control unit in a MG61
(e.g., DG, Distribution Storage (DS), or load controller) is designed as an agent62
[22], and its operation is determined by devices interactions through intelligent63
decision making processes and collaborations.64
MG systems aggregate many DER and loads together as an autonomous en-65
3
●
●
●
●
●
●
●
●
●
●
●
●
●
2004 2006 2008 2010 2012 2014 2016
0 50 100 150 200
Year
Total publications − MAS and SG − Scopus
Figure 1: Number of papers involving MAS and SG until September 13th, 2016
tity [23]. Furthermore, additional components added or improved by consumers66
will be integrated to the SG, imposing new frontiers for MG control and man-67
agement. For example, Plug-in Electric Vehicles (PEV) [24] are being integrated68
to the power grid (specially with the rise of smart charging parks, known as,69
SmartPark [25]), imposing new grid constraints, requirements and goals, settled70
by its users. Coordination and integration of DER in MG systems have been71
the focus of different researches and remain complex tasks [26]. This integration72
may cause lack of efficient control and problems in stability, reliability, power73
quality and security over MG. Thus, as emphasized by Agrawal & Arvind [27],74
MAS seems to have appealing features meeting operation, control requirements75
and goals balance of the entities that integrate the MG systems. Figure 2shows76
the number of publications linking MAS and MG.77
This field has not been researched only by academia, patents have been78
submitted relating MAS to MG applications [28]. Goldsmith [29] registered a79
patent, deposited in the end of 2011, for a MG encompassing a geographic range80
of less than 300 square miles, comprising less than five thousand different power81
sources and less than one hundred thousand different loads. They described a82
computing architecture that facilitates autonomously controlling operations of83
this MG. Computing devices in this decentralized, distributed network perform84
computations using MAS based technologies.85
Many research work found in the literature adopt an agent based solution86
to implementf intelligence, communication and optimization over the emerging87
smart electric grids. However, Sanz, Rodrigues, Soler & Gallejo [30] pointed88
4
●
●
●
●
●
●
●
●
●
●
●
●
●
2004 2006 2008 2010 2012 2014 2016
0 20 40 60 80 100
Year
Total publications − MAS and MG − Scopus
Figure 2: Number of papers involving MAS and MG until September 13th, 2016
out some misunderstandings in the way agent based systems are being applied89
to MG systems, such as: consideration of one instance of each agent type;90
fagents without really decision capabilities; components distribution but only91
decentralization; agent interaction acting as client-server rather than peer-to-92
peer. On the other hand, new research fields involving MAS are evolving from93
SG applications, e.g. Figure 3shows a hybrid multi-agent control model, so-94
called HAM, proposed by Dou et al. [22]. The effectiveness of the hierarchical95
hybrid control in distribution grids with DER units is demonstrated through96
simulation.97
Integrating these emerging hybrid tools in order to assist SG viability is the98
most important task tackled by the researchers. Low-cost based systems, such as99
implemented by Purusothaman [31], are important applications that promotes100
and facilitates microgrids diffusion. Our current work will discuss what has101
been achieved with a main focus on the state-of-the-art in MAS applications102
for supporting MG and Multiple MG systems. We will present trends and103
prospectives envisioned by us and also extracted from the literature.104
Section 2describes applications related to MAS and coordination, control105
and management of SG components. Section 3presents an overview of MAS106
and smart-microgrid systems. Decentralized approaches for MG operation and107
management are presented in Section 3.1 with a special attention into energy108
storage systems, discussed in Section 3.1.1, Section 3.2 focuses on MAS and109
MG in the context of promoting SG security and grid stability. Section 3.3110
discusses the whole of multiple microgrids and MAS. The special case involving111
5
Figure 3: Relations between hybrid multi-agent control model (HAM) and hierarchical control
framework (HCF). [22]
market operation in a system of interconnected MG is highlighted in Section112
3.3.1. The state-of-the-art regarding to stand-alone MG managed by MAS is113
explored in Section 3.4. Demand control system are presented in Section 3.5,114
Section 3.6 reviews some applications related to Power Distribution System115
Reconfiguration (PDSR), focusing on service reconfiguration problems. Section116
4introduces some future applications expected between MAS and the field of117
Smart-Microgrids. Finally, some conclusions are drawn in Section 5.118
2. MAS and Smart Grids overview119
The term Smart Grid has been more oriented to the entire electrical system120
including generation, transmission and distribution [32]. Regarding the distri-121
bution system, several efforts target the increase of manageability and efficiency122
by dividing the smart distribution grid into sub-systems. Figure 4presents a123
future vision of a SG, adapted from European Commission report on SG [33].124
Different sub-systems will compose the future SG, as can be imagined through125
Figure 4. These sub-systems are called “Smart-Microgrids”, or just “Micro-126
grids”, and consist of energy consumers and producers at a small scale, which127
are able to manage themselves, being self-sustainable or in stand-alone state128
(readers may check Kaldellis [34]). The environment depicted involve different129
components idealized for the future power grids, such as: Hydro power stations130
(medium and small); Low emission power plants; Solar power plant; Biomass;131
Wave energy generation (a brief view of these last five RER can be seen in132
Ellabban, Abu-Rub & Blaabjerg [35]); Offshore wind farms [36]; Residential133
photovoltaic generation; Batteries bank (Battery Energy Storage System [37],134
Compressed Air Energy Storage systems [38], Flywheels [39], Thermal Energy135
Storage [40], Pumped-storage hydroelectricity [41], Superconducting Magnetic136
Energy Storage [42]); PEVs [43]; Distribution and management: Transformers,137
HVDC link, underground systems and power transmission, control and commu-138
6
Figure 4: Adaptation from the Future Network Vision of European Commission report on SG
[33]
nication center and satellites. Small wind turbine on buildings rooftops [44] and139
Smart Parks [45] could be also envisioned for this future system.140
The choice of RER by the future power grid is being expected [46], and this141
growth is motivated as a result of the need of reducing environmental impacts, as142
emissions of greenhouse gases [47,48]. The potential for RER is growing vastly143
and it is predicted that it will, in principle, exponentially exceed the world’s144
energy demand [35]. SG infrastructure should also provide new opportunities145
for the grid and its customers for information exchange regarding real-time146
electricity rates and demand profiles [49]. The massive insertion of these RER147
motivates the development of management systems to be able to integrate these148
DER to the SG.149
Studies in the field of DER management, in most cases, demand the inclusion150
of criteria such as fault tolerance and adaptability. Lagorse, Paire & Miraoui [50]151
reported that the designer of the components of these systems generally knows152
each agent response separately. Centralized management system focuses its153
attention solely on the overall reaction of the system. MAS provides new levels154
of problem-solving by employing interacting agents equipped with AI tools.155
In this sense, the use of a paradigm based on MAS has been showing to be156
reasonable and bridges the communication gap between humans and machines.157
As mentioned in the review of Khare & Kumar [51].158
“Presently multi agent structures are the advancement of artificial159
intelligence”.160
Brown [52] emphasized bidirectional communication between devices as the161
most important characteristic for integrating new DER into the energy systems.162
From this communication process and standards (for example IEC61850, as163
can be seen in Figure 9, or ZigBee based protocols [53]), a process of decision164
7
making is taken by different SG components. In this sense, the convenience165
for new MAS applications using agent peer-to-peer interaction instead of client-166
server will confront many opportunities in the near future. The shift to this167
new business model and the implementation of the SG has a starting point168
in installation of Smart Meters (SM) [54], which improve access to electricity169
consumption information, and sensors in residences or commercial buildings.170
SM are a key enabler for communication between SG devices.171
The IEEE Power and Energy Society addressed the existence of agent re-172
search in the rising SG through two reports. McArthur et al. [17] advocated the173
interest in investing in agent technology for Power Grids, concluding that a MAS174
could be used either as a way of building robust and flexible hardware/software175
systems or as a modeling approach. A second report, by McArthur et al. [55]176
emphasized techniques and tools that could favor and support engineers to use177
MAS. The authors indicate the Java Agent Development Framework (JADE)178
[56], which is a FIPA (The Foundation for Intelligent Physical Agents) standard-179
based MAS framework supporting multi-agent development with facilities of180
agent management, as an important agent platform implementation (which has181
being used in different SG applications [57,58,59,60,61]). However, some lim-182
itations for widespread adoption of MAS technology over SG systems were also183
highlighted by McArthur et al. [17], such as: Portability; Real implementation184
and security; Scalability in large scale real problems. An example of the agent185
management system (AMS) with FIPA message transport is shown in Figure 5.186
MAS guidelines implementations over MG were discussed and claimed by Sanz,187
Rodrigues, Soler & Gallejo [30].188
Figure 5: FIPA compliant agent platform model. [57]
A common consensus is that intelligence in conjunction with MAS can189
be implemented by the incorporation of known AI techniques such as: Evo-190
lutionary Computing, Population based and trajectory search Metaheuristics,191
Multi/Many-Objective Optimization techniques, learning and forecasting model,192
among others. The use of AI techniques in modern systems has been investi-193
gating and tested in various real life applications [62]. Key concepts regarding194
agents intelligence were pointed out by Wooldridge [63]. MAS are used as dis-195
tributed AI tools that, differently from classical AI, underpins its research on the196
possibility of learning from social phenomena [64]. It is often cited as the evolu-197
8
tion of distributed control [65], where two or more physical or virtual (software)198
entities interact with each other, capable of tackling sub-problems in order to199
reduce complexity of the main problem. MAS have been applied to regulate,200
coordinate and control SG, as will be presented throughout this current paper.201
The involved entities are namely agents, able to interact between external envi-202
ronments, to cooperate with other agents and have the computational ability of203
solving problems autonomously. Some common agents properties can be found204
in several MAS applications in the field of SG [22,65]. The architecture of a cog-205
nitive agent applied for MG management is reproduced as a flowchart in Figure206
6, designed by Velik & Nicolay [66]. Even if their application was not defined as207
a MAS, many characteristics applicable in distributed models can be extracted208
from their work. Some properties of MAS and its agents are highlighted below:209
•Flexibility – the MAS structure allows advanced plug-and-play capabilities210
and adaptively adjusts the control of MG according to actual conditions211
and targets. Agents are able to present self-adaptive behavior in accor-212
dance to the environment and act accordingly in order to accomplish their213
personal goals (mono or multi objective functions).214
•Fault-tolerance – if one agent fails, the whole system remains communicat-215
ing and able to adapt its new states admitting previous established rules216
and behaviors. Thus, control of individual DGs is robust to disturbances217
and faults in the context of MG.218
•Autonomy – the ability to operate in order to attend specific and individ-219
ual objective functions and also being guided by a global communitarian220
goal, without constant guidance from the user side.221
•Responsiveness: collecting environment information, data base (Figure 6)222
or real-time data acquisition, and completing a decision making process223
provide agents the ability to respond to changes.224
•Pro-activeness – the ability to reason and initiate its own actions in order225
to meet its specific goals, sometimes guided by its own beliefs (i.e., by226
processing information from deterministic or probabilistic forecasts [67,227
68,69]).228
•Social ability – the ability to bargain, collaborate, compete and exchange229
knowledge with other virtual or physical agents.230
•Scalability – extend and expand the functions of a MAS based on SG231
users’ needs is feasible and suitable.232
3. MAS and Smart-Microgrids233
Making electricity grids smarter is a challenging, long-term, and ambitious234
process. There is a broad unanimity of the necessity for smarter microgrids;235
9
Figure 6: Agent architecture. [70]
yet, stakeholders also associated a range of risks and barriers such as: lack of236
investment; disengaged consumers; complexity and data privacy with measures237
to make the grid smarter. MG growth is being motivated inside the cities,238
since stakeholders felt many smart energy system functions are more likely to239
be implemented in urban areas [71]. On the other hand, rendering the situation240
even more complex, more people are expected to still keep moving to urban241
areas, especially in developing countries [72].242
The task of making these decentralized MG smarter involve several chal-243
lenges and sub-problems, some of them are being tackled by the use of MAS244
and will be discussed from now on.245
3.1. Microgrid operation and management246
Effective energy management is a key to achieve efficient operation of the247
SG, extracting the best benefits of its components. Sanz, Rodrigues, Soler &248
Gallejo [30] suggested guidelines to be followed when designing an agent system249
to manage MG.250
A pioneer research done by Dimeas & Hatziargyriou [73] proposed optimiza-251
tion tools for internal operation of a MG interacting with the energy market.252
Their approach took use of three types of distributed controllers, as exemplified253
in Figure 7, and modeled four kinds of agents: production agent, consumption254
agent, power system agent and a coordinating agent. Every DER was designed255
for optimizing its own objectives while taking into account the overall benefit256
through an auction algorithm, whereas production units could accept or decline257
a load offer.258
Coordination agents are designed with smart negotiation algorithm that se-259
lects seller/buyer pairs in connection with generator and load agents, see Fun-260
10
Figure 7: MG control based on three decentralized controllers. [73].
abashi, Tanabe, Nagata & Yokoyama [74]. Ueda & Nagata [75] proposed a261
central SG controller with four types of agents. The controller was also con-262
sidered as a fifth agent, which announces two prices into the system: first, the263
price for peak buying and the price for off-peak buying. and second, batteries264
(prosumer), which might sell enegy to the load agents (buyers/clients) or buy it265
from the grid. Two types of energy generation agents were introduced, a green266
power generation and a gas turbine. The proposal was able to converge and267
operate the energy supply of the system.268
A distributed agent based solution to energy management was presented269
for hybrid energy generation system in Jun, Junfeng, Jie & Ngan [76], similar270
approaches were proposed by [77,78]. A MG composed of a train station, wind271
power plant and district was investigated in Kuznetsova, Li, Ruiz & Zio [79].272
An optimization tool was applied to solve goal-directed actions planning of each273
agent, based on robust optimization concepts. Their framework indicated the274
ability to improve system reliability and decrease power imbalances. MAS was275
also proposed for energy source scheduling for integrating MG systems with276
DER and lumped loads [70]. A management system based on multiple agents277
was designed by Ricalde et al. [80] in order to measure and control the loads278
inside the building or in an islanded power system.279
Other approaches focused on the MG management issue ensuring energy280
supply with high security and quality control, as found in Dou & Liu [22].281
These approaches are detailed in Section 3.2.282
3.1.1. Energy storage microgrid systems283
Energy storage has been widely analyzed for MG systems, a spread range284
of applications exist for Energy Storage Systems (ESS). Tan, Li and Wang [81]285
11
refer to the following: power quality enhancement; assisting microgrid in iso-286
lated operation; active distribution systems and PEVs’ technologies. Its use has287
important benefits, improving dynamic stability, transient stability, voltage sup-288
port and frequency regulation [37,82]. Furthermore, they can also be applied for289
minimizing global cost and environment impacts [83]. Current smart-microgrid290
scenarios may include different renewable energy resources and several types of291
storage units.292
A wide range of applications exist for ESS and the current generation of MG293
are now able to take profit of MAS over it. Battery-less system are being tested294
[84] and idealizes what has been called as green alternatives. Power dispatching295
problems [39] including ESS deals with communications of several different SG296
components, such as energy storage devices, DER and forecasting agents. PEVs297
are one of the most viable technologies for achieving the goals of energy saving298
and environmental protection before the breakthrough in battery technology299
and fuel cell technology [85], its penetration is expected to increase significantly300
in the next 20 years [24]. Bidirectional power flow between PEVs and the301
grid will become essential [86] and coordinating this new wave of plug-and-play302
vehicles is a claiming burden.303
Hu, Saleem, You, Nordstrom, Lind & Ostergaard [87] applied multi-agent304
technology, coded with a co-simulation environment called JACK, for distribu-305
tion grid congestion management considering the integration of electric vehicles.306
They developed a two level hierarchical control method for integrating PEVs307
into the grid. PEVs owners and a distribution system operator were the main308
agents of the test system, communication and agreements were facilitated by309
the introduction of two operators: fleet and the grid capacity market operators.310
Ramachandran, Srivastava & Cartes [88] described a decentralized controller311
using MAS in an electricity market framework. Their goal was to decide optimal312
charging based hourly charging rate of each EV battery. A model with customer313
comfort zone in order to define demand response was considered.314
Switching the operation modes of the storage units based on the MAS by us-315
ing fuzzy logic rules, ensuring a secure and reliable energy supply, was proposed316
by Lagorse, Simoes & Miraoui [89]. The developed control scheme considered317
batteries state-of-charge limits and the size of charging/discharging currents.318
Motivated by this previous work, Yoo, Chung, Lee & Hong [90] improved it319
using a state machine able to respond to the changes in MG environments for320
controlling the output power of the DERs. Figure 8shows a general view about321
the framework proposed by them.322
Energy dispatch and coordination without involving a central controller was323
architectured by Ye, Zhang & Sutanto [91]. The coalition formation-based en-324
ergy dispatch mechanism, communicated through a defined negotiation pro-325
tocol, enables SG components to make decisions autonomously and archiving326
effectiveness compared to other recently developed mechanisms.327
3.2. MG security, stability and protection328
Agrawal & Arvind [27] defined that MG controls should ensure connectivity329
to the main grid or self-isolation. MAS application had already been used to330
12
Figure 8: Configuration of multi-agent based MG energy management with EES [90]
detect faults on utility feeders thereby switching microgrid to island mode [92].331
The transition between these two states should happen, in a rapid and seamless332
fashion. The role of switching a MG between island mode to grid-connected333
mode (vice-versa) is a challenge matter and its feasibility was already evaluated334
through applications related to hierarchical MAS [93,94]. When working in grid335
connected, the microgrid is aid by the grid supplier, excess of deficit of power is336
absorbed by it. On the other hand, maintaining system frequency and stability337
is requested in applications running in islanded mode [57].338
The problem of achieving a stable frequency spectrum in MG under islanded339
operation won particular attention of MAS applications recently [95,96], instead340
of classical centralized control strategies [97]. Maintaining a specific frequency in341
the islanded mode is an important requirement, the control of DGs’ output and342
charge action of DSs are used in supply surplus conditions and load-shedding343
and discharge action of DSs are used in supply shortage conditions. Kim, Ki-344
noshita, Lim & Kim [98] proposed a MAS for load-shedding, which is intentional345
reduction of electricity use, is a critical problem in islanded MG operation based346
on the MAS.347
A multi-agent based protection framework was proposed to enhance the348
stability of smart grids in Rahman, Mahmud, Pota & Hossain [99]. In Rosa,349
Silva & Miranda [100], a MAS technology-based platform was considered as350
potential applications in management and simulation processes for power grids.351
Physical grid parameters and network constraints that can be abstracted to MG352
were considered by the last two mentioned works. In Wilkosz [101], MAS, in353
connection with Petri nets, was used to verify the power-system topology during354
real time to avoid errors in the power network connectivity models.355
13
Mocci, Natale, Pilo & Ruggeri [102] included coordination of active demand356
along with the use of PEVs. They designed agents able to perform optimization357
based on local and global shared information, seeking goals based on Nash’s358
theory on games. A central load coordinator aggregates the demand of the359
main grid and its users, a master agent assisted the operation of independent360
agents checking system operation without degrading power quality and stability.361
Farid [103] focused on resilience, or self-healing, coordination and control of362
future MG designed through MAS. He presented generic MAS design princi-363
ples for resilience in power systems considering, as background, graph theory364
and axiomatic design for large flexible engineering systems [104]. The resilience365
property on normal operation of healthy regions of the grid while disrupted366
and perturbed regions bring themselves back to normal operation. Further-367
more, self-healing system has the ability to operate connected or disconnected368
from the main power grid. Unexpected critical loads can also disrupt the sys-369
tem, thus, some studies dealt with this event in photovoltaic systems [105].370
Autonomous agents, their cooperation, coalition, continuous evolving of faster371
response time, and adaptability are key enabling technologies for such a re-372
silient control scheme. It was concluded that limited contributions to power373
grid resilience has been done until 2015. Multi-agent systems have recently374
been proposed as a key enabling technology for such a resilient control scheme.375
3.3. Multiple microgrids376
High penetration of power at distribution level creates such MMG, specially377
in large geographical regions. The European Research Project More MicroGrids378
[106], characterized the concept of MMG related to a higher level structure,379
mostly composed of medium-voltage components, consisting of several MG and380
DG units connected on several adjacent feeders along with controllable loads.381
In Yuen, Oudalov & Timbus [107], MAS was used to coordinate microgrids to382
participate in ancillary service markets, such as frequency-control-reserve mar-383
kets. Their goal was to analyze aggregated MG in view of technical impacts or384
benefits on the network in market scenarios. Providing Frequency Control Re-385
serves (FCR) can be a profitable and important technical aspect when multiple386
MG are connected and trading.387
The special case involving market operation in multiple interconnected mi-388
crogrid is detailed in the next section.389
3.3.1. Microgrid market operations390
Bertsekas & Casta˜non [108] highlighted that MAS-based architecture en-391
ables and can assist individual microgrids to practice power trading, a type of392
asymmetric assignment problem. This task is mainly done by efficient AI tools.393
Considering multiple interconnected microgrids, Nunna & Doolla [59] considered394
symmetrical assignment problem based on na¨ıve auction algorithm in order to395
match the buyers and sellers in the energy market. Results strongly indicated396
that the agent-based management techniques, such as MAS implemented using397
JADE platform, is effective in resource management among multiple microgrids398
economically and profitably.399
14
Since the first introduction of MAS, and further applied to the market of400
MG, simulation of energy markets has accomplished an important role. Ra-401
machandran et al. [109] simulated multiple microgrid market scenario involving402
load and generation agents with ESS. MAS is designed for risk-based continu-403
ous double auction, achieving consensus among the agents and proving to be an404
useful tool.405
Before Eddy, Gooi & Chen [57], any MAS platform had been implemented406
via integration of microgrid market operations and DERs. Thus, Eddy, Gooi407
& Chen [57] simulated, with the assistance of the JADE plataform, market408
operations with distributed generators (DGs) and price-sensitive loads (such as409
air conditioning loads [110]) participating in a microgrid energy market. The410
MAS comprised several intelligent agents with access to its localized knowledge411
base. Demand agents were designed such that they were capable of regulating412
and controlling its power demand, activating and deactivating load units based413
on energy prices.414
3.4. Stand-alone smart-microgrids415
Stand-alone MGs have been considered as an efficient way being standardized416
for providing electricity in remote areas. Playing an important role in solving417
power supply problems [111], specially in remote areas (such as islands), these418
are the microgrid systems composed of RER and storage units.419
In the example presented in Figure 9, Zhao, Xue, Zhang, Wang & Zhao [5]420
defined seven types of agents for a stand-alone PV-small hydro hybrid MG. In421
their proposal, a small-hydro generation plant is controlled by the Frequency422
regulation agent (FRA), diesel generators are controlled by the Dispatchable DG423
agent (DDA), the PV system is controlled by the Intermittent DG agent (IDA),424
the BESS is controlled by the Energy storage agent (ESA), and the controllable425
loads are controlled by the Demand management agent (DMA). Thus, individual426
agents were implemented according to their defined tasks (forecasting ability,427
frequency control, among others) and the characteristics of the systems/devices428
that they were designed for. Coordination agents Schedule agent (SA) and429
Operation agent (OA) were the central rule decision makers following a client-430
server procedure.431
3.5. Microgrid demand control432
Torriti [112] remarked that increased awareness regarding consumption should433
bring conservation impacts and flatten peak demand. On the other hand, Sor-434
rel [113] pointed out some challenges in reducing energy demand, due to the435
strong correlation between increased wealth and increased energy consumption.436
Goulden, Bedwell, Rennick-Egglestone, Rodden & Spence [114] discussed the437
concepts of “energy consumer” and “energy citizen”, pointing out that we should438
recognize that SG users are actively engaged with energy, and it is critical to439
much of what is proposed by demand side management. They advocated the440
contrasting vision of an active citizen who becomes a “manager”, a potential441
MG prosumer.442
15
Figure 9: Real-time PXI-RTDS MAS simulation platform for a PV-small hydro hybrid mi-
crogrid [5].
Karfopoulos et al. [65] demonstrated MG users satisfaction in household in443
Spain, by the use of ICT and operational tools for distributed demand manage-444
ment. Consumers were intended to participate in grid support without affecting445
their level of satisfaction. An integrated home energy management system en-446
abling the provision of demand response services from residential customers was447
proposed, which was also able to operate under critical/emergency grid opera-448
tional conditions.449
Multi-agent reinforcement learning was used for coordination of consumer450
agents in an energy management tool proposed by Raju, Sankar & Milton [115].451
The consumer was modeled as an agent continuously interacting with the envi-452
ronment and learning how to take optimal actions. The main goal of the MAS453
was to achieve the long term objective of reducing total MG power consumption454
from grid.455
Self-demand control building models are been investigated tackling different456
house components [116]. Smart and energy-efficient building are also the focus457
of some MAS applications [117], such as building heat distribution control [118].458
A multi-agent home automation system for power management was idealized by459
Abras, Pesty, and Ploix & Jacomino [119] and is an interesting guideline for this460
kind of MAS application.461
3.6. Service restoration462
Prado et al. [120] mentioned the PDSR as a class of SG problems comprising463
service restoration, power loss reduction, and expansion planning, which are,464
nowadays, usually formulated as complex multi-objective and multi-constrained465
optimization problems. This class of problems requires time-consuming tasks to466
archive applicable solutions. The service restoration problem due to faults has467
been tackled by several works in the literature. An example of fault detection468
and reconfiguration is presented in Figure 10.469
For obvious reasons, SG should be stable and converge in case of any fault470
or when it falls within the above mentioned problems. One of the first MAS471
applications in service restoration was performed in 2000 by Nagata, Watan-472
abe, Ohno & Sasaki [121], consisting in a system with Bus Agents (BAGS) and473
16
Figure 10: Power grid service restoration. [120]
Facilitator Agent (FAG). A BAG was developed to decide a sub-optimal tar-474
get configuration after faults (interacting with other BAGs), while a FAG was475
acting as a manager for the decision process. Several other works in the liter-476
ature proposed using the distributed MAS approaches to solve the distribution477
system service restoration problems [122,123,58], also involving priority-based478
distribution-system restoration and contigency analysis [124,125].479
Saraiva & Asada [126] reconfigured the grid topology in order to satisfy the480
operation constraints, according to the data processed by agents dispersed in481
the grid.482
When dealing if MMG, owing to the complexity of the large networks DER,483
loads and storage units, the system requires the adoption of a hierarchical control484
scheme for providing flexibility and controllability. Thus, Resende, Gil & Lopez485
[127] introduced a hierarchical control system with easy abstraction for agent486
based system, allowing the coordination among DG units and MG. Peng, Song,487
Wang & Wang [128] advised that service restoration associated to MG scenarios488
should aim at important load power supplies. On the basis of hierarchical489
control strategy, they proposed three layer MAS architecture with: generation490
agent, load Agent and switch agent at the bottom layer; MG coordination and491
management agent in the middle layer; and, finally, a control agent in the top492
layer, checking technical operations.493
Recently, Yu, Von-Wun & Tsai [58] designed switch agents modeled as intel-494
ligent agents and organized as local power committees. Their approach captured495
the essence of Holonic Multi-Agent Systems (HMAS) [129], which provide self-496
adaptation and self-organization abilities able to assist management of large and497
complex systems, such as the case of service restoration in SG. The local com-498
mittees were in charge of reaching a consensus among many proposed service499
restoration solutions after faults were detected and isolated, as exemplified in500
Figure 10.501
4. Future MAS in the context of Smart-Microgrid502
4.1. Applications for allowing mini/microgrid autonomy503
An enhanced vision of the electric power grid should consider efficient power504
dispatching optimizers. This kind of demand-side management was explored505
by Atzeni et al. [130], considering a day-ahead optimization process regulated506
17
by a centralized agent. Mariani et al. [39] tackled the problem of optimal507
power dispatching in a smart-microgrid scenario seeking to minimize system to-508
tal costs. Mohammadi, Soleymani & Mozafari [131] generate a similar scenario509
considering uncertainties over the forecasting of consumption and renewable en-510
ergy generation. Coelho et al. [132] followed their ideas and idealized a power511
dispatching model, in a microgrid community scenario, with PEVs. In their512
work, a hybrid forecasting model [133] was used in order to generate probabilis-513
tic forecasts, which are more indicated for dealing with RER prediction. This514
field of smart energy scheduling in MG still request more attention for the next515
years. There is a visible lack between state-of-the-art centralized optimization516
techniques compared to decentralized approaches. Models that deal with dis-517
tributed agents might consider uncertainties over agents response or use robust518
optimization techniques [134].519
MAS applications have an important role in establishing communication520
between autonomous batteries, a truly system integration should, specially, in-521
clude PEVs and Unmanned Aerial Vehicle (UAV). Lausenhammer, Engel &522
Green [135] designed a system demand response, considering a fundamental523
game theoretic Java-based MAS framework assisted by PEVs storage units. As524
expected, with the increasing penetration of PEVs and small UAVs, the use of525
self-generated energy will increase. However, a scenario considering constraints526
and desires regarding to particular batteries (individual agents), in order to ob-527
tain mutual consensus of energy storage, remains an open problem. HCF and528
HAM systems (as described in Figure 3) could perform a mutual work in finding529
optimal schedules, letting the agents adapt and modify it in real-time.530
Some approaches in the literature incorporated the reduction of greenhouse531
gas emissions, as part of a multi-objective optimization problem [136,83,137].532
An open field for MAS applications is related to ecological environment mon-533
itoring and energy generation [138]. Decentralized systems have an incredible534
potential for contributing to energy sustainability, as pointed out by Hu, Li,535
Cao, Fang, He & Zhang [139]. Agents specialized in improving air quality,536
water drink-ability, growing of trees, lightning, and several other ecological in-537
dicators can now be mutual considered and negotiated between MG users, grid538
managers and infrastructure owners.539
The MG systems topology are dealing with different protocols between neigh-540
boring MG, these emerging systems are in the scope of community energy plan-541
ning, a topic that has been considered by academics [140]. The context of542
achieving mutual consensus between shared communitarian energy in MG will543
require specific desires set by users and citizens of each neighbor, in particular,544
in the future Smart Cities [141]. These naturally multi-objective optimization545
problems will require specific goals to be optimized for each agent. Future gen-546
erations of MAS should be prepared for a scenario with more bargaining and547
distinct objective functions between grid users.548
Testing MG systems are still being explored and several characteristics can549
still be sought for optimization and assisted by MAS. As emphasized by Lidula550
& Rajapakse [15], generic simulation models, reflecting MG properties, are still551
requested and would facilitate further researches focusing on improving transient552
18
stability performance, system protection, fault tolerance, novel control strategies553
and standard guidelines designs for MG.554
Prado et al. [120] highlighted that the performance obtained by metaheuris-555
tics applied for service restoration, in large-scale distribution systems, are dra-556
matically affected by the data structures used to represent electrical topology of557
the power grid system. For these cases, decentralized approaches for handling558
service restoration are an open field of research. The idea of a decentralized559
power flow calculus could increase the applications of MAS based tools over560
this class of problems [142]. More elaborated agents strategies and organiza-561
tions need to be explored for resolving the system after multiple faults, solving562
potentially subtle conflicts.563
As cited by Sanz, Rodrigues, Soler & Gallejo [30], new high performance564
control algorithms able to be executed in parallel, integrated, with standard565
power management mechanisms are required. In particular, greener alternative566
with low energy consumption [143], such as Graphics Processing Unit, could be567
integrated into the SM. Thus, efforts should be devoted to carefully design the568
new SM devices [144], implementing them with smart negotiation and commu-569
nication tools from MAS.570
The scenario of multiple microgrid is an eminent field of research, in which571
different characteristic can be optimized through efficient management and572
scheduling. Distribution networks become increasingly complex as the power573
industry moves towards decentralization [61] and the amount of barter between574
MG members might also increase. The action or system of exchanging goods575
or services without using money might also grow as different characteristics of576
the network might be optimized. Agents might start to consider the amount of577
resources that will come from another member of the grid instead of calculating578
only total profit or cost.579
4.2. System integration in emerging scenarios580
The development of MG systems in developing countries [145,146,47,147,581
148,149,48,150,151] is now well established and calls attention due to several582
challenges. As an example, let us consider the Amazon rain forest. Stand-alone583
photovoltaic plants are being deployed in riverside communities. Among several584
benefits, the native inhabitants highlight the reduction of noise and pollution585
in comparison to diesel generators. The communication between neighbor MG586
might reduce the need of investing in huge infrastructure, which would increase587
deforestation and might require money from investors. The role of negotiat-588
ing energy is not only useful for spreading energy around the world, but also589
for promoting and maintaining a health environment, without private sector590
interference where it is not necessary.591
However, even with an increase of RER generation and the self-sustainability592
of these communities, new smart mechanism of logistics and power dispatching593
are requested for implementing a low carbon environment. As a contradiction594
to the effort made, in the Amazon rain forest, the transport is still mainly done595
by small boats equipped with diesel generators. Some boats are now coming596
19
equipped with batteries. Will the advance of MAS contribute to the communi-597
cation between these small boats? Planning and scheduling the amount of oil598
that should be carried is a special type of the power energy dispatching prob-599
lem. Designing agents for interacting with these boats might be an interesting600
solution for achieving an efficient and skilled oil use.601
Out of the demographic areas, incorporation of RER into urban and every-602
day scenarios in emerging countries is another reality. Even residential loads603
are being disaggregated and handled with nonintrusive load monitoring [152],604
MAS applications in the autonomous control of residential houses is an open re-605
search area. A complex field of study with different combinatorial optimization606
problems, due to large set of small components and autonomous equipments, is607
promising not only for industries, but also for start-ups companies. The latter608
can now find a receptive market while trying to improve MG customers life609
quality.610
Alternative smart-microgrid environments are been quoted for RER imple-611
mentation, such as green roofs [153]. Tabrizi, Whale, Lyons & Urmee [44]612
pointed out that installation of small wind turbines in rooftops is not only fea-613
sible but also suggested by architects and project developers. This choice is a614
potential innovative way for incorporating sustainable energy generation into615
building design. Efforts in the context of improving wind energy production in616
urban areas were also pointed out by Ishugah, Li, Wang & Kiplagat [154], and617
also technical works aimed at improving its use in turbulent urban wind environ-618
ment [155]. Sarralde, Quinn, Wiesmann & Steemers[156] discussed solar energy619
and urban morphology, focusing on increasing the renewable energy potential620
of 4718 neighborhoods in London. Recently, Labeodan, Aduda, Boxem & Zeiler621
[157] detailed an overview study of the application of MAS in smart building622
operations and its interaction with smart grid. These smart decentralized tools623
would assist householders with better energy consumption management and624
blaze the trail for efficient autonomous green houses.625
5. Conclusions626
According to Ball [158]:627
“There are many arguments for and against the use of autonomous-628
agents in ambient intelligence and intelligent environments. Some629
researchers maintain that it is vital to restrict autonomy of agents630
so that users have complete control over the system; whereas, many631
others maintain that there is a greater benefit to be gained by em-632
ploying autonomous-agents to take some of the work load off the633
user and increase user convenience”.634
Although the opinions and concerns of people regarding autonomy in SG635
systems, specially in the context of mini/microgrids, can differ greatly from636
person to person, we believe in the flourishing of MAS applications for au-637
tonomous control of the future power grids. We suggest that further studies638
20
still grow in direction of greater energy grid autonomy. In particular, decen-639
tralizing the infrastructure, giving more weight to society wishes, as well as640
facilitating maintenance, reducing costs and opening doors for innovative ideas641
and the development of low-cost devices embedded with AI tools. Such auton-642
omy, managed by agents, guided by their own desires and rules, promotes the643
transition to a more decentralized scenario, in which users will be allowed to644
participate in the system more actively. On the other hand, several combinato-645
rial optimization problems will be opened to be improved and discussed along646
the next coming years. Novel negotiation protocols may arise in order to in-647
crease efficiency in such complex problems. On the other hand, those protocols648
might also be extended to other applications of our daily life.649
As verified along this paper, it was reported that from operation, manage-650
ment, security and efficient use of RER, MAS have been studied and usually651
presented feasible and reasonable performance. Thus, we finally emphasize the652
trend of implementing MAS in real-time MG systems, in particular, throughout653
the use of low-cost based technologies, such as Arduinos [31], Raspberry PI [159]654
and GPUs, which are flexible and posses enough power to be embedded with655
several high performance algorithms using AI techniques.656
Acknowledgment657
Vitor N. Coelho would like to thank the support given by CAPES, FAPERJ658
(inside the scope of the project “Autonomia em redes inteligentes assistida por659
ferramentas da inteligˆencia computacional”) and FP7 CORDIS, “New Horizons660
for Multi Criteria Decision Making”, for supporting the development of this661
work. Frederico G. Guimar˜aes was supported by the Brazilian agency CNPq662
(312276/2013-3 and 306850/2016-8) and FAPEMIG. Luis S. Ochi and Igor M.663
Coelho were supported by FAPERJ.664
References665
[1] C. T. Tugcu, I. Ozturk, A. Aslan, Renewable and non-renewable energy con-666
sumption and economic growth relationship revisited: Evidence from G7 coun-667
tries, Energy Economics 34 (6) (2012) 1942 – 1950. doi:http://dx.doi.org/668
10.1016/j.eneco.2012.08.021.669
[2] S. Hall, T. J. Foxon, Values in the smart grid: The co-evolving political economy670
of smart distribution, Energy Policy 74 (0) (2014) 600 – 609. doi:http://dx.671
doi.org/10.1016/j.enpol.2014.08.018.672
[3] M. Fadaeenejad, A. Saberian, M. Fadaee, M. Radzi, H. Hizam, M. AbKadir,673
The present and future of smart power grid in developing countries, Renewable674
and Sustainable Energy Reviews 29 (0) (2014) 828 – 834. doi:http://dx.doi.675
org/10.1016/j.rser.2013.08.072.676
[4] R. H. Lasseter, P. Paigi, Microgrid: a conceptual solution, in: Power Electronics677
Specialists Conference, 2004. PESC 04. 2004 IEEE 35th Annual, Vol. 6, 2004,678
pp. 4285–4290. doi:10.1109/PESC.2004.1354758.679
21
[5] B. Zhao, M. Xue, X. Zhang, C. Wang, J. Zhao, An MAS based energy manage-680
ment system for a stand-alone microgrid at high altitude, Applied Energy 143 (0)681
(2015) 251 – 261. doi:http://dx.doi.org/10.1016/j.apenergy.2015.01.016.682
[6] D. Olivares, A. Mehrizi-Sani, A. Etemadi, C. Canizares, R. Iravani, M. Kaz-683
erani, A. Hajimiragha, O. Gomis-Bellmunt, M. Saeedifard, R. Palma-Behnke,684
G. Jimenez-Estevez, N. Hatziargyriou, Trends in microgrid control, Smart Grid,685
IEEE Transactions on 5 (4) (2014) 1905–1919. doi:10.1109/TSG.2013.2295514.686
[7] F. Katiraei, M. R. Iravani, Power management strategies for a microgrid with687
multiple distributed generation units, IEEE transactions on power systems 21 (4)688
(2006) 1821–1831.689
[8] H. Farhangi, The path of the smart grid, Power and Energy Magazine, IEEE690
8 (1) (2010) 18–28. doi:10.1109/MPE.2009.934876.691
[9] D. Xenias, C. J. Axon, L. Whitmarsh, P. M. Connor, N. Balta-Ozkan, A. Spence,692
UK smart grid development: An expert assessment of the benefits, pitfalls and693
functions, Renewable Energy 81 (0) (2015) 89 – 102. doi:http://dx.doi.org/694
10.1016/j.renene.2015.03.016.695
[10] G. Kyriakarakos, A. I. Dounis, S. Rozakis, K. G. Arvanitis, G. Papadakis, Poly-696
generation microgrids: A viable solution in remote areas for supplying power,697
potable water and hydrogen as transportation fuel, Applied Energy 88 (12)698
(2011) 4517 – 4526. doi:http://dx.doi.org/10.1016/j.apenergy.2011.05.699
038.700
[11] G. Kyriakarakos, D. D. Piromalis, A. I. Dounis, K. G. Arvanitis, G. Pa-701
padakis, Intelligent demand side energy management system for autonomous702
polygeneration microgrids, Applied Energy 103 (2013) 39 – 51. doi:http:703
//dx.doi.org/10.1016/j.apenergy.2012.10.011.704
[12] C.-S. Karavas, G. Kyriakarakos, K. G. Arvanitis, G. Papadakis, A multi-agent705
decentralized energy management system based on distributed intelligence for706
the design and control of autonomous polygeneration microgrids, Energy Con-707
version and Management 103 (2015) 166–179.708
[13] A. Rogers, S. D. Ramchurn, N. R. Jennings, Challenges for autonomous agents709
and multi-agent systems research, in: Twenty-Sixth AAAI Conference on Arti-710
ficial Intelligence (AAAI-12), Toronto, CA, 2012, pp. 2166–2172.711
[14] C. Nguyen, A. Flueck, Agent based restoration with distributed energy storage712
support in smart grids, Smart Grid, IEEE Transactions on 3 (2) (2012) 1029–713
1038. doi:10.1109/TSG.2012.2186833.714
[15] N. W. A. Lidula, A. D. Rajapakse, Microgrids research: A review of experimental715
microgrids and test systems, Renewable and Sustainable Energy Reviews 15 (1)716
(2011) 186 – 202. doi:http://dx.doi.org/10.1016/j.rser.2010.09.041.717
[16] G. Merabet, M. Essaaidi, H. Talei, M. Abid, N. Khalil, M. Madkour, D. Ben-718
haddou, Applications of multi-agent systems in smart grids: A survey, in: Mul-719
timedia Computing and Systems (ICMCS), 2014 International Conference on,720
2014, pp. 1088–1094. doi:10.1109/ICMCS.2014.6911384.721
22
[17] S. McArthur, E. Davidson, V. Catterson, A. Dimeas, N. Hatziargyriou, F. Ponci,722
T. Funabashi, Multi-agent systems for power engineering applications – part i:723
Concepts, approaches, and technical challenges, Power Systems, IEEE Transac-724
tions on 22 (4) (2007) 1743–1752. doi:10.1109/TPWRS.2007.908471.725
[18] M. R. B. Khan, R. Jidin, J. Pasupuleti, Multi-agent based distributed control726
architecture for microgrid energy management and optimization, Energy Con-727
version and Management 112 (2016) 288–307.728
[19] J. Jimeno, J. Anduaga, J. Oyarzabal, A. G. de Muro, Architecture of a microgrid729
energy management system, European Transactions on Electrical Power 21 (2)730
(2011) 1142–1158. doi:10.1002/etep.443.731
URL http://dx.doi.org/10.1002/etep.443732
[20] H. Jiayi, J. Chuanwen, X. Rong, A review on distributed energy resources and733
microgrid, Renewable and Sustainable Energy Reviews 12 (9) (2008) 2472 –734
2483. doi:http://dx.doi.org/10.1016/j.rser.2007.06.004.735
[21] Y. A. R. I. Mohamed, E. F. El-Saadany, Adaptive decentralized droop con-736
troller to preserve power sharing stability of paralleled inverters in distributed737
generation microgrids, IEEE Transactions on Power Electronics 23 (6) (2008)738
2806–2816. doi:10.1109/TPEL.2008.2005100.739
[22] C. xia Dou, B. Liu, Hierarchical management and control based on MAS for740
distribution grid via intelligent mode switching, International Journal of Elec-741
trical Power & Energy Systems 54 (2014) 352 – 366. doi:http://dx.doi.org/742
10.1016/j.ijepes.2013.07.029.743
[23] W.-D. Zheng, J.-D. Cai, A multi-agent system for distributed energy resources744
control in microgrid, in: Critical Infrastructure (CRIS), 2010 5th International745
Conference on, 2010, pp. 1–5. doi:10.1109/CRIS.2010.5617485.746
[24] R. Romo, O. Micheloud, Power quality of actual grids with plug-in electric ve-747
hicles in presence of renewables and micro-grids, Renewable and Sustainable748
Energy Reviews 46 (0) (2015) 189 – 200. doi:http://dx.doi.org/10.1016/j.749
rser.2015.02.014.750
[25] F. Mwasilu, J. J. Justo, E.-K. Kim, T. D. Do, J.-W. Jung, Electric vehicles751
and smart grid interaction: A review on vehicle to grid and renewable energy752
sources integration, Renewable and Sustainable Energy Reviews 34 (2014) 501753
– 516. doi:http://dx.doi.org/10.1016/j.rser.2014.03.031.754
[26] T. Logenthiran, D. Srinivasan, D. Wong, Multi-agent coordination for DER in755
microgrid, in: Sustainable Energy Technologies, 2008. ICSET 2008. IEEE Inter-756
national Conference on, 2008, pp. 77–82. doi:10.1109/ICSET.2008.4746976.757
[27] M. Agrawal, A. Mittal, Overview of multi-agent approach for micro-grid energy758
management, Journal of Electrical and Electronics Engineering, IOSR-JEEE 9759
(2014) 01 – 07. doi:http://dx.doi.org/10.1016/j.rser.2014.03.031.760
[28] S. R. G, Methods and systems for cyber-physical security modeling, simulation761
and architecture for the smart grid (08 2013).762
URL http://worldwide.espacenet.com/publicationDetails/biblio?FT=D&763
date=20130808&DB=&locale=en_EP&CC=WO&NR=2013115936A1&KC=A1&ND=1764
23
[29] S. Goldsmith, Computing architecture for autonomous microgrids, WO Patent765
App. PCT/US2011/062,717 (06 2012).766
URL http://google.com/patents/WO2012078433A2?cl=en767
[30] J. J. Gomez-Sanz, S. Garcia-Rodriguez, N. Cuartero-Soler, L. Hernandez-768
Callejo, Reviewing microgrids from a multi-agent systems perspective, Energies769
7 (5) (2014) 3355–3382. doi:10.3390/en7053355.770
[31] S. D. Purusothaman, R. Rajesh, K. K. Bajaj, V. Vijayaraghavan, Implementa-771
tion of arduino-based multi-agent system for rural indian microgrids, in: 2013772
IEEE Innovative Smart Grid Technologies-Asia (ISGT Asia), IEEE, 2013, pp.773
1–5.774
[32] J. Ekanayake, K. Liyanage, J. Wu, A. Yokoyama, N. Jenkins, Smart grid: tech-775
nology and applications, John Wiley & Sons, 2012.776
[33] E. Commission, Smart grids – vision and strategy for europe’s electricity net-777
works of the future, Tech. Rep. EUR 22040, European Commission (2006).778
[34] J. K. Kaldellis, Stand-alone and hybrid wind energy systems: technology, energy779
storage and applications, Elsevier, 2010.780
[35] O. Ellabban, H. Abu-Rub, F. Blaabjerg, Renewable energy resources: Current781
status, future prospects and their enabling technology, Renewable and Sustain-782
able Energy Reviews 39 (0) (2014) 748 – 764. doi:http://dx.doi.org/10.783
1016/j.rser.2014.07.113.784
[36] N. Ederer, Evaluating capital and operating cost efficiency of offshore wind785
farms: A DEA approach, Renewable and Sustainable Energy Reviews 42 (2015)786
1034 – 1046. doi:http://dx.doi.org/10.1016/j.rser.2014.10.071.787
[37] Y. Levron, J. Guerrero, Y. Beck, Optimal power flow in microgrids with energy788
storage, Power Systems, IEEE Transactions on 28 (3) (2013) 3226–3234.789
[38] S. C. Manchester, L. G. Swan, D. Groulx, Regenerative air energy storage for790
remote wind-diesel micro-grid communities, Applied Energy 137 (0) (2015) 490791
– 500. doi:http://dx.doi.org/10.1016/j.apenergy.2014.06.070.792
[39] R. Rigo-Mariani, B. Sareni, X. Roboam, C. Turpin, Optimal power dispatching793
strategies in smart-microgrids with storage, Renewable and Sustainable Energy794
Reviews 40 (0) (2014) 649 – 658. doi:http://dx.doi.org/10.1016/j.rser.795
2014.07.138.796
[40] G. Comodi, A. Giantomassi, M. Severini, S. Squartini, F. Ferracuti, A. Fonti,797
D. N. Cesarini, M. Morodo, F. Polonara, Multi-apartment residential microgrid798
with electrical and thermal storage devices: Experimental analysis and simula-799
tion of energy management strategies, Applied Energy 137 (0) (2015) 854 – 866.800
doi:http://dx.doi.org/10.1016/j.apenergy.2014.07.068.801
[41] B. Zakeri, S. Syri, Electrical energy storage systems: A comparative life cycle802
cost analysis, Renewable and Sustainable Energy Reviews 42 (0) (2015) 569 –803
596. doi:http://dx.doi.org/10.1016/j.rser.2014.10.011.804
24
[42] P. Tinador, Superconducting magnetic energy storage; status and perspective,805
in: IEEE/CSC&ESAS European Superconductivity news forum, no. 3, 2008.806
[43] J. Garc´ıa-Villalobos, I. Zamora, J. S. Mart´ın, F. Asensio, V. Aperribay, Plug-in807
electric vehicles in electric distribution networks: A review of smart charging808
approaches, Renewable and Sustainable Energy Reviews 38 (2014) 717 – 731.809
doi:http://dx.doi.org/10.1016/j.rser.2014.07.040.810
[44] Rooftop wind monitoring campaigns for small wind turbine applications: Effect811
of sampling rate and averaging period.812
[45] G. Venayagamoorthy, P. Chakravarty, Optimal fuzzy logic based coordination813
controller for improved transient stability of a smart grid, in: Fuzzy Systems814
(FUZZ-IEEE), 2014 IEEE International Conference on, 2014, pp. 346–353. doi:815
10.1109/FUZZ-IEEE.2014.6891824.816
[46] L. E. Jones, Renewable energy integration: practical management of variability,817
uncertainty, and flexibility in power grids, Academic Press, 2014.818
[47] A. O. P. Jr, R. C. da Costa, C. do Vale Costa, J. de Moraes Marreco, E. L. L.819
Rovere, Perspectives for the expansion of new renewable energy sources in brazil,820
Renewable and Sustainable Energy Reviews 23 (0) (2013) 49 – 59. doi:http:821
//dx.doi.org/10.1016/j.rser.2013.02.020.822
[48] M. Welsch, M. Bazilian, M. Howells, D. Divan, D. Elzinga, G. Strbac, L. Jones,823
A. Keane, D. Gielen, V. M. Balijepalli, A. Brew-Hammond, K. Yumkella, Smart824
and just grids for sub-saharan africa: Exploring options, Renewable and Sus-825
tainable Energy Reviews 20 (0) (2013) 336 – 352. doi:http://dx.doi.org/10.826
1016/j.rser.2012.11.004.827
[49] S. Kahrobaee, R. A. Rajabzadeh, L.-K. Soh, S. Asgarpoor, Multiagent study828
of smart grid customers with neighborhood electricity trading, Electric Power829
Systems Research (2014) 123 – 132doi:http://dx.doi.org/10.1016/j.epsr.830
2014.02.013.831
[50] J. Lagorse, D. Paire, A. Miraoui, A multi-agent system for energy management832
of distributed power sources, Renewable Energy 35 (1) (2010) 174 – 182. doi:833
http://dx.doi.org/10.1016/j.renene.2009.02.029.834
[51] A. R. Khare, B. Y. Kumar, Multiagent structures in hybrid renewable power835
system: A review, Journal of Renewable and Sustainable Energy 7 (6). doi:836
http://dx.doi.org/10.1063/1.4934668.837
[52] R. Brown, Impact of smart grid on distribution system design, in: Power and838
Energy Society General Meeting - Conversion and Delivery of Electrical Energy839
in the 21st Century, 2008 IEEE, 2008, pp. 1–4. doi:10.1109/PES.2008.4596843.840
[53] N. Batista, R. Mel´ıcio, J. Matias, J. C. ao, Photovoltaic and wind energy systems841
monitoring and building/home energy management using zigbee devices within842
a smart grid, Energy 49 (2013) 306 – 315. doi:http://dx.doi.org/10.1016/843
j.energy.2012.11.002.844
25
[54] M. P. McHenry, Technical and governance considerations for advanced metering845
infrastructure/smart meters: Technology, security, uncertainty, costs, benefits,846
and risks, Energy Policy 59 (2013) 834 – 842. doi:http://dx.doi.org/10.847
1016/j.enpol.2013.04.048.848
[55] S. McArthur, E. Davidson, V. Catterson, A. Dimeas, N. Hatziargyriou, F. Ponci,849
T. Funabashi, Multi-agent systems for power engineering applications 2014 –850
part ii: Technologies, standards, and tools for building multi-agent systems,851
Power Systems, IEEE Transactions on 22 (4) (2007) 1753–1759. doi:10.1109/852
TPWRS.2007.908472.853
[56] F. Bellifemine, A. Poggi, G. Rimassa, Jade: A fipa2000 compliant agent de-854
velopment environment, in: Proceedings of the Fifth International Confer-855
ence on Autonomous Agents, ACM, New York, NY, USA, 2001, pp. 216–217.856
doi:10.1145/375735.376120.857
[57] Y. Foo Eddy, H. Gooi, S. Chen, Multi-agent system for distributed management858
of microgrids, Power Systems, IEEE Transactions on 30 (1) (2015) 24–34. doi:859
10.1109/TPWRS.2014.2322622.860
[58] W.-Y. Yu, V.-W. Soo, M.-S. Tsai, Power distribution system service restoration861
bases on a committee-based intelligent agent architecture, Engineering Applica-862
tions of Artificial Intelligence 41 (2015) 92 – 102. doi:http://dx.doi.org/10.863
1016/j.engappai.2015.01.011.864
[59] H. Kumar Nunna, S. Doolla, Multiagent-based distributed-energy-resource man-865
agement for intelligent microgrids, Industrial Electronics, IEEE Transactions on866
60 (4) (2013) 1678–1687. doi:10.1109/TIE.2012.2193857.867
[60] K. Wu, H. Zhou, A multi-agent-based energy-coordination control system for868
grid-connected large-scale wind – photovoltaic energy storage power-generation869
units, Solar Energy 107 (2014) 245 – 259. doi:http://dx.doi.org/10.1016/j.870
solener.2014.05.012.871
[61] M. Kouluri, R. Pandey, Intelligent agent based micro grid control, in: Intelligent872
Agent and Multi-Agent Systems (IAMA), 2011 2nd International Conference on,873
2011, pp. 62–66. doi:10.1109/IAMA.2011.6049007.874
[62] S. J. Russell, P. Norvig, J. F. Canny, J. M. Malik, D. D. Edwards, Artificial875
intelligence: a modern approach, Vol. 2, Prentice hall Upper Saddle River, 2003.876
[63] M. Wooldridge, Intelligent agents: The key concepts, in: ECCAI Advanced877
Course on Artificial Intelligence, Springer, 2001, pp. 3–43.878
[64] M. Wooldridge, An Introduction to MultiAgent Systems, 2nd Edition, John879
Wiley & Sons, 2009.880
[65] E. Karfopoulos, L. Tena, A. Torres, P. Salas, J. G. Jorda, A. Dimeas, N. Hatziar-881
gyriou, A multi-agent system providing demand response services from residen-882
tial consumers, Electric Power Systems Research 120 (0) (2015) 163 – 176, smart883
Grids: World’s Actual Implementations. doi:http://dx.doi.org/10.1016/j.884
epsr.2014.06.001.885
26
[66] R. Velik, P. Nicolay, A cognitive decision agent architecture for optimal energy886
management of microgrids, Energy Conversion and Management 86 (2014) 831887
– 847. doi:http://dx.doi.org/10.1016/j.enconman.2014.06.047.888
[67] T. Hong, J. Wilson, J. Xie, Long term probabilistic load forecasting and nor-889
malization with hourly information, Smart Grid, IEEE Transactions on 5 (1)890
(2014) 456–462.891
[68] Y. Zhang, J. Wang, X. Wang, Review on probabilistic forecasting of wind power892
generation, Renewable and Sustainable Energy Reviews 32 (0) (2014) 255 – 270.893
doi:http://dx.doi.org/10.1016/j.rser.2014.01.033.894
[69] R. Weron, Electricity price forecasting: A review of the state-of-the-art with a895
look into the future, International Journal of Forecasting 30 (4) (2014) 1030 –896
1081.897
[70] T. Logenthiran, D. Srinivasan, A. M. Khambadkone, Multi-agent system for en-898
ergy resource scheduling of integrated microgrids in a distributed system, Elec-899
tric Power Systems Research 81 (1) (2011) 138 – 148. doi:http://dx.doi.org/900
10.1016/j.epsr.2010.07.019.901
[71] D. Xenias, C. J. Axon, L. Whitmarsh, P. M. Connor, N. Balta-Ozkan, A. Spence,902
UK smart grid development: An expert assessment of the benefits, pitfalls and903
functions, Renewable Energy 81 (0) (2015) 89 – 102. doi:http://dx.doi.org/904
10.1016/j.renene.2015.03.016.905
[72] U. Al-mulali, H. G. Fereidouni, J. Y. Lee, C. N. B. C. Sab, Exploring the relation-906
ship between urbanization, energy consumption, and CO2 emission in MENA907
countries, Renewable and Sustainable Energy Reviews 23 (2013) 107 – 112.908
doi:http://dx.doi.org/10.1016/j.rser.2013.02.041.909
[73] A. Dimeas, N. Hatziargyriou, Operation of a multiagent system for microgrid910
control, Power Systems, IEEE Transactions on 20 (3) (2005) 1447–1455. doi:911
10.1109/TPWRS.2005.852060.912
[74] T. Funabashi, T. Tanabe, T. Nagata, R. Yokoyama, An autonomous agent for913
reliable operation of power market and systems including microgrids, in: Electric914
Utility Deregulation and Restructuring and Power Technologies, 2008. DRPT915
2008. Third International Conference on, 2008, pp. 173–177. doi:10.1109/DRPT.916
2008.4523397.917
[75] Y. Ueda, T. Nagata, Consideration of smart grid operations by multi-agent,918
Energy Procedia 14 (2012) 738 – 743, 2011 2nd International Conference on919
Advances in Energy Engineering (ICAEE). doi:http://dx.doi.org/10.1016/920
j.egypro.2011.12.1004.921
[76] Z. Jun, L. Junfeng, W. Jie, H. Ngan, A multi-agent solution to energy manage-922
ment in hybrid renewable energy generation system, Renewable Energy 36 (5)923
(2011) 1352 – 1363. doi:http://dx.doi.org/10.1016/j.renene.2010.11.032.924
[77] R. Roche, B. Blunier, A. Miraoui, V. Hilaire, A. Koukam, Multi-agent systems925
for grid energy management: A short review, in: IECON 2010 - 36th Annual926
Conference on IEEE Industrial Electronics Society, 2010, pp. 3341–3346. doi:927
10.1109/IECON.2010.5675295.928
27
[78] X. Zhou, Y. Gu, Y. Ma, L. Cui, S. Liu, Hybrid operation control method for929
micro-grid based on mas, in: Progress in Informatics and Computing (PIC),930
2010 IEEE International Conference on, Vol. 1, 2010, pp. 72–75. doi:10.1109/931
PIC.2010.5687398.932
[79] E. Kuznetsova, Y.-F. Li, C. Ruiz, E. Zio, An integrated framework of agent-based933
modelling and robust optimization for microgrid energy management, Applied934
Energy 129 (0) (2014) 70 – 88. doi:http://dx.doi.org/10.1016/j.apenergy.935
2014.04.024.936
[80] L. J. Ricalde, E. Ordonez, M. Gamez, E. N. Sanchez, Design of a smart grid937
management system with renewable energy generation, in: Computational Intel-938
ligence Applications In Smart Grid (CIASG), 2011 IEEE Symposium on, 2011,939
pp. 1–4. doi:10.1109/CIASG.2011.5953346.940
[81] X. Tan, Q. Li, H. Wang, Advances and trends of energy storage technology in941
microgrid, International Journal of Electrical Power & Energy Systems 44 (1)942
(2013) 179 – 191. doi:http://dx.doi.org/10.1016/j.ijepes.2012.07.015.943
[82] M. Baran, I. El-Markabi, A multiagent-based dispatching scheme for distributed944
generators for voltage support on distribution feeders, Power Systems, IEEE945
Transactions on 22 (1) (2007) 52–59. doi:10.1109/TPWRS.2006.889140.946
[83] C. Colson, M. Nehrir, C. Wang, Ant colony optimization for microgrid multi-947
objective power management, in: Power Systems Conference and Exposition,948
2009. PSCE’09. IEEE/PES, 2009, pp. 1–7.949
[84] G. Kyriakarakos, D. D. Piromalis, K. G. Arvanitis, A. I. Dounis, G. Papadakis,950
On battery-less autonomous polygeneration microgrids: Investigation of the951
combined hybrid capacitors/hydrogen alternative, Energy Conversion and Man-952
agement 91 (2015) 405–415.953
[85] P. Zhang, F. Yan, C. Du, A comprehensive analysis of energy management954
strategies for hybrid electric vehicles based on bibliometrics, Renewable and955
Sustainable Energy Reviews 48 (0) (2015) 88 – 104. doi:http://dx.doi.org/956
10.1016/j.rser.2015.03.093.957
[86] Intelligent unit commitment with vehicle-to-grid – a cost-emission optimization,958
Journal of Power Sources 195 (3) (2010) 898 – 911.959
[87] J. Hu, A. Saleem, S. You, L. Nordstrom, M. Lind, J. Ostergaard, A multi-960
agent system for distribution grid congestion management with electric vehicles,961
Engineering Applications of Artificial Intelligence 38 (0) (2015) 45 – 58. doi:962
http://dx.doi.org/10.1016/j.engappai.2014.10.017.963
[88] B. Ramachandran, S. K. Srivastava, D. A. Cartes, Intelligent power management964
in micro grids with EV penetration, Expert Systems with Applications 40 (16)965
(2013) 6631 – 6640. doi:http://dx.doi.org/10.1016/j.eswa.2013.06.007.966
[89] J. Lagorse, M. Simoes, A. Miraoui, A multiagent fuzzy-logic-based energy man-967
agement of hybrid systems, Industry Applications, IEEE Transactions on 45 (6)968
(2009) 2123–2129. doi:10.1109/TIA.2009.2031786.969
28
[90] C.-H. Yoo, I.-Y. Chung, H.-J. Lee, S.-S. Hong, Intelligent control of battery970
energy storage for multi-agent based microgrid energy management, Energies971
6 (10) (2013) 4956–4979. doi:10.3390/en6104956.972
[91] D. Ye, M. Zhang, D. Sutanto, Decentralised dispatch of distributed energy re-973
sources in smart grids via multi-agent coalition formation, Journal of Parallel974
and Distributed Computing 83 (2015) 30 – 43. doi:http://dx.doi.org/10.975
1016/j.jpdc.2015.04.004.976
[92] M. Pipattanasomporn, H. Feroze, S. Rahman, Multi-agent systems in a dis-977
tributed smart grid: Design and implementation, in: Power Systems Con-978
ference and Exposition, 2009. PSCE ’09. IEEE/PES, 2009, pp. 1–8. doi:979
10.1109/PSCE.2009.4840087.980
[93] Z. Xiao, T. Li, M. Huang, J. Shi, J. Yang, J. Yu, W. Wu, Hierarchical mas981
based control strategy for microgrid, Energies 3 (9) (2010) 1622–1638. doi:982
10.3390/en3091622.983
[94] X. Dou, X. Quan, Z. Wu, M. Hu, K. Yang, J. Yuan, M. Wang, Hybrid multi-984
agent control in microgrids: Framework, models and implementations based on985
IEC 61850, Energies 8 (1) (2014) 31–58. doi:10.3390/en8010031.986
[95] Z. Wu, W. Gu, Active power and frequency control of islanded microgrid based987
on multi-agent technology [j], Electric Power Automation Equipment 11 (2009)988
57–61.989
[96] W. Liu, W. Gu, B. Zhao, A novel multi-agent based control approach for fre-990
quency stabilization of islanded microgrids, in: Innovative Smart Grid Tech-991
nologies Europe (ISGT EUROPE), 2013 4th IEEE/PES, 2013, pp. 1–5. doi:992
10.1109/ISGTEurope.2013.6695328.993
[97] J. Peas Lopes, C. Moreira, A. Madureira, Defining control strategies for micro-994
grids islanded operation, Power Systems, IEEE Transactions on 21 (2) (2006)995
916–924. doi:10.1109/TPWRS.2006.873018.996
[98] H.-M. Kim, T. Kinoshita, Y. Lim, T.-H. Kim, A bankruptcy problem approach997
to load-shedding in multiagent-based microgrid operation, Sensors 10 (10) (2010)998
8888–8898. doi:10.3390/s101008888.999
[99] M. Rahman, M. Mahmud, H. Pota, M. Hossain, A multi-agent approach for1000
enhancing transient stability of smart grids, International Journal of Electrical1001
Power & Energy Systems 67 (0) (2015) 488 – 500. doi:http://dx.doi.org/10.1002
1016/j.ijepes.2014.12.038.1003
[100] M. A. da Rosa, A. M. L. da Silva, V. Miranda, Multi-agent systems applied1004
to reliability assessment of power systems, International Journal of Electrical1005
Power & Energy Systems 42 (1) (2012) 367 – 374. doi:http://dx.doi.org/10.1006
1016/j.ijepes.2012.03.048.1007
[101] K. Wilkosz, A multi-agent system approach to power system topology verifica-1008
tion, in: Proceedings of the 8th International Conference on Intelligent Data1009
Engineering and Automated Learning, IDEAL’07, Springer-Verlag, Berlin, Hei-1010
delberg, 2007, pp. 970–979.1011
29
[102] S. Mocci, N. Natale, F. Pilo, S. Ruggeri, Demand side integration in LV smart1012
grids with multi-agent control system, Electric Power Systems Research 1251013
(2015) 23 – 33. doi:http://dx.doi.org/10.1016/j.epsr.2015.03.021.1014
[103] A. M. Farid, Multi-agent system design principles for resilient coordination &1015
control of future power systems, Intelligent Industrial Systems 1 (3) (2015) 255–1016
269. doi:10.1007/s40903-015-0013-x.1017
[104] A. Farid, Static resilience of large flexible engineering systems: Axiomatic design1018
model and measures, Systems Journal, IEEE PP (99) (2015) 1–12. doi:10.1109/1019
JSYST.2015.2428284.1020
[105] M. Pipattanasomporn, H. Feroze, S. Rahman, Securing critical loads in a pv-1021
based microgrid with a multi-agent system, Renewable Energy 39 (1) (2012)1022
166–174.1023
[106] European research project more microgrids, http://microgrids.power.ece.1024
ntua.gr, accessed: 2015-10-21.1025
[107] C. Yuen, A. Oudalov, A. Timbus, The provision of frequency control reserves1026
from multiple microgrids, Industrial Electronics, IEEE Transactions on 58 (1)1027
(2011) 173–183. doi:10.1109/TIE.2010.2041139.1028
[108] D. Bertsekas, D. Casta˜non, A forward/reverse auction algorithm for asymmet-1029
ric assignment problems, Computational Optimization and Applications 1 (3)1030
(1992) 277–297. doi:10.1007/BF00249638.1031
[109] B. Ramachandran, S. Srivastava, C. Edrington, D. Cartes, An intelligent auc-1032
tion scheme for smart grid market using a hybrid immune algorithm, In-1033
dustrial Electronics, IEEE Transactions on 58 (10) (2011) 4603–4612. doi:1034
10.1109/TIE.2010.2102319.1035
[110] S. Horowitz, B. Mauch, F. Sowell, Forecasting residential air conditioning1036
loads, Applied Energy 132 (2014) 47 – 55. doi:http://dx.doi.org/10.1016/1037
j.apenergy.2014.06.029.1038
[111] B. Zhao, X. Zhang, J. Chen, C. Wang, L. Guo, Operation optimization of1039
standalone microgrids considering lifetime characteristics of battery energy stor-1040
age system, Sustainable Energy, IEEE Transactions on 4 (4) (2013) 934–943.1041
doi:10.1109/TSTE.2013.2248400.1042
[112] J. Torriti, People or machines? assessing the impacts of smart meters and load1043
controllers in italian office spaces, Energy for Sustainable Development 20 (2014)1044
86 – 91. doi:http://dx.doi.org/10.1016/j.esd.2014.01.006.1045
[113] S. Sorrell, Reducing energy demand: A review of issues, challenges and ap-1046
proaches, Renewable and Sustainable Energy Reviews 47 (0) (2015) 74 – 82.1047
doi:http://dx.doi.org/10.1016/j.rser.2015.03.002.1048
[114] M. Goulden, B. Bedwell, S. Rennick-Egglestone, T. Rodden, A. Spence, Smart1049
grids, smart users? the role of the user in demand side management, Energy1050
Research & Social Science 2 (2014) 21 – 29. doi:http://dx.doi.org/10.1016/1051
j.erss.2014.04.008.1052
30
[115] L. Raju, S. Sankar, R. Milton, Distributed optimization of solar micro-grid using1053
multi agent reinforcement learning, Procedia Computer Science 46 (2015) 231 –1054
239. doi:http://dx.doi.org/10.1016/j.procs.2015.02.016.1055
[116] D. Zhang, N. Shah, L. G. Papageorgiou, Efficient energy consumption and op-1056
eration management in a smart building with microgrid, Energy Conversion1057
and Management 74 (2013) 209 – 222. doi:http://dx.doi.org/10.1016/j.1058
enconman.2013.04.038.1059
[117] S. Smitha, F. Chacko, Intelligent energy management in smart and sustain-1060
able buildings with multi-agent control system, in: Automation, Computing,1061
Communication, Control and Compressed Sensing (iMac4s), 2013 International1062
Multi-Conference on, 2013, pp. 190–195. doi:10.1109/iMac4s.2013.6526406.1063
[118] M. Mokhtar, M. Stables, X. Liu, J. Howe, Intelligent multi-agent system for1064
building heat distribution control with combined gas boilers and ground source1065
heat pump, Energy and Buildings 62 (0) (2013) 615 – 626. doi:http://dx.doi.1066
org/10.1016/j.enbuild.2013.03.045.1067
[119] S. Abras, S. Pesty, S. Ploix, M. Jacomino, An anticipation mechanism for power1068
management in a smart home using multi-agent systems, in: Information and1069
Communication Technologies: From Theory to Applications, 2008. ICTTA 2008.1070
3rd International Conference on, 2008, pp. 1–6. doi:10.1109/ICTTA.2008.1071
4530305.1072
[120] R. S. Prado, R. C. P. Silva, O. M. Neto, F. G. G. aes, D. S. Sanches, J. ao Bosco1073
A. London Jr., A. C. Delbem, Differential evolution using ancestor tree for ser-1074
vice restoration in power distribution systems, Applied Soft Computing 23 (2014)1075
498 – 508. doi:http://dx.doi.org/10.1016/j.asoc.2014.06.005.1076
[121] T. Nagata, H. Watanabe, M. Ohno, H. Sasaki, A multi-agent approach to power1077
system restoration, in: Power System Technology, 2000. Proceedings. PowerCon1078
2000. International Conference on, Vol. 3, 2000, pp. 1551–1556 vol.3. doi:1079
10.1109/ICPST.2000.898202.1080
[122] J. Solanki, S. Khushalani, N. Schulz, A multi-agent solution to distribution1081
systems restoration, Power Systems, IEEE Transactions on 22 (3) (2007) 1026–1082
1034. doi:10.1109/TPWRS.2007.901280.1083
[123] Y.-T. Pan, M.-S. Tsai, Development a bdi-based intelligent agent architecture1084
for distribution systems restoration planning, in: Intelligent System Applications1085
to Power Systems, 2009. ISAP ’09. 15th International Conference on, 2009, pp.1086
1–6. doi:10.1109/ISAP.2009.5352961.1087
[124] A. Oonsivilai, K. A. Greyson, Power system contingency analysis using multia-1088
gent systems, World Academy of Science, Engineering and Technology 60 (2009)1089
355–360.1090
[125] C.-H. Lin, H.-J. Chuang, C.-S. Chen, C.-S. Li, C.-Y. Ho, Fault detection, isola-1091
tion and restoration using a multiagent-based distribution automation system,1092
in: Industrial Electronics and Applications, 2009. ICIEA 2009. 4th IEEE Con-1093
ference on, 2009, pp. 2528–2533. doi:10.1109/ICIEA.2009.5138663.1094
31
[126] F. O. Saraiva, E. N. Asada, Sistema multiagente aplicado ao gerenciamento1095
topol´ogico de sistema el´etricos de distribu´ı¸c˜ao inteligentes, in: Anais do XIX1096
Congresso Brasileiro de Autom´atica – CBA 2012., 2012, pp. 1490–1497.1097
[127] F. O. Resende, N. J. Gil, J. A. P. Lopes, Service restoration on distribution sys-1098
tems using multi-microgrids, European Transactions on Electrical Power 21 (2)1099
(2011) 1327–1342. doi:10.1002/etep.404.1100
[128] P. Li, B. Song, W. Wang, T. Wang, Multi-agent approach for service restoration1101
of microgrid, in: Industrial Electronics and Applications (ICIEA), 2010 the 5th1102
IEEE Conference on, 2010, pp. 962–966. doi:10.1109/ICIEA.2010.5515722.1103
[129] P. Leitao, P. Valckenaers, E. Adam, Self-adaptation for robustness and cooper-1104
ation in holonic multi-agent systems, in: A. Hameurlain, J. K¨ung, R. Wagner1105
(Eds.), Transactions on Large-Scale Data- and Knowledge-Centered Systems I,1106
Vol. 5740 of Lecture Notes in Computer Science, Springer Berlin Heidelberg,1107
2009, pp. 267–288. doi:10.1007/978-3-642-03722- 1_11.1108
[130] I. Atzeni, L. G. O. nez, G. Scutari, D. P. Palomar, J. R. Fonollosa, Demand-1109
side management via distributed energy generation and storage optimization,1110
IEEE Transactions on Smart Grid 4 (2) (2013) 866–876. doi:10.1109/TSG.1111
2012.2206060.1112
[131] S. Mohammadi, S. Soleymani, B. Mozafari, Scenario-based stochastic operation1113
management of microgrid including wind, photovoltaic, micro-turbine, fuel cell1114
and energy storage devices, International Journal of Electrical Power & Energy1115
Systems 54 (0) (2014) 525 – 535. doi:http://dx.doi.org/10.1016/j.ijepes.1116
2013.08.004.1117
[132] V. N. Coelho, I. M. Coelho, B. N. Coelho, M. W. Cohen, A. J. Reis, S. M. Silva,1118
M. J. Souza, P. J. Fleming, F. G. Guimaraes, Multi-objective energy storage1119
power dispatching using plug-in vehicles in a smart-microgrid, Renewable Energy1120
89 (2016) 730 – 742. doi:http://dx.doi.org/10.1016/j.renene.2015.11.084.1121
[133] V. N. Coelho, I. M. Coelho, B. N. Coelho, A. J. Reis, R. Enayatifar, M. J. Souza,1122
F. G. Guimaraes, A self-adaptive evolutionary fuzzy model for load forecasting1123
problems on smart grid environment, Applied Energy 169 (2016) 567 – 584.1124
doi:http://dx.doi.org/10.1016/j.apenergy.2016.02.045.1125
[134] E. Kuznetsova, C. Ruiz, Y.-F. Li, E. Zio, Analysis of robust optimization1126
for decentralized microgrid energy management under uncertainty, Interna-1127
tional Journal of Electrical Power & Energy Systems 64 (2015) 815 – 832.1128
doi:http://dx.doi.org/10.1016/j.ijepes.2014.07.064.1129
[135] W. Lausenhammer, D. Engel, R. Green, Utilizing capabilities of plug in elec-1130
tric vehicles with a new demand response optimization software framework:1131
Okeanos, International Journal of Electrical Power & Energy Systems 75 (2016)1132
1 – 7. doi:http://dx.doi.org/10.1016/j.ijepes.2015.08.014.1133
[136] E. Alvarez, A. Lopez, J. G´omez-Aleixandre, N. de Abajo, On-line minimiza-1134
tion of running costs, greenhouse gas emissions and the impact of distributed1135
generation using microgrids on the electrical system, in: Sustainable Alter-1136
native Energy (SAE), 2009 IEEE PES/IAS Conference on, 2009, pp. 1–10.1137
doi:10.1109/SAE.2009.5534847.1138
32
[137] H. Kanchev, D. Lu, B. Francois, V. Lazarov, Smart monitoring of a microgrid1139
including gas turbines and a dispatched pv-based active generator for energy1140
management and emissions reduction, in: Innovative Smart Grid Technologies1141
Conference Europe (ISGT Europe), 2010 IEEE PES, 2010, pp. 1–8. doi:10.1142
1109/ISGTEUROPE.2010.5638875.1143
[138] F. Bousquet, C. L. Page, Multi-agent simulations and ecosystem management:1144
a review, Ecological Modelling 176 (3-4) (2004) 313 – 332. doi:http://dx.doi.1145
org/10.1016/j.ecolmodel.2004.01.011.1146
[139] Z. Hu, C. Li, Y. Cao, B. Fang, L. He, M. Zhang, How smart grid contributes1147
to energy sustainability, Energy Procedia 61 (2014) 858 – 861, international1148
Conference on Applied Energy, ICAE2014. doi:http://dx.doi.org/10.1016/1149
j.egypro.2014.11.982.1150
[140] Z. Huang, H. Yu, Z. Peng, M. Zhao, Methods and tools for community energy1151
planning: A review, Renewable and Sustainable Energy Reviews 42 (0) (2015)1152
1335 – 1348. doi:http://dx.doi.org/10.1016/j.rser.2014.11.042.1153
[141] S. Karnouskos, T. N. De Holanda, Simulation of a smart grid city with software1154
agents, in: Computer Modeling and Simulation, 2009. EMS’09. Third UKSim1155
European Symposium on, IEEE, 2009, pp. 424–429.1156
[142] T. Erseghe, Distributed optimal power flow using admm, Power Systems, IEEE1157
Transactions on 29 (5) (2014) 2370–2380. doi:10.1109/TPWRS.2014.2306495.1158
[143] D. B. Kirk, W. H. Wen-mei, Programming massively parallel processors: a1159
hands-on approach, 2nd Edition, Morgan Kaufmann, 2012.1160
[144] G. Pepermans, Valuing smart meters, Energy Economics 45 (2014) 280 – 294.1161
doi:http://dx.doi.org/10.1016/j.eneco.2014.07.011.1162
[145] P. D´ıaz, J. Mas´o, Evolution of production and the efficient location of renewable1163
energies. the case of china, Energy Procedia 40 (0) (2013) 15 – 24, european1164
Geosciences Union General Assembly 2013, {EGUDivision}Energy, Resources1165
& the Environment, {ERE}.doi:http://dx.doi.org/10.1016/j.egypro.1166
2013.08.003.1167
[146] H.-T. Pao, H.-C. Fu, Renewable energy, non-renewable energy and economic1168
growth in brazil, Renewable and Sustainable Energy Reviews 25 (0) (2013) 3811169
– 392. doi:http://dx.doi.org/10.1016/j.rser.2013.05.004.1170
[147] G. S. Alem´an-Nava, V. H. Casiano-Flores, D. L. C´ardenas-Ch´avez, R. D´ıaz-1171
Chavez, N. Scarlat, J. Mahlknecht, J.-F. Dallemand, R. Parra, Renewable en-1172
ergy research progress in mexico: A review, Renewable and Sustainable Energy1173
Reviews 32 (0) (2014) 140 – 153. doi:http://dx.doi.org/10.1016/j.rser.1174
2014.01.004.1175
[148] J. Yuan, J. Shen, L. Pan, C. Zhao, J. Kang, Smart grids in china, Renewable1176
and Sustainable Energy Reviews 37 (0) (2014) 896 – 906. doi:http://dx.doi.1177
org/10.1016/j.rser.2014.05.051.1178
33
[149] C.-C. Lin, C.-H. Yang, J. Z. Shyua, A comparison of innovation policy in the1179
smart grid industry across the pacific: China and the {USA}, Energy Policy1180
57 (0) (2013) 119 – 132. doi:http://dx.doi.org/10.1016/j.enpol.2012.12.1181
028.1182
[150] P. Acharjee, Strategy and implementation of smart grids in india, Energy Strat-1183
egy Reviews 1 (3) (2013) 193 – 204, future Energy Systems and Market Integra-1184
tion of Wind Power. doi:http://dx.doi.org/10.1016/j.esr.2012.05.003.1185
[151] A. Mahmood, N. Javaid, A. Zafar, R. A. Riaz, S. Ahmed, S. Razzaq, Pak-1186
istan’s overall energy potential assessment, comparison of lng, TAPI and IPI gas1187
projects, Renewable and Sustainable Energy Reviews 31 (0) (2014) 182 – 193.1188
doi:http://dx.doi.org/10.1016/j.rser.2013.11.047.1189
[152] K. X. Perez, W. J. Cole, J. D. Rhodes, A. Ondeck, M. Webber, M. Baldea,1190
T. F. Edgar, Nonintrusive disaggregation of residential air-conditioning loads1191
from sub-hourly smart meter data, Energy and Buildings 81 (2014) 316 – 325.1192
doi:http://dx.doi.org/10.1016/j.enbuild.2014.06.031.1193
[153] O. Saadatian, K. Sopian, E. Salleh, C. Lim, S. Riffat, E. Saadatian,1194
A. Toudeshki, M. Sulaiman, A review of energy aspects of green roofs, Re-1195
newable and Sustainable Energy Reviews 23 (2013) 155 – 168. doi:http:1196
//dx.doi.org/10.1016/j.rser.2013.02.022.1197
[154] T. Ishugah, Y. Li, R. Wang, J. Kiplagat, Advances in wind energy resource1198
exploitation in urban environment: A review, Renewable and Sustainable Energy1199
Reviews 37 (2014) 613 – 626. doi:http://dx.doi.org/10.1016/j.rser.2014.1200
05.053.1201
[155] F. To ja-Silva, A. Colmenar-Santos, M. Castro-Gil, Urban wind energy exploita-1202
tion systems: Behaviour under multidirectional flow conditions – opportunities1203
and challenges, Renewable and Sustainable Energy Reviews 24 (2013) 364 – 378.1204
doi:http://dx.doi.org/10.1016/j.rser.2013.03.052.1205
[156] J. J. Sarralde, D. J. Quinn, D. Wiesmann, K. Steemers, Solar energy and1206
urban morphology: Scenarios for increasing the renewable energy potential1207
of neighbourhoods in london, Renewable Energy 73 (2015) 10 – 17, sustain-1208
able Development in Building and Environment (SuDBE) 2013. doi:http:1209
//dx.doi.org/10.1016/j.renene.2014.06.028.1210
[157] T. Labeodan, K. Aduda, G. Boxem, W. Zeiler, On the application of multi-1211
agent systems in buildings for improved building operations, performance and1212
smart grid interaction – a survey, Renewable and Sustainable Energy Reviews1213
50 (2015) 1405 – 1414. doi:http://dx.doi.org/10.1016/j.rser.2015.05.081.1214
[158] M. Ball, V. Callaghan, A study of agent autonomy in intelligent environments,1215
Journal of Ambient Intelligence and Smart Environments 1 (0) (2012) 1 – 38.1216
[159] M. Richardson, S. Wallace, Getting started with raspberry PI, ” O’Reilly Media,1217
Inc.”, 2012.1218
34
