ENERGY MANAGEMENT ALGORITHMS IN SMART GRIDS: STATE OF THE ART AND EMERGING TRENDS

International Journal of Artificial Intelligence and Applications (IJAIA), Vol. 7, No. 4, July 2016
DOI: 10.5121/ijaia.2016.7403 25
ENERGY MANAGEMENT ALGORITHMS IN SMART
GRIDS: STATE OF THE ART AND EMERGING
TRENDS
Joelle Klaimi1, 2
,Rana Rahim-Amoud1, 3
,Leila Merghem-Boulahia2
and Akil Jrad1, 3
1
EDST, LaSTRe, Lebanese University, Lebanon.
2
ICD/ERA (UMR CNRS 6281), Troyes University of Technology, France.
3
Faculty of sciences, Lebanese University, Lebanon.
ABSTRACT
The electric grid is radically evolving into the smart grid, which is characterized by improved energy
efficiency of available resources. The smart grid permits interactions among its computational and physical
elements thanks to the integration of Information and Communication Technologies (ICTs). ICTs provide
energy management algorithms and allow renewable energy integration and energy price minimization.
Given the importance of renewable energy, many researchers developed energy management (EM)
algorithms to minimize renewable energy intermittency. EM plays an important role in the control of users'
energy consumption and enables increased consumer participation in the market. These algorithms provide
consumers with information about their energy consumption patterns and help them adopt energy-efficient
behaviour. In this paper, we present a review of the state of the energy management algorithms. We define
a set of requirements for EM algorithms and evaluate them qualitatively. We also discuss emerging tools
and trends in this area.
KEYWORDS
Smart grids, Energy management, Renewable resources, Storage systems, Multi-agent systems
1. INTRODUCTION
There is a growing worldwide interest in the evolution of the smart grid [1], a modern power grid
that supports bidirectional communication between energy providers and consumers for fine-
grained metering, control, and feedback. One of the key features of the smart grid is enhancing
energy efficiency and manageability of available resources [2]. According to the European
Commission Task Force on Smart Grid, the smart grid is defined as an electricity network that
can intelligently integrate the action of all users connected to it - generators, consumers and those
that do both - in order to ensure economically efficient and sustainable power system with low
losses, high level of quality, security of supply and safety [3]. In fact, fossil fuels represent the
major energy resources in most countries, these resources release toxic gas that have severe
consequences on the habitats and affect human health [4]. Furthermore, fossil resources are
exhaustible and no longer available once used. To offer a greener and a cheaper solution, the
smart grid incorporates the distributed renewable resources such as wind, solar and hydroelectric.
The renewable resources reduce fuel consumption and CO2 emissions. Moreover, the integration
of distributed renewable resources in the smart grids minimizes system losses and increases the
reliability, efficiency and security of electricity supply to consumers [5, 6]. (1) These resources
are inexhaustible. Even (1), they are however non-scheduled resources because they depend on
weather and natural phenomena which makes it complex to balance supply and demand [7]. In

26
this context, the smart grid helps consumers (residential, commercial, etc.) to be energy suppliers;
they are then called prosumers. In addition, this new grid permits the two-way flow of
information and electricity between consumers and electric power companies. Figure 1 shows this
two-way flow in all communication infrastructures in smart grid architecture. These
infrastructures include home area networks (HANs), business area networks (BANs),
neighborhood area networks (NANs), data centers, and substation automation integration systems
[8].
Figure 1 Illustration of smart grid communication network [8]
The two-way communication in the smart grid architecture is responsible for transmitting power
grid sensing and measuring status, as well as the control messages [9] which helps consumers to
control their energy usage [10]. In order to control the grid energy usage and to give utility
companies the access to the requested information, smart grid includes new smart entities like
smart meters [11]. Entities communication can be wired (power line communication, etc.), or
wireless (cellular networks, sensor networks, etc.) [12]. To provide near real-time information to
utilities and end users, a smart grid integrates distributed computing and communication systems.
In addition, this smart grid introduces several technologies such as information and
communication technologies to ensure reliability, demand management, storage, distribution and
transport of electrical energy [13]. Facing the necessity of renewable resources integration,
energy management algorithms represent a great solution to make a decision about power
delivering in efficient ways. In order to reduce consumers' energy bill, energy management
provides autonomous actions in controlling different electrical grid subsystems (energy
generation, distribution and consumption).
In the literature, we can find many definitions of energy management. Based on the definitions
given in [14, 15, 16, 17, 18, 19], the energy management can be defined as a set of strategies and
functions that can adjust and optimize the energy use. These set of functions increase the energy
efficiency and coordinate the energy resources. It is the process of observing, controlling and
conserving electricity usage in a building, a neighborhood, etc. Energy management should be
able to optimize costs and to minimize the risk of loss of production excess.
To make intelligent decisions about energy management functions, multi-agent algorithms have
been adopted by several researchers. The Multi-Agent System (MAS) is a paradigm used when
the problem to resolve is complex. The MAS is an important tool for self-control in widely
distributed systems whose characteristics are very dynamic [20]. A MAS is composed of a set of
autonomous and intelligent entities called agents deployed in a shared environment and acting
intelligently upon that environment, these agents interact and cooperate in order to reach a

27
specific objective [21]. Each agent in MAS has a degree of smartness depending on storage and
management role and/or functions [22].
In order to meet consumers' needs, the smart grid exchanges a huge amount of expensive data
[23]. In this regard, cloud computing is a response for the vast data exchange in the smart power
grid. Cloud computing presents many advantages in on-demand self-service, ubiquitous network
access and location independent resource pooling. Furthermore, cloud providers own their
immense data centers responsible for computation, storage capacities, applications and services
[23, 24]. As a brief definition, the cloud computing is a procedure to share resources. By using the
cloud infrastructure, users can access their applications anytime and from anywhere through a
device connected to the network [25]. Cloud computing discharges the users from a number of
constraints. However, the cloud computing infrastructure is not suitable to latency-sensitive
applications. This fact encourages researchers to explore fog computing. Fog computing is the
extension of the cloud computing paradigm to the edge of the network. It presents several
advantages comparing to the cloud computing: low latency, wide-spread geographical
distribution, very large number of nodes, etc. [26, 27, 28]. In fact, cloud and fog computing
architectures have been studied by many researchers that they could have a role in the smart grid
energy management.
The purpose of this article is to present some energy management algorithms used in smart grids
or microgrids. The rest of this paper is organized as follows. We present an overview of demand
response programs and energy storage in section 2. An explanation of multi-agent concept, cloud
computing and fog computing paradigm will be given in section 3. In section 4 we present some
algorithms of energy management in smart grids and microgrids. We conclude our paper in the
section 5.
2. ENERGY MANAGEMENT REQUIREMENTS
The introduction of renewable resources has generated problems of intermittency and irregularity.
The demand response is necessary for the stability of the electricity network operation. Utilities
have to offer tariffs that motivate consumers to save energy and/or shift loads to off-peak periods.
Furthermore, buffering and storing energy should be a solution for renewable energies problems.
In 2012, Sofana et al. and Law et al. published an extensive review of demand response and
energy management modeling and feedback techniques [29, 30]. More recently, Vardakas et al.
[31] and Deng et al. [32] offered a detailed review of feedback methods for energy management
and demand response in smart grid architecture. In the next subsection, the demand response
programs in the smart grids will be defined.
2.1. DEMAND RESPONSE PROGRAMS
The Demand Response (DR) is an opportunity for the consumers' participation in electricity
markets. The smart grids can give an important support for the integration of demand response
[33]. Based on [34, 35, 36], the DR consists in changing the normal electrical use of end users in
response to price changes. The DR has an important role to play in electricity market, it can be a
solution to minimize the demands especially in peak hours or when system reliability is
jeopardized. In other words, this DR program aims to smooth consumption in peak hours and
reduce the cost of energy by shifting demands from peak hours to off-peak hours. Consumers
who participate to the DR have two ways to change their electricity usage [34, 36]: (1) reduce
energy consumption during critical peak periods when prices are high without changing their
consumption during off-peak hours, this way involves a temporary loss of comfort. (2) Shift some
of the peak demand operations to off-peak periods. Moreover, there are different classifications of

28
the DR programs that can be found in literature and can be divided into three categories: the rate-
based or price DR programs, the incentive or event-based DR programs and the demand reduction
bids [34]. Firstly, in the rate-based DR program, DR is implemented over utility tariffs in the
markets according to variable electricity prices in order to motivate consumers to adjust their
consumption patterns and to react to the fluctuations in the electricity prices. In fact, consumers
pay a high price during high peak periods and a low price during off-peak periods [34]. Secondly,
the event-based DR program rewards all consumers who minimize their consumption upon
request. Additionally, this program can be a response to many conditions like local or regional
grid congestion, system economics, or operational reliability requirements, local or system
temperature, etc. [34]. Thirdly, in the demand reduction bids DR program, consumers initiate and
send demand reduction bids to the utility. These bids consist in the available demand reduction
capacity and the requested price [34].
The smart grid consumers use "Demand Response Technologies" which are by definition smart
grid technologies. Those technologies include products or services that help in the active
monitoring and dynamic control of electricity usage. The smart-meters are one of the most well-
known examples, these meters measure quantities and calculate costs. They can also measure in a
time interval prices, among others, and send them to the consumers in order to reduce and shift
usage on peak periods. Furthermore, the meters also allow new and better information to be
generated and used by both the consumer and the electricity providers. This information can be
presented to consumers via in home display devices, which help consumers understanding their
electricity usage. Other technologies such as smart thermostats, dynamic lighting controls and
dynamic energy storage systems are used as DR technologies. Thus, the deployment of the
demand response can lead to higher and more sustainable levels of energy efficiency and lower
toxic gas emissions (CO2, etc.).
Energy efficiency usually refers to devices or practices that provide the same level of output or
benefit by using less energy. It usually focuses on reducing overall energy use, not just at certain
times. The DR improves the overall efficiency of the electricity system but differs from the
traditional energy efficiency by being more dynamic, meaning that it can be used to meet rising
demands.
In fact, the DR programs may offer a broad range of potential benefits on system operation and
energy efficiency. The participant bill savings can be considered as the main benefit of the DR
program [34, 35, 36]. On the other hand, some concerns may discourage the consumers from
participating in DR programs as summarized in [35]: (1) the uncertainty of the price. (2) The
undefined quantity of load that might be available for reduction during an event. (3) The
economic viability of participating in a DR program. (4) The willingness to maintain occupant
comfort during a DR event.
The energy storage systems can be a solution that helps the DR programs to resolve the
intermittency problem of renewable resources. In the next subsection we will present an overview
of the energy storage systems.
2.2. ENERGY STORAGE
Renewable energy resources, such as solar or wind, offer a greener solution compared to
traditional energy resources such as fossil fuels. However, their intermittent nature makes it
difficult to balance demand and supply, which is essential in a power grid. Both electricity
generation from the renewable energy resources and electricity demand are stochastic processes
[37]. The Energy Storage Devices (ESD) should be a solution for random energy problem as they

29
can smooth out variations in generation and demand [37]. The energy storage has long been
identified as an important technology to leverage supply and demand. In an operational context,
the energy storage can be used to store energy excess when supply exceeds demand and then
dispatch it when it is needed which improves the energy efficiency. The energy storage can
alleviate the need to generate power exactly at the time it is needed. Furthermore, it can smooth
out the variations of energy utility due to random power demand and uncertain energy supply.
The energy variations smoothing is desirable due to economic considerations and to the
increasing incorporation of intermittent renewable resources. Intuitively, the total electricity cost
can be reduced by recharging the battery from the electric power grid when the electricity price is
low while discharging it during the high electricity price period [16]. Consumers may also benefit
from lower-cost off-peak energy by using storage technologies in order to postpone some
consumption operations to an overnight shift. There are many types of energy storage devices:
mechanical, thermo-dynamic, electrochemical and electro-magnetic, each with its own
characteristics. In fact, these characteristics are shown in Figure 2 and include [38]:
• ESD size that is the maximum amount of energy that can be stored in an ESD.
• Storage charging and discharging rate limit which is the limit on the charge or discharge
power.
• ESD efficiency is defined by the degree of reducing each energy unit to be used later.
• Self-discharge that is the leaks of the stored energy over time.
• Depth-of-Discharge (DoD) is used to describe how deeply the battery is discharged.
• Temperature-dependency that is the temperature that can affect the charging and
discharging rates and the lifetime of a battery.
• State-of-Health (SoH) that is either expressed in terms of capacity fade or charging power
fades.
Figure 2 Storage systems characteristics
In residential areas, energy storage is practically synonymous with batteries [39]. Due to the
popularity of the emerging PHEVs (Plug-in Hybrid Electric Vehicle), the battery of a PHEV may
be treated as temporary electricity storage for the residential consumers [19]. Given the
importance of the energy storage system, the lack of its integration is one of the drawbacks of
current energy management algorithms. The development of energy management algorithms and
tools can include storage to optimize the energy cost. Furthermore, other service tools for smart
grid energy management such as multi-agent systems, cloud and fog computing will be shown in
the following section.
3. SERVICE TOOLS FOR ENERGY MANAGEMENT IN SMART GRIDS
Many service tools have an important role in the smart grid energy management (e.g. multi-agent
systems, cloud computing, fog computing) and will be presented in the rest of this section.

30
3.1. THE MULTI-AGENT SYSTEMS
In order to explore the potential benefits of MAS in energy management, the concepts of agents
and multi-agent systems need to be understood. In this context, we will discuss and define these
paradigms in this section.
3.1.1. WHAT IS AN AGENT?
To make an intelligent decision about energy management in the smart grid, some researchers
used intelligent entities called agents. Intelligent agents are a new paradigm for developing
software applications. Recently, agents have been the focus of intense research in the fields of
computer science and artificial intelligence. Agents are pieces of software operating on a network
as a representative of a person or a device in order to perform pre-defined tasks, and allow
embedding intelligent system techniques in a large distributed information system. Intelligent
agents can also be utilized in the planning and operation of decentralized supply and demand. As
shown in Figure 3, an agent is defined as an entity that perceives its environment through sensors
and acts upon it through actuators [40].
Figure 3 Basic agent behaviour
Furthermore, we can define the agent environment simply by everything external to the agent.
The environment may be physical (e.g., the power system), and therefore observable through
physical sensors, or it may be virtual representing the computing environment (e.g., data
resources, computing resources, and other agents), observable through system calls, program
invocation, and messaging [41]. A multi-agent system is a system composed of by two or more
intelligent agents. In the next paragraph we will detail the multi-agent systems.
3.1.2. WHAT IS A MAS?
The multi-agent systems are now needed for a range of applications including diagnostics,
condition monitoring, power system restoration, market simulation, network control and
automation [41]. An agent cooperates with other agents to solve problems autonomously [42, 43,
44]. According to Roche et al. [45] and McArthur et al. [41], agents have three main
characteristics. Firstly they are reactive so they can react to the environment rapidly. Secondly,
these agents are objective-oriented so they are pro-active. Thirdly, these agents are sociable so
they are able to negotiate and communicate together in their own language. Negotiation is a key
issue for the multi-agent system to harmonize knowledge and resources and solve the conflicts
based on a communication language. Through negotiation, agents solve some questions [46, 47,
48]. Many researchers used MAS in energy management but others believed in the effectiveness
of cloud computing and fog computing for the smart grids energy management.

31
3.2. CLOUD COMPUTING AND FOG COMPUTING ARCHITECTURES
3.2.1 CLOUD COMPUTING CONCEPT
Cloud computing is defined as a model that presents on-demand facilities and shared resources
over the Internet. Cloud computing is implemented on software resources such as stand-alone
operating systems and hardware infrastructures like servers, storages and networks [49]. There
are three distinct types of services in the cloud computing: (1) the Infrastructure as a Service
(IaaS), (2) the Platform as a Service (PaaS) and (3) the Software as a Service (SaaS). The
providers of IaaS offer physical and virtual resources such as computers, storage and virtual
machines to users. Load balancing in cloud computing is performed using IaaS. PaaS is the
service that develops and delivers programming models to IaaS while SaaS is responsible for
helping users to run their applications without installing the software on their own computers.
Furthermore, the use of the cloud computing offers many advantages such:
• Elastic nature: cloud computing devices can change their memory capacity according to
the users demand.
• Shared architecture: information can be shared between all users to reduce service costs.
• Metering architecture: metering architecture enables users to pay for their consumed
resources only, thereby, it helps in cost optimization.
• Internet services: cloud computing supports the existing network infrastructure.
Figure 4 represents the cloud computing infrastructure. Moreover, the cloud computing can have
three deployment models. Firstly, the private cloud is the model owned by a private organization.
In this model, information is shared only within the organization. Secondly, the public cloud is
owned by a service provider. Thirdly, the hybrid cloud is a composition of two clouds techniques.
Figure 5 shows the hierarchical cloud computing deployment models. In fact, the cloud
applications can be served from virtual energy storage and data storage devices. In addition, the
components of the smart grid can communicate with the cloud and take decisions concerning
energy management. Finally, the use of cloud services in smart grids energy management field
avoids some problems (e.g. energy management, information management, security, etc.)
encountered in this area [25]. For a more distributed system, the concept of fog computing was
evolved. In the following paragraph we will explain the fog computing paradigm.
Figure 4 Cloud Computing infrastructure [25] Figure 5 Cloud Computing
deployment models
3.2.2. FOG COMPUTING PARADIGM
Fog computing is an extension of cloud computing to the edge of the network. In other words, fog
computing is a cloud computing but close to the "ground" [27]. In fact, fog computing consists of

32
computing, storage and networking services between end users and traditional cloud computing.
Furthermore, in fog computing, the services can be hosted at the end devices such as set-top-
boxes or access points. By comparing it to cloud, fog computing presents a number of advantages
that make it a non-trivial extension of the cloud. These advantages are as follows: the low latency,
the high geographical distribution, the large number of nodes supported and the real time
interaction [26]. For sake of clarity, an example of fog devices interaction is shown in Figure 6. In
this example, there is a simple three levels hierarchy: (1) the cloud level, (2) the fog level and (3)
the smart devices level. In this framework, each smart device is attached to one of fog servers. In
addition, fog servers could be interconnected and each of them is linked to the cloud. Authors in
[27] highlight the role of Fog computing in the smart grids applications. In the smart grid, energy
load balancing must run on the network edge devices (smart meters and microgrids). These Fog
devices automatically switch to alternative energy (solar or wind), based on demand, availability
and lowest price. Figure 7 presents an example of fog computing use into the smart grid. As
shown in Figure 7, fog collectors located at the edge handle the data generated by the grid sensors
and devices. In addition, these collectors send control commands to the actuators. They also filter
the data to be exploited locally, and send the rest to the higher tiers for visualization, real-time
reports and transaction analysis. Finally, this example shows that fog supports ephemeral storage
at the lowest tier to semi-permanent storage at the highest tier. After presenting some tools that
can help in smart grid energy management, we will present in the following section some energy
management algorithms for smart grids based on these tools.
4. POWER MANAGEMENT IN MODERN ELECTRICAL INFRASTRUCTURE
Tomorrow's consumers should negotiate and interact with their external environment. In addition,
keeping the balance between demand and supply will constitute an important issue in the future
electrical grid, hence the need for energy management tools. Several works studied the energy
management in the smart grids or the microgrids domain but they have some drawbacks like the
lack of energy storage systems, etc. A microgrid is an electrical system that incorporates
renewable energy as a primary power resource, battery storage and loads. Microgrids are the
optimal way to integrate renewable resources on the community level and allow for consumer
participation in the electricity market [50]. Microgrids aim to improve energy production and
delivery for local consumers, while facilitating a more stable electrical infrastructure with benefits
towards the environmental emissions, the energy conservation, and the operational cost [15].
Furthermore, a microgrid will operate in two distinct states: interconnected with a larger utility
network or islanded from it [15].
Figure 6 Example of fog computing Figure 7 Fog computing example in
deployment use [27] the smart grid context [27]
In this section some energy management algorithms in the microgrids or in the smart grids will be
presented. The rest of this section is organized as follows. In the subsection 4.1, we will show the

33
energy prediction and the real time control use in energy management in the smart grid context.
The subsection 4.2 will summarize some energy management algorithms that use the multi-agent
technologies in the smart grids. In the subsection 4.3, we will explore some energy management
algorithms using cloud computing paradigm. In the last subsection 4.4, we will present a
qualitative comparison of all energy management algorithms shown in the previous subsections.
4.1. ENERGY PREDICTION AND REAL TIME CONTROL FOR POWER MANAGEMENT
As explained previously, the penetration of the renewable energy is an increasing phenomenon
and the intermittent nature of these resources creates an imbalance between production and
consumption. Moreover, the use of non-scheduled renewable energy makes it difficult to balance
supply and demand.
4.1.1. PREDICTIONS AND ENERGY MANAGEMENT
Energy management algorithms analysis must include prediction process in order to improve the
optimization results [51]. Most predictions are done over 24 time slots, each corresponding to an
hour of a day. This day-ahead prediction is often adopted to predict weather conditions, load
levels and market prices.
In the smart grid context where decision is based on information exchange about supply and
demands, the smart grid decision makes the load and renewable energy forecasting a key
component. In fact, each region has its own resource production and energy consumption based
on year time and day time. Based on the previous electrical power production demands, the
behavior of the load can be forecasted and estimated. Several factors influence the load in the
electrical network: the weather situation (temperature, cloud coverage, etc.), the economic
activity (huge modifications of load forecasting are necessary during the holiday periods), the
standard working hours, etc. Classic methods of load forecasting are based on meteorological
information and historic consumption data. These predicted values are used to make a more
flexible system. The forecast information are exactly those used by the smart grid to improve the
system efficiency [10, 52, 53]. Several researchers used the information predictions in the context
of smart grid energy management. They use mathematical algorithms (integer, quadratic,
dynamic and stochastic programming, game theory, evolutionary algorithms, etc.) in order to
predict information [13].
Some algorithms have used storage systems in order to help in smart grid energy management
[54, 55]. Authors in [54] used a system composed of many storage agents as shown in Figure 8.
In this algorithm, the authors used the predicted grid price and the predicted users demand to store
the needed energy in the storage systems. Furthermore, each storage agent chooses the best
moment to buy energy from the grid, which corresponds to the period when the predicted grid
price is low. In addition, all consumers buy their demands from storage agents so they do not
access to the grid. In this algorithm, the authors aim to minimize the energy price but in the other
hand, they do not use the renewable resources and they do not consider the case when the
consumer can buy the energy directly from the grid and minimize the transportation losses.
Similarly, some research efforts have focused on whole home energy consumption predictions at
time frames of a day [56, 57]. A comprehensive energy management solution, however, requires
a broad understanding of prediction accuracy at different entities (e.g. appliance and home) as
well as different time horizons (hour, day, or week). A broad understanding of predictability may
be used to inform the design of energy management solutions by allowing a designer to tune the

34
system to tolerate prediction error, or to make smart decisions about integrating or excluding
appliances known to have unpredictable usage patterns.
Another optimization load control scheme based on Linear Programming (LP) was proposed in
[58]. The main goal of this research is to minimize the households' bill. To realize this goal, the
authors proposed an optimization-based residential load control scheme that uses real time pricing
market (RTP). A price predictor unit estimates the future prediction prices by using weighted
averaging filter to past prices. This scheme combines the prediction of price and consumption in
order to schedule demands and make an optimal decision.
Subramanian et al. [59] proposed a day-ahead price-optimization algorithm for electrical utilities
in the smart grid framework. This research has the ability to predict the day-ahead pricing plans in
order to manage residential electricity loads. This study uses historical data from the real-world
resources to calibrate the underlying model and to measure the quality of the model fit. The
proposed algorithm reduces costs by optimally shifting peak loads to off-peak periods, thereby
reducing the need for spot-market electricity usage. The research in [60] proposed predictive
control model of energy management in the smart grid to control and minimize the cost function.
A scheduling of battery usage is determined based on many parameters. These parameters can be
defined as variable cost values, power consumption and generation profiles, as well as functional
constraints under uncertainty due to variations in the power generator model. This algorithm does
not use a storage system.
In order to reduce the intermittency of photovoltaic resources, Kanchev et al. [14] developed an
energy management algorithm based on predicted data. The authors used two storage
technologies at the consumer level to enable grid demand management and renewable energy
integration. Batteries are the first storage technologies. They are used to ensure an energy reserve
for the grid operator. In addition, supercapacitors are considered as the second storage
technologies in this algorithm and are used to balance fast power variations coming from the
photovoltaic generator and from the primary frequency control. A strategy has been presented to
drive storage systems according to the solar energy resources and grid requirements. In this
proposed strategy, a deterministic operational power planning has been proposed to perform the
day-ahead power scheduling for the conventional and photovoltaic generators. The presented
scheme relies on photovoltaic power predictions and load forecasting. The scheme also sets out
plans for the use of the distributed battery storage.
Figure 8 The storage agent architecture by Alberola et al. [54]

35
In the researches done in [14, 58, 59, 60], the developed algorithm does not adjust the predicted
values when a perturbation occurs (i.e. new consumer in the network, the failure of a production
unit, etc.).
Unlike previous researches, Rose et al. [53] developed a new scoring rule mechanism called Sum
of Others' plus Max (SOM). This new scheme contains a centralized agent that aggregates all
predicted information. Based on aggregated information, the centralized agent buys all needed
energy to the system. In order to encourage agents to report their true demands to the center, this
last rewards agents from a budget that is equal to the savings made by using the agents'
information. By using the proposed algorithm (SOM), the center distributes the savings to the
agents in a budget-balanced way. As the algorithm is centralized, this may calls into question the
scalability of the algorithm and constitutes bottlenecks.
Instead of centralization used in [53], Wang et al. [52] used a distributed system in order to
control the energy storage systems based on predicted photovoltaics output. The authors argued
that the prediction of the power generation and load power consumption is very important to
develop a storage control algorithm for a better energy management in the context of households.
This algorithm adjusts at each time slot all the prediction values when a disturbance occurs and
the predicted value does not match the real value. Based on predicted values, the authors
developed an algorithm that helps to plan the future charging and discharging schemes for storage
systems in the next periods and chooses the best way to charge the battery. However, this
algorithm does not use intelligent interactions between consumers in order to reduce bill.
Similary, Cecati et al. [61] proposed an energy management algorithm that reduces the energy
cost for consumers and resolves the problem of renewable resources intermittency in the smart
grid context. The proposed algorithm uses the generation and load forecasts to determine the
optimum way to purchase energy. Firstly, the energy management system schedules a one day-
ahead predicted values. Secondly, a real time intra-day optimization operation is proposed. This
real time optimization can modify the predicted values in order to consider the operation
requirements. To resolve the problem of the high data integration into the existing energy
management systems, the authors use a learning architecture based on fully connected neural
networks. In this way, they can lower the computational time needed to predict the optimal
schedule of generation. Moreover, they can allow benefits such as short processing time, required
by real-time operation, and a simple implementation. However, this algorithm does not use the
intelligent interaction between consumers in order to reduce bill.
4.1.2. REAL TIME CONTROL STRATEGY FOR ENERGY MANAGEMENT
Generally, some perturbations may occur on the system causing a dispatch between the predicted
values and the real ones. Real time energy management is a procedure used in the smart grid or in
the microgrid to manage the grid usage, the storage and the production in order to satisfy all
energy demands with a minimum cost. During real-time operations, the generation and the load
must be matched.
An algorithm based on the predictions and the real-time energy management is proposed by
Hooshmand et al. [62]. This research work uses the energy predictions and the real-time energy
management in order to relax the issue of sudden unforecasted unbalances between supply and
demand. The energy storage systems are used in this algorithm in order to minimize the costs.
The proposed system consists of the four following nodes' types: (1) the demand node D(t) that is
the node energy consumption. (2) The node of renewable energy which is made up of wind
energy, solar energy, etc. (3) The grid node that is able to buy exceeded energy from producers or
to sell the needed energy to end users. (4) The storage node stores unused renewable energy or

36
imports energy with low prices from the grid. In this research, the authors use the forecasted load
profile and renewable energy generation to minimize the cost of energy. In this context, they
proposed a novel Model Predictive Control (MPC) which is by definition a control system. It will
be responsible for energy management using current and future information of the load profiles of
renewable energy, battery state of charge SOC, etc. MPC can be defined as a class of control
policies which uses a model that projects the behavior of the system under control. The proposed
MPC corrects any prediction error, solves the optimization problem, minimizes real-time
generation of energy and optimizes energies costs. However, this research does not take into
consideration the sudden uses of the energy (e.g new user in the system). In order to reduce
severe environment impacts, Sechilariu et al. [63] proposed an energy management algorithm in
the smart grid. In their research, the authors focus on building-integrated microgrid design and
applied an energy storage and smart grid communication. The researchers use a photovoltaic
resource in order to reduce toxic gas emissions. The main goal of this work is to balance the
supply and the demand in a real-time. In this context, a power balancing strategy using smart grid
interactions was proposed. The proposed strategy reduces grid peak consumption, avoids
undesirable grid power injection and makes full use of local photovoltaic production. However, in
this research the authors do not use predictions in order to minimize users' bill.
The research in [16] focuses on a study to minimize the expected electricity cost with real-time
electricity pricing. This research proposes an algorithm that uses energy storage to harvest
excessive renewable energy generation for later use. In addition, the energy storage charges the
battery when the electricity price is low while discharging it when the electricity price is high.
The authors increase battery capacity by using optimization techniques. This research focuses on
elastic energy demands (computers, televisions, etc.) and proposed to queue those demands and
serve them with a delay to guarantee the finite worst case delay for any buffered energy demand.
However, this algorithm does not use energy predictions in order to optimize bill.
An important number of energy management algorithms use a multi-agent system for smarter
solutions, we will define some of those algorithms in the next paragraphs.
4.2 ENERGY MANAGEMENT USING MULTI-AGENT SYSTEMS
Several researchers adopted multi-agent systems for energy management in the smart grids or the
microgrids. Roche et al. [64] proposed an energy management algorithm in the smart grids that
uses the MAS. In this scheme, the authors used a multi-agent system shown in Figure 9.
The proposed multi-agent architecture consists of several agents, divided into three categories:
• Control Agents: they include physical components like turbines and storage. A turbine
agent is created for each turbine in the microarray and resides in the turbine control
system. It communicates with the SCADA agent, sends its current state through its
actions, and receives instructions in return.
• Management System Agents: SCADA, dispatching and planning. The SCADA agent
plays the role of a supervisor and business controller. Based on the data received from the
SCADA agent, the dispatch agent calculates the optimum operating point for each of the
micro-network agents. It also maintains a database containing the needed information to
determine the correct distribution of powers by using economic and environmental data.
The planning agent is responsible for planning in advance, if possible optimal and
operation of the set, with production forecasts, demand and prices issued by agents.
• Auxiliary staff: it includes prediction agent that predicts users' data and production.

37
Figure 9 Illustration of Roche et al. system agents Figure 40 The system components by
Nagata et al. algorithm [42]
The authors in [64] ensure the energy efficiency but they do not negotiate the energy purchase to
optimize the consumers' bill. An energy management algorithm for the smart grids was proposed
by Logenthiran et al. [65]. The objective of this research is to develop and simulate a multi-agent
system that enables an electricity market for the operation of a microgrid in both islanded and
grid-connected modes. In this research, the microgrid maximizes its profit by optimizing the
power exchange with the main grid. During the islanded mode operation, the microgrid satisfies
the local energy demand using its local power production and minimizes load losses. During grid-
connected mode, the microgrid sells power to internal load of the microgrid and also exchanges
power with the main grid at market price. When the amount of power produced by microgrid
resources is not enough, or too expensive to cover the local load, power is bought from the
upstream network and sold to the consumers at the same price. The multi-agent system in this
algorithm is used in order to plan in real-time the produced energy and the storage system use.
However, this algorithm does not allow negotiating the energy purchase to minimize the
consumers’ bills.
Unlike the proposal described in [65], an algorithm for energy management in smart grid using a
mutli-agent system and a negotiation algorithm was proposed by Nagata et al. [42]. The authors
proposed this algorithm in order to reduce the energy costs. The proposed system components are
shown in Figure 10. The selling price (SP0) and the buying price (BP0) are two parameters
announced by the grid operator. The smart grid controller unit (SGC) in this system aims to
negotiate and optimize the buying prices for consumers. Moreover, production units and
consumption units belong to the smart grid. These units adjust their set point (SPi and BPi) after
negotiation with other units based on the grid price, their operational costs, and the load demand.
The system is constructed with six types of agents:
1. The Grid Agent (GridAG): it has two principal roles: buying kWh from the smart grid or
selling kWh to it.
2. The Generator Agent (GAG): its role is to maximize the total benefit.
3. The Seller Agent (SAG): created by the generator agent to sell kWh to the market.
4. The Load Agent (LAG): its role is to create several buyer agents.
5. The Buyer Agent (BAG): its role is to buy kWh from the market.
6. The smart grid Control Agent (SGC): its role is to optimize the smart grid operation by
using a negotiation algorithm.

38
It is worth noting that, in this algorithm, each SAG and BAG is assigned to a specific amount of
power. The centralization and the lack of storage systems can be considered as disadvantages of
this research.
Colson et al. [15] proposed an energy management algorithm for microgrid, using a multi-agent
system. The authors integrate three basic agent types: the producer agent, the consumer agent and
the observer agent. The responsibilities of the producer agent can be defined as:
• Monitoring power.
• Determining the cost of the power supplied by the component.
• Providing information to other agents.
The consumer agent is responsible for:
• Monitoring the amount of consumed power.
• Determining the amount of the instantaneous active and reactive power.
• Negotiating the purchase of energy.
While responsibilities of the observer agent are:
• Monitoring specific parameters within the network of micro-network.
• Providing information to other agents regarding the state of the node.
In this algorithm, consumer agents use negotiation to get the energy at the lowest price. The lack
of the energy storage system integration is the drawback of this algorithm.
Similarly, an agent-based algorithm was proposed in [66]. In this work, the authors proposed an
agent-based algorithm allowing more efficient use of energy while minimizing cost and taking
into account the integration of distributed renewable energy resources. This proposed system is
composed of three types of agents:
1. Generator Company agents (GC): they represent the big generator companies that are the
main resources of energy.
2. Prosumer agents: they represent small resources of renewable energy distributed near the
consumers. They obviously produce and consume energy. They produce a smaller
quantity of electricity than GC.
3. Consumer agents: the regular consumer agent is an agent that cannot produce energy. It is
only a purchaser of energy.
In this proposal, prosumers prefer to sell the unused quantity of energy. Furthermore, the
consumers act to determine and negotiate the purchase of energy in order to reduce their bill. The
goal of each consumer agent is to maximize its profit in terms of unit price paid per day. Each
consumer divides its data into three priorities. Firstly, the vital loads are satisfied. The consumer
negotiates the cost and chooses the resource that minimizes its own cost of purchase. In this
proposal, the demands of consumers are first met with supply from renewable resources. If the
renewable resources cannot meet the total energy requirements of all consumers, then the
traditional resources are used to meet these demands. In this algorithm, the authors do not use the
energy storage systems that would improve outcomes and reduce the consumers’ bill.
Similarly, Vytelingum et al. [67] argued that using micro-storage devices at home to save power
via the Smart Grid is reducing the burning of fossil fuels. In this research, all residential
consumers are modeled as autonomous agents (SmartHome), responsible for shifting their
consumption in order to maximize the residential revenue in addition to several social welfare
factors. They proposed an agent-based micro-storage management technique with a provided
general framework within which to analyze the Nash equilibrium of an electricity grid and devise

39
new agent-based storage learning strategies that adapt to the market conditions. However, this
research does not take into consideration users' comfort.
Ramchurn et al. [68] proposed a MAS-based control mechanism to incorporate consumer
preferences in managing demand. All agents, in this algorithm, represent consumers and can
communicate with the grid and optimize energy consumption. This algorithm helps in controlling
a system containing green energy suppliers operating within the grid and a number of individual
homes. This approach reduces consumers' energy bill, but does not take into consideration users'
comfort.
The authors in the research described in [69, 70] proposed a novel energy management algorithm
for the smart grids using multi-agent systems, intelligent storage systems, and a negotiation
algorithm. In this work, the authors integrate the storage systems in order to minimize the access
to the grid and the consumers' bills. The main goal of the research developed in [70] is to use
renewable resources (i.e. wind) and resolve the problem of their intermittency. In addition, in
[70], the authors aim to optimize the consumer bill so they adopted a negotiation algorithm to
choose the combination of consumers and producers that minimizes the energy bill. In this
research, the authors proposed a multi-agent system composed of four types of agents shown in
Figure 11:
1. The grid agent, introduced in the smart grid, which role is to satisfy the energy lack and
buy the exceeded energy production.
2. The storage agent, introduced in a centralized battery, aims to control the energy storage.
3. The prosumer agent, introduced in renewable resources, controls the distribution of the
energy it produces.
4. The consumer agent, introduced in smart homes, negotiates the energy purchase with
other consumers and prosumers.
In this algorithm, the consumers negotiate the energy purchase from all resources (producers, grid
and storage system). Each consumer is selfish and chooses the combination that minimizes its
own bill and sends its proposal to the other consumers in its network. Finally, the consumers
choose the proposal that has the maximal utility function (i.e. the minimal cost). In this research,
the authors minimize consumers’ bill and use optimally all renewable energy production. The
next section highlights cloud computing use in the smart grids energy management.
Figure 51 Illustration of the proposed system of the researches presented in [69, 70]
4.3. ENERGY MANAGEMENT USING CLOUD COMPUTING ARCHITECTURE
Many authors used the cloud computing in the smart grid energy management. In this section we
will present some of the proposed algorithms.

40
Ghamkhari et al. [71] presented an optimization-based strategy for data centers. The authors
developed a mathematical model that aims to minimize the data center's energy.
The authors in [72] proposed a Building Energy Management algorithm called BEMS using smart
grids and cloud computing. The aim of this work is to design a management model by using the
cloud concept. The authors implement their algorithm in a building. In addition, the proposed
system consists of three components: the Power Monitoring Entity that aims to monitor and
control device's power consumption, the Environmental Information Entity that aims to collect
environmental information, and the System Manager that collects an manages data and provides
services to users. Furthermore, the authors used a cloud computing three-layered architecture.
Firstly, the service layer is composed of a service priority manager that is responsible for the
priority management of services. In addition, this layer allocates the resources to the high priority
demand. Furthermore, the service analyzer layer analyzes the requested service and determines
the required resources. Finally, the service profile database stores the information from services
already requested to response quickly without data analysis. Secondly, the management layer is
responsible for allocating resources, gathering data and categorizing the situational events for
database storage and managing the power and environmental information. Thirdly, the physical
layer aims to manage heterogeneous devices such as mobile phones, laptops, etc.
Table 1. Review of energy management algorithms characteristics
Algorithm Use
MAS
Use
Storage
device
Centralisa
tion
Use
predicted
data
Update
predicted
data
Use
negotiati
on
Alberola et al. [54] ✓ ✓ X ✓ X X
Mohsenian et al.
[58]
X ✓ X ✓ X X
Subramanian et al.
[59]
X X X ✓ X X
Prodan et al. [60] X ✓ X ✓ X X
Kanchev et al. [14] X ✓ X ✓ X X
Rose et al. [53] ✓ ✓ ✓ ✓ X X
Wang et al. [52] X ✓ X ✓ ✓ X
Cecati et al. [61] X X X ✓ X X
Hooshmand et al.
[62]
X ✓ X ✓ ✓ X
Sechilariu et al.
[63]
X ✓ X X X X
Guo et al. [16] X ✓ X ✓ X X
Roche et al. [64] ✓ ✓ X ✓ X X
Logenthiran et al.
[65]
✓ ✓ X X X X
Nagata et al. [42] ✓ X ✓ ✓ X ✓
Colson et al. [15] ✓ X X X X ✓
Nabouch et al. [66] ✓ X X X X ✓
Klaimi et al. [70] ✓ ✓ ✓ ✓ X ✓
Ngoc et al. [56, 57] ✓ X X ✓ X X
Vytelingum et al.
[55]
✓ ✓ X ✓ X X
Ramchurn et al.
[68]
✓ ✓ X ✓ X X

41
4.4. REVIEW OF THE DIFFERENT ALGORITHMS
Quantitative analysis of the summarized algorithms is presented in this section. The centralization
is the drawback of the algorithm presented in [53]. The research presented in [61] has a lack in
storage systems use. The algorithm in [62] solves the optimization problem, minimizes real-time
generation of energy and optimizes energies costs. Moreover, this research does not take into
consideration the sudden energy use. Furthermore, the algorithm described in [65] is used in order
to plan in real-time the produced energy and the storage system use. However, this algorithm does
not negotiate the energy purchase to minimize the consumers’ bills. Moreover, the research of
Nagata et al. [42] optimizes consumers' bills but the centralization and the lack of storage systems
can be considered as disadvantages of this research. In addition, Colson et al. [15] presents an
algorithm that uses negotiation to get the energy at the lowest price. This algorithm has a
drawback which is lack of the integration of storage system. Nabouch et al. [66] have not
integrated storage systems that may improve outcomes and reduce the consumers’ bills.
The Table 1 presents a review of all summarized energy management algorithms presented in this
paper.
5. CONCLUSION
The integration of the smart grids technologies will be an extremely important task in the near
future. The main target is to develop general and exceptionally flexible integration strategies for
the use of distributed energy resources and/or storage systems. Therefore, there is an evident need
to investigate the feasibility/efficiency of integrating different distributed energy storage systems
in combination with distributed energy resources and their influence on the penetration of
renewable energy. Moreover, it is clearly important to develop standardized control strategies for
energy management in the smart grid systems including the storage and renewable production
side.
In this work we have presented the need for energy management for efficient usage of electricity
in the smart grid. We presented the current electricity grids that absolutely must adapt to become
more intelligent, that is to say, integrate new information technologies and communication. ICTs
permit interactions among the computational and the physical elements of the smart grid.
Furthermore, ICTs will meet the challenges of integrating renewable power production by
providing energy management algorithms.
The massive integration of renewable energy resources in electricity networks is an issue that
aims to be solved. The main goal to reach is the management of the intermittent power into the
power grids. Finally, we presented also several algorithms of energy management using the
storage and/or multi-agent systems and many algorithms that use the cloud computing
architecture. This presentation has been concluded by a brief qualitative discussion of these
proposals.
ACKNOWLEDGEMENTS
This work is supported by grants from the Troyes University of Technology and the Lebanese
University and is partly supported by the Champagne-Ardenne region research project SOLOTEC
and the Lebanese-French research project CEDRE Number 32928XK.

42
REFERENCES
[1] V.C. Gungor et al. Smart grid technologies: Communication technologies and standards. Industrial
Informatics, IEEE Transactions on, 529-539, Nov 2011.
[2] C. Gao et al. A review of voltage control in smart grid and smart metering technologies on
distribution networks. In Universities' Power Engineering Conference (UPEC), Proceedings of 2011
46th International, pages 1-5, Sept 2011.
[3] J.O. Petinrin et al. Smart power grid: Technologies and applications. In Power and Energy (PECon),
2012 IEEE International Conference on, pages 892-897, Dec 2012.
[4] J. Ou, et al. Mapping global fossil fuel combustion co2 emissions at high resolution by integrating
nightlight, population density, and traffic network data. IEEE Journal of Selected Topics in Applied
Earth Observations and Remote Sensing, 1674-1684, 2016.
[5] R. Atia et al. Sizing and analysis of renewable energy and battery systems in residential microgrids.
IEEE Transactions on Smart Grid, 1-10, 2016.
[6] S. Gottwalt et al. Modeling and valuation of residential demand exibility for renewable energy
integration. IEEE Transactions on Smart Grid, PP(99):1-10, 2016.
[7] H. Chen et al. Key technologies for integration of multitype renewable energy sources 2014;research
on multi-timeframe robust schedulingdispatch. IEEE Transactions on Smart Grid, 471-480, 2016.
[8] Y. Yan et al. A survey on smart grid communication infrastructures: Motivations, requirements and
challenges. IEEE Communications Surveys Tutorials, 5-20, 2013.
[9] Y. Feng et al. Energy efficient self-sustaining wireless neighborhood area network design for smart
grid. Smart Grid, IEEE Transactions on, 220-229, Jan 2015.
[10] C.W. Potter et al. Building a smarter smart grid through better renewable energy information. In
Power Systems Conference and Exposition, 2009. PSCE '09. IEEE/PES, pages 1-5, March 2009.
[11] N. Matta et al. Pribacc: In-network sensor data processing for efficient smart grid monitoring
applications. In Global Information Infrastructure Symposium, 2013, pages 1-6, Oct 2013.
[12] N. Matta et al. A wireless sensor network for substation monitoring and control in the smart grid. In
Green Computing and Communications (GreenCom), 2012 IEEE International Conference on, pages
203-209, Nov 2012.
[13] K. Mets et al. Distributed multi-agent algorithm for residential energy management in smart grids. In
Network Operations and Management Symposium (NOMS), 2012 IEEE, pages 435-443, April 2012.
[14] H. Kanchev et al. Energy management and operational planning of a microgrid with a pv-based active
generator for smart grid applications. Industrial Electronics, IEEE Transactions on, 4583-4592, Oct
2011.
[15] C.M. Colson et al. Algorithms for distributed decision-making for multi-agent microgrid power
management. In Power and Energy Society General Meeting, 2011 IEEE, pages 1-8, July 2011.
[16] G. Yuanxiong et al. Optimal power management of residential customers in the smart grid. Parallel
and Distributed Systems, IEEE Transactions on, 23(9):1593-1606, Sept 2012.
[17] N. Javaid et al. A survey of home energy management systems in future smart grid communications.
In Broadband and Wireless Computing, Communication and Applications (BWCCA), 2013 Eighth
International Conference on, pages 459-464, Oct 2013.
[18] L. Yang et al. Cost-effective and privacy-preserving energy management for smart meters. IEEE
Transactions on Smart Grid, 6(1):486-495, 2015.
[19] W. Zhang et al. Distributed online optimal energy management or smart grids. IEEE Transactions on
Industrial Informatics, 717-727, 2015.
[20] R. Belkacemi et al. Experimental implementation of multi-agent system algorithm for distributed
restoration of a smart grid system. In SOUTHEASTCON 2014, IEEE, pages 1-4, March 2014.
[21] A. Sardouk et al. Agent strategy data gathering for long life wsn. International Journal of Computer
Networks & Communications, 71-90, 2010.
[22] C. M. Elliott.Distributed multi-agent systems-a literature survey and inquisitive discussion. In 2014
IEEE Aerospace Conference, pages 1-7, 2014.
[23] B. Bitzer et al. Cloud computing framework for smart grid applications. In Power Engineering
Conference (UPEC), 2013 48th International Universities', pages 1-5, Sept 2013.
[24] A.-H. Mohsenian-Rad et al. Coordination of cloud computing and smart power grids. In Smart Grid
Communications (SmartGridComm), 2010 First IEEE International Conference on, pages 368-372,
Oct 2010.

43
[25] S. Bera et al. Cloud computing applications for smart grid: A survey. Parallel and Distributed
Systems, IEEE Transactions on, PP(99):1-1, 2014.
[26] F. Bonomi et al. Fog computing and its role in the internet of things. In Proceedings of the First
Edition of the MCC Workshop on Mobile Cloud Computing, MCC '12, pages 13-16, 2012.
[27] I. Stojmenovic et al. The fog computing paradigm: Scenarios and security issues. In Computer
Science and Information Systems (FedCSIS), 2014 Federated Conference on, pages 1-8, Sept 2014.
[28] K. Hong et al. Mobile fog: A programming model for large-scale applications on the internet of
things. In Proceedings of the Second ACM SIGCOMM Workshop on Mobile Cloud Computing,
MCC '13, pages 15-20, 2013.
[29] S. Sofana et al. A smart survey on demand response potential in global energy market. Indian Journal
of Science and Technology, 2015.
[30] Y. W. Law et al. Demand response architectures and load management algorithms for energy efficient
power grids: A survey. In Knowledge, Information and Creativity Support Systems (KICSS), 2012
Seventh International Conference on, pages 134-141, 2012.
[31] J. S. Vardakas et al. A survey on demand response programs in smart grids: Pricing methods and
optimization algorithms. IEEE Communications Surveys Tutorials, 152-178, 2015.
[32] R. Deng et al. A survey on demand response in smart grids: Mathematical models and approaches.
IEEE Transactions on Industrial Informatics, 570-582, 2015.
[33] P. Faria et al. Demand response programs definition using demand price elasticity to define
consumers aggregation for an improved remuneration structure. In Innovative Smart Grid
Technologies Europe (ISGT EUROPE), 2013 4th IEEE/PES, pages 1-5, Oct 2013.
[34] S. Pierluigi. Demand response and smart grids--a survey. Renewable and Sustainable Energy
Reviews, 461 - 478, 2014.
[35] M.C. Falvo et al. Electric vehicles integration in demand response programs. In Power Electronics,
Electrical Drives, Automation and Motion (SPEEDAM), 2014 International Symposium on, pages
548-553, June 2014.
[36] M.H. Albadi et al. A summary of demand response in electricity markets. Electric Power Systems
Research, 1989 - 1996, 2008.
[37] Y. Ghiassi-Farrokhfal et al. Toward a realistic performance analysis of storage systems in smart grids.
Smart Grid, IEEE Transactions on, 402-410, Jan 2015.
[38] A. A. Jamali et al. Energy storage systems and their sizing techniques in power system-a review. In
2015 IEEE Conference on Energy Conversion (CENCON), pages 215-220, Oct 2015.
[39] A. Kawashima et al. Energy management systems based on real data and devices for apartment
buildings. In Industrial Electronics Society, IECON 2015 - 41st Annual Conference of the IEEE,
pages 3212-3217, 2015.
[40] M.H. Cintuglu et al. Real-time implementation of multiagent-based game theory reverse auction
model for microgrid market operation. Smart Grid, IEEE Transactions on, 1064-1072, March 2015.
[41] S.D.J. McArthur et al. Multi-agent systems for power engineering applications part i: Concepts,
approaches, and technical challenges. Power Systems, IEEE Transactions on, 1743-1752, Nov 2007.
[42] T. Nagata et al. A multi-agent approach to smart grid energy management. In IPEC, 2012 Conference
on Power Energy, pages 327-331, Dec 2012.
[43] G.H. Merabet et al. Applications of multi-agent systems in smart grids: A survey. In Multimedia
Computing and Systems (ICMCS), 2014 International Conference on, pages 1088-1094, April 2014.
[44] A. Sharma et al. A multi-agent-based power system hybrid dynamic state estimator. Intelligent
Systems, IEEE, 52-59, May 2015.
[45] R. Roche et al. Multi-agent systems for grid energy management: A short review. In IECON 2010-
36th Annual Conference on IEEE Industrial Electronics Society, 3341-3346, Nov 2010.
[46] T. Hai-Jun et al. An extended contract-net negotiation model based on task coalition and genetic
algorithm. In Machine Learning and Cybernetics, 2007 International conference on, volume 2, pages
879-884, Aug 2007.
[47] X. Dong et al. Multi-attribute negotiation model based on internal factors argumentation. In
Management Science and Engineering (ICMSE), 2013 International Conference on, pages 20-27, July
2013.
[48] G. Rohbogner et al. What the term agent stands for in the smart grid definition of agents and multi-
agent systems from an engineer's perspective. In Computer Science and Information Systems
(FedCSIS), 2012 Federated Conference on, pages 1301-1305, Sept 2012.

44
[49] C. Sivapragash et al. Advanced cloud computing in smart power grid. In Sustainable Energy and
Intelligent Systems (SEISCON 2012), IET Chennai 3rd International on, pages 1-6, Dec 2012.
[50] P. Toma et al. Sizing photovoltaic-wind smart microgrid with battery storage and grid connection. In
Automation, Quality and Testing, Robotics, 2014 IEEE International Conference on, pages 1-5, May
2014.
[51] Chen Changsong et al. Energy trading model for optimal microgrid scheduling based on genetic
algorithm. In Power Electronics and Motion Control Conference, 2009. IPEMC '09. IEEE 6th
International, pages 2136-2139, 2009.
[52] W. Yanzhi et al. Adaptive control for energy storage systems in households with photovoltaic
modules. Smart Grid, IEEE Transactions on, 5(2):992-1001, March 2014.
[53] H. Rose et al. A scoring rule-based mechanism for aggregate demand prediction in the smart grid. In
Proceedings of the 11th International Conference on Autonomous Agents and Multiagent Systems -
Volume 2, AAMAS '12, pages 661-668, 2012.
[54] J. M. Alberola et al. A self-configurable agent-based system for intelligent storage in smart grid.
volume 365 of Communications in Computer and Information Science, pages 240-250. 2013.
[55] P. Vytelingum et al. Theoretical and practical foundations of large-scale agent-based micro-storage in
the smart grid. J. Artif. Int. Res., 765-813, 2011.
[56] Ngoc Cuong Truong et al. Towards appliance usage prediction for home energy management. In
Proceedings of the Fourth International conference on Future Energy Systems, e-Energy '13, pages
287-288, 2013.
[57] Ngoc Cuong et al. Forecasting multi-appliance usage for smart home energy management. In 23rd
International Joint Conference on Artificial Intelligence (IJCAI 2013), April 2013.
[58] A.-H. Mohsenian-Rad et al. Optimal residential load control with price prediction in real-time
electricity pricing environments. Smart Grid, IEEE Transactions on, 120-133, Sept 2010.
[59] S. Subramanian et al. Dynamic price optimization models for managing time-of-day electricity usage.
In Smart Grid Communications (SmartGridComm), 2013 IEEE International Conference on, pages
163-168, Oct 2013.
[60] P. Ionela et al. A model predictive control framework for reliable microgrid energy management.
International Journal of Electrical Power & Energy Systems, 399 - 409, 2014.
[61] C. Cecati et al. Optimal operation of smart grids with demand side management. In Industrial
Technology (ICIT), 2013 IEEE International Conference on, pages 2010-2015, Feb 2013.
[62] A. Hooshmand et al. A novel cost-aware multi-objective energy management method for microgrids.
In Innovative Smart Grid Technologies (ISGT), 2013 IEEE PES, pages 1-6, Feb 2013.
[63] S. Manuela et al. Building-integrated microgrid: Advanced local energy management for forthcoming
smart power grid communication. Energy and Buildings, 236 - 243, 2013.
[64] R. Roche et al. Algorithmes hybrides pour la gestion intelligente de l'énergie dans les smart grids. In
Journées Francophones sur la planification, la décision et l'apprentissage pour le contrôle des
systèmes - JFPDA 2012, page 14 p, May 2012.
[65] T. Logenthiran et al. Multi-agent system for market based microgrid operation in smart grid
environment. Dept. of Elec. & Comp. Science, National University of Singapore, 2011.
[66] D. El Nabouch et al. An agent-based approach for efficient energy management in the context of
smart houses. In Highlights on Practical Applications of Agents and Multi-Agent Systems, volume
365, pages 375-386, 2013.
[67] P. Vytelingum et al. Agent-based micro-storage management for the smart grid. May 2010.
[68] S. Ramchurn et al. Agent-based homeostatic control for green energy in the smart grid. ACM
Transactions on Intelligent Systems and Technology, 35:1-35:28, July 2011.
[69] J. Klaimi et al. An energy management approach for smart-grids using intelligent storage systems. In
Digital Information and Communication Technology and its Applications (DICTAP), 2015 Fifth
International Conference on, pages 26-31, April 2015.
[70] J. Klaimi et al. An agent-based approach for energy management in smart-grids. In Highlights of
Practical Applications of Agents, Multi-Agent Systems, and Sustainability-The PAAMS Collection,
pages 225-236. 2015.
[71] M. Ghamkhari et al. Energy and performance management of green data centers: A profit
maximization approach. Smart Grid, IEEE Transactions on, 1017-1025, June 2013.

45
[72] H. Insung et al. Cloud computing-based building energy management system with zigbee sensor
network. In Innovative Mobile and Internet Services in Ubiquitous Computing (IMIS), 2012 Sixth
International Conference on, pages 547-551, July 2012.
[73] M. E. El-hawary. The smart grid: state-of-the-art and future trends. Electric Power Components and
Systems, 239-250, 2014.
Authors
Joelle Klaimi received her Telecommunication Engineer degree from the Lebanese
University- First Branch, Tripoli, Lebanon, in 2012. She also received her M2R in Medical
Technology and Industrial Systems (TSMI) in 2012 from the Lebanese University,
Doctoral school of sciences and technology. Currently, she is a PhD student at the
Lebanese university, Lebanon and Troyes University of Technology, France. Her research
interests include distributed systems, agent cooperation and energy efficiency.
Rana RAHIM-AMOUD received with distinction her Computer science and
Telecommunication Engineer degree from the Lebanese University - First Branch, Tripoli,
Lebanon, in 2002. She then obtained her Master degree (DEA) in Networking and
Telecommunication in 2003 from the USJ University and the Lebanese University, Beirut,
Lebanon, and her PhD degree in January 2008 from the University of Technology of
Troyes (UTT) - France. She was a postdoctoral researcher at the UTT from October 2008 to October 2009.
She obtained her HDR (Habilitation à Diriger des Recherches) in 2016. She is currently Assistant professor
at the Lebanese University. Her research interests include Multi-agent Systems, System management,
Quality of Service, WSN Networks, IoT Networks, Smart Grids. She is author of more than 30 papers
published in leading journals and conference proceedings. She is the General Chair of the New Generations
of Networks Workshop (NGN). She is also TPC member for many international conferences and has served
as a reviewer for a number of international journals and conferences.
Dr. Leïla Merghem-Boulahia received an engineering degree in computer science from
the University of Sétif, Algeria, in 1998, an M.S. degree in artificial intelligence and a
Ph.D. in computer science from the University of Paris 6, France, in 2000 and 2003,
respectively. She received the “Habilitation à diriger des recherches” degree in Computer
Science from the University of Compiègne in 2010. She is an associate professor at the
University of Technology of Troyes (UTT) in France still 2005. Her main research topics
include multi-agent systems, quality of service management, autonomic networks, cognitive and sensor
networks and smart grids. Dr. Merghem-Boulahia authored or co-authored more than 80 international
journals and conference papers. She received the best paper award of the IFIP WMNC’2009 and
GIIS’2013. She also acted as a TPC member of many conferences and workshops (IEEE Globecom, IEEE
ICC, IEEE MASS…) and has served as a reviewer for internationally well-known journals (IEEE
Communications letters, Communication Networks, International Journal of Network Management, Annals
of Telecommunications).
Akil Jrad was born in Kalamoun, Lebanon, on September 4, 1970. He received the M.
Sc. degree in physics from the Lebanese University, Tripoli, Lebanon in 1994, the Ph.D.
degree in electronics from the University of Savoy, France, in 1999. From 1999 to 2001,
he was with the Laboratory of Microwaves and Characterization, University of Savoy,
France, where he was involved with the conception and realization of NLTL’s for the
generation and measurement of ultra-fast microwave signals and frequency multipliers. Currently he is a
professor at the Lebanese University. He is also member of the laboratory of the Electronic Systems,
Telecommunications and Networks (LaSTRe), EDST, Lebanese University, where his current research
interests include nonlinear microwave and millimeter-wave circuit analysis and design. He is author of
more than 50 papers published in leading journals and conference proceedings. He is TPC member for
many international conferences and has served as a reviewer for a number of international journals and
conferences.

ENERGY MANAGEMENT ALGORITHMS IN SMART GRIDS: STATE OF THE ART AND EMERGING TRENDS

More Related Content

What's hot (20)

Similar to ENERGY MANAGEMENT ALGORITHMS IN SMART GRIDS: STATE OF THE ART AND EMERGING TRENDS (20)

Recently uploaded (20)

ENERGY MANAGEMENT ALGORITHMS IN SMART GRIDS: STATE OF THE ART AND EMERGING TRENDS