Optimizing of Consumption Energy in Smart building

NEDIOUI MOHAMED ABDELHAMID1, BRAHIM LEJDEL1, ELISEO CLEMENTINI2

1Department of Computer science, Faculty of Exact Sciences, University of El-Oued, El-Oued,

ALGERIA

2University of L’Aquila, L’Aquila,

ITALY

Abstract- Intelligent technologies have advanced significantly over the past two decades and have integrated

with cities to enhance citizen lives. Additionally, the amount of energy consumed varies depending on the

weather, the number of occupants, and the type of building—commercial, residential, or administrative. In

contrast, the citizen must make a trade-off between the building's environmental impact, comfort levels, and

energy use.

In this essay, we'll suggest a smart model that enables management, control, and regulation of energy usage in

accordance with a set of standards. As a result, this approach enables real-time calculation, regulation, and

optimization of energy usage as well as comfort for the occupants. As a result, the person can learn about their

energy consumption without having to read electricity measurements or wait for a billing cycle. Additionally,

this method enables energy resource conservation and increases system output even during periods of high

demand.

Keywords- Smart Energy System, Energy consumption, multi-agent system, genetic algorithm.

Received: July 15, 2022. Revised: September 2, 2022. Accepted: September 24, 2022. Published: October 14, 2022.

1 Introduction

Enhancing citizen quality of life is the goal of smart

cities (SC). It has been rising in relevance on

policymakers' agendas, [1]. As a result, the City is

an urban region made up of a variety of

interconnected elements such as the population,

various networks such as the electricity or water

network, buildings, etc. Actually, the smart city is

built to make the best use of resources like energy,

water, and internet. We can install sensors and

lighting that can collect and deliver information to

reduce energy consumption. In order to create a

smart city, the streets must be connected. The

brains of smart cities are these electronically

interconnected streetlights. There are

structures throughout the city that have off-peak

times. For instance, if a residential building houses

students who are in class or studying throughout

the day, the demand for energy will be reduced. But

in the evening, things are different. Even if they use

a lot of energy throughout the day in commercial or

administrative buildings, they use less in the

evening. Additionally, we can detect a peak off

hour and low usage in the same building. In order

to manage, regulate, and optimize the energy usage

of all these different types of buildings, we must

develop a smart model. In order to control and

manage energy usage while preserving occupant

comfort, smart buildings need an interior

environment control system, [2].

In this study, we suggest using a multi-agent

system that allows for the distribution of various

duties among the agents while allowing for real-

time energy consumption optimization using

genetic algorithms by each agent, allowing for

quick adaptation to building consumption.

As a result, we will employ some agents to act as

the building's meter. Additionally, we assign a

resource agent—a type of agent that can interact

with other actors or meter agents—to each energy

resource that supplies energy to streets in order to

control and optimize energy use. In order to

identify the optimum solution that can control

energy consumption for all types of buildings, all

agents can work together and engage in

negotiation. Then, we create a GIS system that

enables us to know the location of a building and

any information related to it, such as the building's

ID, energy usage, billing information, etc.

Following a thorough investigation of the

matter, we discover that there are three variables

that can impact energy consumption, including

system peak times, energy prices, peak energy

demand, and the amount of energy consumed each

hour, among other things. Additionally, there are

two types of energy-consuming equipment that we

commonly utilize, including lighting systems and

WSEAS TRANSACTIONS on SYSTEMS and CONTROL

DOI: 10.37394/23203.2022.17.47

Nedioui Mohamed Abdelhamid,

Brahim Lejdel, Eliseo Clementini

E-ISSN: 2224-2856

428

Volume 17, 2022

HVAC (Heating, Ventilation, and Air

Conditioning) systems.

We have seen significant advancements in

intelligent technology over the past two decades,

which have fused in cities to enhance the quality of

life for citizens. With the aid of these intelligent

building technologies, tenants' lives can be made

easier. Generally speaking, smart buildings are

anticipated to use cutting-edge computing systems

and intelligent technology to accomplish the best

possible trade-offs between total occupant comfort

and energy consumption, [3]. In this context, we

can point out that the primary energy consumers in

residential, office, and commercial buildings are

HVAC and lighting systems. Demand-driven

control measures like turning off or dimming smart

lighting systems, controlling ventilation, and

regulating the amount of heating and cooling

provided to buildings using actual building

occupancy data all contribute to improving the

energy performance in buildings, [4]. Up to 60% of

the energy for buildings is used by these systems.

Depending on how well the structure functions, the

remaining energy is used by various types of

equipment, [5]. Because they can handle distributed

and adaptive conditions, multi-agent systems are

the best method for modeling complex systems. In

order to control environmental factors and resolve

potential conflicts that could arise between energy

consumption and customer comfort, we will apply

this approach in this study to manage, regulate, and

optimize the energy consumption in residential

buildings.

A multi-agent system has been primarily used

to coordinate the use of building electrical devices

and heating as an HVAC-L system in order to

reduce energy demand in a smart building that

enables to optimize the energy consumption and

energy cost in order to increase the comfort level of

building occupants. In this section, we evaluate the

current state of the art for using a multi-agent

system to balance energy usage with the

satisfaction of building inhabitants.

In order to reduce building energy consumption

and maintain occupant comfort, Hagras et al.

suggest using a system to learn how a building

reacts thermally to both interior and external

occupancy loads, [6].

Following that, Liao and Barooah create a

Multi-Agent System to simulate the actions of

every building tenant and produce reduced-order

graphical representations from simulations of the

agent-based model, [7].

A Multi-Agent System was also created by

Joumaa et al. to manage proactive and reactive

control of building HVAC and lighting systems.

They choose the agent-based methodology because

distributed systems can simulate building energy

usage. These systems rely on HVAC-L system and

sensor interaction with facility systems and

appliances,[8].

We might also mention the work of Azar and

Menassa, who create a novel agent-based

methodology that enables us to model the energy

consumption of commercial buildings. This

simulation takes into account different types of

occupants and potential changes to occupant

behavior as a result of their interactions with the

built environment and one another, [9].

All of these prior multi-agent systems look for

ways to better manage building systems and energy

resources while also reducing building energy

consumption through direct interaction and

coordination with building occupants. However,

there are no methods that have been created to

maximize energy consumption and meet occupant

comfort requirements. In this article, we'll

introduce a hybrid technique called the multi-agent

system and genetic agent (SMA-GA), which

enables each agent to discover the best strategy for

maximizing energy efficiency and passenger

comfort.

This paper is organized as the following. First, we

outline our suggested strategy, which is based on

the Multi-Agent System and Genetic Agent

methods (SMA-GA). Then, we include a summary

and any recommendations for follow-up work.

2 Materials and Methods

2.1 Smart Building

The idea of a smart building is actually first

introduced in intelligent systems that can regulate

energy usage and occupant comfort. In order to

maximize energy efficiency, occupant comfort,

resource savings, and system overall productivity,

smart buildings manage, regulate, and control their

indoor environment using intelligent technologies.

By utilizing current technologies, these systems

seek to reduce energy consumption, conserve

resources, and meet occupant comfort demands.

HVAC-L systems and other intelligent energy

appliances help reduce energy use. Saving

resources entails protecting them from threats. By

adding HVAC-L values that adjust the building

environment in accordance with occupant

preferences, demand from occupants is satisfied.

On the other hand, when a building is intelligently

controlled to accommodate tenant preferences by

WSEAS TRANSACTIONS on SYSTEMS and CONTROL

DOI: 10.37394/23203.2022.17.47

Nedioui Mohamed Abdelhamid,

Brahim Lejdel, Eliseo Clementini

E-ISSN: 2224-2856

429

Volume 17, 2022

modifying the lighting and heating/cooling level,

these preferences must be carefully evaluated and

discovered through occupant input or behaviors.

The corresponding agent must be able to detect and

learn from behaviors in order to use this skill.

Through learning the activities of the agents and

the corresponding responses of the inhabitants, the

process is carried out interactively based on the

reinforcement mechanism.

To meet the needs of energy efficiency and

occupant satisfaction, a variety of activities must be

finished for minimizing energy consumption in

buildings and evaluating occupant comfort in

reaction to changes in the building environment.

However, the relationship between energy use and

occupant comfort is typically inverse. Therefore,

resolving the conflicts between energy usage and

occupant comfort is one of a smart building's main

objectives. Typically, environmental criteria used

to assess occupant comfort levels include indoor

temperature, ventilation, air conditioning, and

lighting level. In addition to preserving the use of

devices against hazards, the decision of the control

strategy is crucial to lowering energy consumption

in the building. The system efficiency is also

influenced by the energy consumption. Application

of an intelligent controller that attempts to reduce

energy usage without lowering occupant comfort

can accomplish these goals. A smart system-

equipped home is shown in Figure 1.

Fig. 1: A smart building.

2.2 Multi-agent System and Genetic

Algorithm

2.2.1 Why We Use the Multi-Agent System in

Smart Cities

An agent is a software system that is embedded in a

certain environment and has the ability to act

autonomously in order to achieve its intended

goals, [10]. The multi-agent system is utilized in

this study because it offers a number of benefits in

the areas of energy consumption and occupant

comfort. The advantages of multi-agent systems

that provide autonomy to address energy

consumption issues and meet occupant comfort

demands. It thus offers a structure that is

appropriate for these systems. Additionally, they

offer a number of crucial traits like mobility,

flexibility, cooperation, and bargaining. In other

words, this autonomous agent senses its

surroundings on the one hand while changing it on

the other hand by its behaviors. As a result, an

agent can quickly adapt to a changing environment.

2.2.2 Why We Use the Genetic Agent

In [11], genetic algorithms are created to mimic the

phenomenon of adaptation in live things. They are

an optimization method based on the ideas of

genetics and natural selection. Within a fair amount

of time, it examines a vast number of potential

solutions for the best one (the process of evolution

takes place in parallel). Each of these solutions has

a set of specifications that fully characterize the

solution. Then, with each parameter consisting of

one or more "chromosomes," this collection of

parameters can be thought of as the "genome" of

the person. They enable a population of solutions to

gradually converge on the best option. They will

accomplish this via a system for selecting from the

population of people (potential solutions). The

chosen individuals will be crossed with one another

(crossover), and some of them will mutate by

avoiding local optima as much as they can. Both

issues are primarily handled by genetic algorithms,

[12].

1. There are several parameters that need to be

optimized at once or the search space is huge.

2. A precise mathematical model cannot

easily capture the nature of the issue.

We connect Multi-Agent Systems with Genetic

Algorithms to allow the agent to freely choose the

optimum course of action (SMA-GA). As a

consequence, our proposed model is based on the

three concepts listed below.

1. A building agent is a software agent that has

the ability to manage, control, and exchange

pertinent data with nearby agents.

2. The genetic algorithm uses genetic

patrimony that has been altered across agents

as an input. This genetic heritage represents

data on the HVAC-L system's energy

WSEAS TRANSACTIONS on SYSTEMS and CONTROL

DOI: 10.37394/23203.2022.17.47

Nedioui Mohamed Abdelhamid,

Brahim Lejdel, Eliseo Clementini

E-ISSN: 2224-2856

430

Volume 17, 2022

consumption that was gathered by the sensor

system.

3. To discover the best configuration for the

current situation, a genetic algorithm is

applied. This algorithm considers two goal

functions: energy consumption and

occupant comfort.

2.3 Building Agent

As we previously mentioned, a smart building has

the ability to manage, control, and regulate its

indoor environment to optimize energy

consumption, ensure the comfort of the occupants,

save energy resources, and boost system

productivity. Therefore, the primary goal of a smart

building is to resolve any conflicts that may arise

between satisfying the comfort needs of its

inhabitants and consuming less energy. Both the

environmental conditions and the occupant's

preferences for the surroundings have an impact on

their level of comfort satisfaction. Environmental

factors may be used as indices to construct the

function of occupant comfort by taking into

account both the actual value of the factors and the

preferences of the occupants to determine how

comfortable the building's occupants are. As a

result, three components have been included in the

design of the building agent to control and regulate

its indoor climate and reduce energy usage. An

optimizer, a simulator, and a comfort model make

up these elements.

2.3.1 An Optimizer

A genetic algorithm is used. In this study, GA runs

100 times in each time step to maximize the

likelihood of reaching the global optimization, save

energy resources, and satisfy occupant preferences

because heuristic techniques cannot be guaranteed

to identify the global optimal solution within the

finite number of iterations. The possibility of

getting better results from running the optimization

method more often should theoretically increase,

but this will certainly require more processing time.

Following numerous experiments, it was shown

that 100 runs is a practical quantity for balancing

the solution quality and computing time cost.

2.3.2 Simulator

Each building agent has a simulator that is utilized

in tandem to determine the inhabitants' level of

comfort and their optimal energy usage under the

current circumstances. To achieve a suitable

balance between discovery time and system

performance, the simulator's output could be

adjusted. The optimizer estimates the satisfaction of

occupants' comfort by repeatedly running energy

flow simulations for each iteration. The best

occupant comfort level is then used to develop the

following generation of general occupant comfort

levels, and this process is repeated over a number

of generations to obtain the optimum candidate

comfort level.

2.3.3 Comfort Model

The comfort model enables computer-based indoor

climate control in smart buildings to optimize

energy use, ensure occupant comfort, conserve

energy resources, and boost system productivity.

Building agents must assess energy usage and

occupant comfort levels in reaction to changes in

the indoor environment in order to reach a balance

between energy efficiency and occupant comfort.

However, the relationship between energy use and

occupant comfort is typically inverse, [13]. An

agent building's primary objective is to resolve

conflicts between lowering energy usage and

raising occupant comfort. The degree of occupant

comfort is influenced by the environment's state as

well as occupant preferences. The actual value of

the relevant environmental characteristics and the

occupants' preferences of these parameters may be

used as indices to develop the function of

occupant's comfort in order to evaluate the

occupants' comfort in their living environment. The

indoor temperature, light intensity, air conditioning,

and ventilation levels within the building are

typically utilized as measures to assess occupant

comfort.

2.4 Building Agent's Optimizer

As we already said, the building agent includes an

optimizer and simulator that work together to

determine the HVAC-L system parameters in order

to maximize energy usage while still satisfying

occupant comfort requirements. The use of genetic

algorithms has a significant advantage over systems

that rely on predefined values because each

building agent enables a genetic algorithm to

discover the values of the HVAC-L system, which

may not resemble any predefined values but may

be appropriate values for the current interior

conditions. The optimizers should strike a good

balance between the time it takes to find solutions

and the amount of energy used. Thus, each building

agent uses a genetic algorithm to identify the best

HVAC-L system values that may be attributed to

each system in order to maximize energy efficiency

and occupant comfort.

WSEAS TRANSACTIONS on SYSTEMS and CONTROL

DOI: 10.37394/23203.2022.17.47

Nedioui Mohamed Abdelhamid,

Brahim Lejdel, Eliseo Clementini

E-ISSN: 2224-2856

431

Volume 17, 2022

2.4.1 Chromosome Structure

We must define the genes and the chromosome

structure in order to use the genetic algorithm. The

gene's identification and a set of HVAC-L system

parameters that can be used to perform the best

energy consumption and satisfy the comfort level

of the occupants can be used to describe the gene.

We code the genes using various forms. In order to

code the identifiers, we first utilize strings. Then,

we use real numbers to code the temperatures,

ventilation, air conditioning, and lighting system

values. The gene's structure is shown in Figure 2.

Fig. 2: Gene structure of the room.

A sequence of genes from various room agents

that can provide the best response to the HVAC-L

system values are represented by the chromosome.

The optimizer use a genetic algorithm with its

various traditional processes, like selection,

crossover, and mutation, to determine the optimal

chromosome from the population. The optimizer

repeatedly runs the energy consumption simulator

for each HVAC-L system in a particular

generation, allowing each building agent to select

the optimal option from the population. The values

discovered by their optimizers are used as sub-

values to provide the ultimate solution of the

present energy consumption of the HVAC-L

system after interaction and negotiation between

the various building agents. The HVAC-L system's

top candidate values are found after a number of

generations. Figure 3 shows an example of a

solution that is found by the building Agent.

Fig. 3: The chromosome of the solution.

2.4.2 Genetic Algorithm Steps

a) Initialization

Each chromosome's initialization for participation

in the genetic algorithm's population is controlled

by the initialization operator. The genetic material

from which all novel solutions will arise is present

in this chromosome. In this work, the Steady State

will be used to start the generation process and

choose the genetic algorithm's population for the

following generation. First, Steady State clones the

initial chromosomes to produce a population of

individuals. Subsequently, during the course of

evolution, a temporary population of individuals is

produced, added to the previous population, and

then the worst individuals are eliminated in order to

bring the present population back to its original

size. According to this technique, the newly

produced offspring may or may not stay within the

new population, depending on how they compare to

the current population members.

b) Crossover

The process for creating a child from two parent

chromosomes is defined by the crossover operator.

The crossover operator creates new people as

offspring who share traits from both parents. The

likelihood of crossover influences the frequency of

crossover at each generation. The single point

crossover strategy will be used in this method for

all tests. At each generation, 50% of the new

population was created by joining two pieces of

each chromosome's parents to create a new

chromosome, which is how the data for all

experiments described in this paper were formed.

The crossover operator is depicted in Figure 4.

Fig. 4: Crossover operator.

c) Mutation

The mutation operator plays a crucial role. The

process for modifying the chromosome is

described. When a child experiences mutation, a

gene may be randomly altered with a low

likelihood. It offers a modest degree of random

search that makes convergence at the world's best

WSEAS TRANSACTIONS on SYSTEMS and CONTROL

DOI: 10.37394/23203.2022.17.47

Nedioui Mohamed Abdelhamid,

Brahim Lejdel, Eliseo Clementini

E-ISSN: 2224-2856

432

Volume 17, 2022

possible easier. The amount of each genome's

genetic material that is altered or mutated depends

on the chance of mutation. A portion of the

chromosome is altered if the mutation is carried

out. The search effort would be harmed if the

mutation happened too frequently. The results in

this study were produced using a 1% mutation

probability that was established experimentally

using a single vector HVAC-L system case. In

picture 5, we show a random mutation.

Fig. 5: Mutation operator.

d) Evaluation of solutions

We can argue that the objective function of every

discrete optimization problem, whose goal is to

deliver a metric for any given solution that

expresses its relative quality, is what determines if

the problem is successful. The goal function

employed in this method for solving the energy

consumption issue in buildings calculates and adds

the penalties related to temperature, illumination,

indoor air quality, and ventilation inside our state

representation. As a result, we will utilize objective

functions to assess potential energy-saving

measures and look at the weighted link between

actual measured values for the following

parameters: temperature, ventilation, indoor air

quality, lighting level, and occupant comfort level.

Several definitions that model the underlying

structure of the problem are necessary for the

objective functions used to evaluate solutions,

specifically:

–

 

RRRRR ,........,, 321



is the set of all

room in the building,

–

 

HHHHH ,........,, 321



is the set of

all heating systems in the building,

–

 

LLLLL ,........,, 321



is the set of all

illumination systems in the building,

–

 

AAAAA ,........,, 321



is the set of all

air conditions in the building,

–

 

VVVVV ,........,, 321



is the set of all

ventilation system in the building,

– Hm, Lm, Am, and Vm are the measured values

of the temperature, the illumination, and

the indoor air quality and ventilation

respectively.

– Hc, Lc, Ac, and Vc are the comfort values of

the temperature, the illumination, and the

indoor air quality, respectively.

– N1, N2, N3, N4 is the all number of the

temperature, the illumination, and the

indoor air quality and ventilation system

respectively.

– [Tmin, Tmax] represent the interval time

where the values of the three parameters

were measured.

– [Cmin, Cmax] represent the comfort range.

This range can be defined by customers.

– [Emin, Emax] represent the consumption

energy range.

Two essential components of our SMA-GA are

the allotted energy to the HVAC system EH and the

assigned energy to the lighting system EL.

In this situation, the two significant functions f(C)

and f(E) allow us to assess the effectiveness and

performance of the suggested strategy. Building

agent calculates these two functions.

In order to assess the effectiveness and

efficiency of our system, the goals of this

optimization mechanism are to optimize occupant

comfort (C) and reduce energy usage (E). First,

there is

HCCCCf ***)( 321 

(1)

The user-defined weighting variables, C1, C2, and

C3, highlight the significance of three comfort

factors and address any equipment conflicts. These

variables accept values between [0, 1]. Depending

on the situation and the time of year, occupants can

choose their own desired values. As we previously

stated, the design of the control method should take

WSEAS TRANSACTIONS on SYSTEMS and CONTROL

DOI: 10.37394/23203.2022.17.47

Nedioui Mohamed Abdelhamid,

Brahim Lejdel, Eliseo Clementini

E-ISSN: 2224-2856

433

Volume 17, 2022

occupancy duration into consideration because it

has a significant impact on energy savings. The

building agent activates the optimizer during hours

of occupancy in order to adjust the set point and

achieve the desired indoor visual comfort while

using the least amount of energy.

In the absence of occupants, the agent building

maintains the blind position and shuts off all

resource lights to conserve energy. The objective

function is described in equation 1 and its

maximization is the aim of optimization. The

control goal is accomplished in part via the comfort

value ratio, which is selected by the occupants and

is displayed on a graphic user interface. As a result,

it enables both an increase in occupant comfort and

energy efficiency.

The building's inside environment can be

controlled with the help of the second goal

function. The goal of this function is to reduce the

HVAC-L system's overall energy usage. Thus, the

following is a definition of this objective function.

LHVAC EEEf )(

(2)

EHVAC and EL represent the consumption energy of

the HVAC system and the lighting system,

respectively.

2.5 System Architecture

Our system primarily uses three different sorts of

agents: profile agents, room agents, and building

agents. These agents are capable of working

together in a building setting. The profile agent gets

its data from the inhabitants who can describe their

level of comfort, whereas the room agents get their

data from sensors. Throughout the entire structure,

sensors are dispersed to track the effectiveness and

operation of various technologies that have been

installed. The sensor network can provide three

different types of data: environmental data,

occupancy data, and energy data. Environmental

data refers to the environmental characteristics of

the building, such as the temperature inside and

outside, the amount of light, the air conditioning,

and the ventilation. The number of occupants, their

presence or absence, and their preferences for

occupancy are all common components of

occupancy statistics. Energy data primarily focuses

on the status of energy supplied, including the price

of electricity and the accessibility of other energy

resources. Different room agents will use these

measured data to determine how their respective

occupants behave. The many systems that are

installed in the room can be controlled by the room

agent, who can also adjust the gas, lighting, and

HVAC systems' settings, among others. Our SMA-

system GA's design is depicted in Figure 6.

Fig. 6: The architecture of our Multi-agent system.

2.6 Negotiation and Cooperation

The multi-agent system and genetic algorithm

(SMA-GA) requires us to present an effective

process of negotiation, cooperation, and

coordination amongst various agents in order to

simulate an ideal energy consumption. We are

aware that a single agent is unable to complete

some complex tasks, such as solving the energy

problem, due to individual limitations or because,

even though it can, its performance and efficiency

are drastically inferior to those achieved through

the cooperation and coordination of many agents,

[14]. The room agents bargain, each seeking to

collect enough energy needed to discover the

optimal ratio between measured values of

temperature, illumination, ventilation, and air

quality in the indoor environment and the comfort

values of the occupants, in order to resolve

consumption energy conflict. Therefore, it's crucial

to keep their impacts in check when conflicts arise

between these many parameters. In this situation,

negotiating strategies help the interested room

agents settle their disputes by finding concessions

between the residents' degree of comfort and the

building's energy consumption. The room agents

can resolve multiple problems at once and avoid

the emergence of new ones thanks to this

negotiation. Additionally, it might be required if

there is a quarrel between two or more building

agents who are from separate houses. When there

are disputes between buildings over who has the

WSEAS TRANSACTIONS on SYSTEMS and CONTROL

DOI: 10.37394/23203.2022.17.47

Nedioui Mohamed Abdelhamid,

Brahim Lejdel, Eliseo Clementini

E-ISSN: 2224-2856

434

Volume 17, 2022

right amount of energy to keep their residents

comfortable, we also employ negotiation to solve

the problem of energy usage. These residential

building agents bargain with one another to

determine the best plan that may be used to satisfy

the building's need for energy while also satisfying

the comfort level of the residents. As a result, under

our suggested method, the building agents bargain

with one another to get to the best agreement.

Through negotiation, they are able to resolve

several conflicts at once and prevent the emergence

of new ones. Figure 7 shows an illustration of a

negotiation between two building agents, Building

Agents 1 and 2. In order to reach an agreement,

these two agents propose a course of action that can

be built around measurable and occupant comfort

values. Each time a cycle occurs, one building

agent proposes a plan to the other building agents,

who might accept or reject the idea. Negotiations

come to a close if they agree; if not, the other

building agent makes a proposal during the

subsequent round.

Fig. 7: Negotiation between two agents.

Coordination between building or room agents to

determine the best course of action for every

system to resolve energy conflicts is another

definition of cooperation. In this essay, we discuss

using collaboration to resolve conflicts that arise

between energy use and occupant comfort. Each

room agent works together with the other room

agents to discover the best solution that enables to

decrease energy consumption and raise the degree

of comfort for the inhabitants. To enable the

building agent to analyze this data and run its

evolutionary algorithm, each room agent gives the

building agent some HVAC-L data. When the

HVAC-L system's optimizer has found a sequence

of values, each building agent communicates the

values to the other agents in a series of requests.

The next step is for each agent to review the

requests on its list, handle them, and attempt to

determine the HVAC-L system's final optimal

sequence of values that will allow it to resolve the

conflict and prevent future conflicts. Depending on

their existing circumstances, the building agents

may accept or reject the request of other agents.

Thus, the building agent requests a solution from its

neighbors and waits to hear back from them. After

analyzing their responses, the building agent

decides whether or not a solution is feasible. If a

solution is workable, it notifies its neighbors who

agreed to the solution that it is viable. A

straightforward arrangement of agent collaboration

is shown in Figure 8.

Fig. 8: Cooperation between Agents.

The signals sent and received by agents during the

resolution of disputes between room agents and

building agents in order to determine the best

energy consumption. Table 1 provides a summary

of these messages.

Table 1. Messages exchanged between agents.

Message

Comments

Inform (Distance,

energy, time)

The building agent

receives a message

from the room agent.

It has the necessary

quantity of energy.

Inform

(Conflict_Resolution)

Message delivered by

the agent to a

neighbor with whom

it has a disagreement.

WSEAS TRANSACTIONS on SYSTEMS and CONTROL

DOI: 10.37394/23203.2022.17.47

Nedioui Mohamed Abdelhamid,

Brahim Lejdel, Eliseo Clementini

E-ISSN: 2224-2856

435

Volume 17, 2022

The message includes

the outcome of the

agent's use of its

dispute resolution

process.

Notify (Conflict)

Message delivered by

the building's agent to

one of its neighbors

informing them of the

emergence of a fresh

dispute.

Notify (Satisfaction)

Message delivered by

a room agent to a

building agent

informing them of the

completion of the

procedure.

Help (Demand)

Message from a room

agent to a building

agent asking if there

is any way they can

assist.

Help (Response)

Message issued by a

building agent to a

station agent in a

room asking if they

can assist.

Negotiate(Demand)

Message sent to

another building

agent by one asking

the other whether a

negotiation is

possible.

Negotiate(Response)

Message sent to

another building

agent by one asking

whether it is possible

to negotiate with the

other agency.

Inform(Confirmation)

Agent transmits a

message It notifies its

neighbors who agreed

to the solution that it

is workable by

sending them a

confirmation.

3 Results and Discussion

In this section, we give a case study that

demonstrates how to create the various system

agents and demonstrate inter-agent cooperation. To

implement the various agents, including the

building agent, profile agent, and room agent, we

use Jade (http://jade.tilab.com/). Additionally, we

use Java (https://www.java.com/fr/) to develop the

evaluation function, crossover operator, and

mutation operator among other elements of the

genetic algorithm. A comfortable atmosphere for

all building residents is the goal of the smart

building, which is a residential structure. The room

agents first use the sensor to learn the HVAC-L

data so that it may be fed into the genetic

algorithms. Through a graphic interface, the

occupants (users) can describe their preferences to

the profile agent. The building agent uses a genetic

algorithm to determine the HVAC-L system's ideal

settings, which allows for the optimization of

energy usage and improvement of occupant

comfort.

The data utilized in the simulation were

collected from June 1 through July 30, with 20

iterations per hour of computing. Because south

zones utilize more energy during this time, the

simulation was started in June. Temperature,

illumination, and air quality are the three variables

we will use to determine the comfort function f(C),

hence the vector of weighting factors is (C1, C2,

C3) = (1, 1, 1).

We employ the air quality index, which is

shown appropriately in Table 2, in our simulation.

Table 2. Air quality index, [15].

Air quality index