A Systematic Review about Requirements Engineering Processes

for Multi-Agent Systems

Giovane D’Avila Mendonc¸a

, Iderli Pereira de Souza Filho

and Gilleanes Thorwald Araujo Guedes

Software Engineering Post-Graduation Program, Pampa Federal University, Av. Tiaraju, 810, Alegrete, Brazil

Keywords:

Requirements Engineering, Multi-Agent Systems, BDI Model, Systematic Review.

Abstract:

Requirements engineering is a crucial phase for the software development process, including multi-agent

systems. This particular kind of software is composed by agents, autonomous and pro-active entities which

can collaborate among themselves to achieve a given goal. However, multi-agent systems have some particular

requirements that are not normally found in other software. Taking this into consideration, this paper aims to

determine the actual state of the development processes which support requirements engineering for multi-

agent systems by means of a systematic review, highlighting the requirements engineering coverage and its

support to the BDI model.

1 INTRODUCTION

Agents technology is a software paradigm that pro-

vides agents abstractions for distribute and heteroge-

neous open systems development (Gan et al., 2020).

An agent is an autonomous, ﬂexible, and pro-active

process (Vicari and Gluz, 2007) that can act in its en-

vironment without being commanded by external en-

tities (Wooldridge and Jennings, 1995). Thus, soft-

ware agents are characterized as being autonomous,

having social skills, reactivity and pro-activity (Haj-

duk et al., 2018). Moreover, agents can collaborate to

achieve their goals (Deloach and Wood, 2000).

Multi-agent systems (MAS) are composed by

a number of agents interacting among themselves

(Wooldridge, 2009). This kind of system has received

great attention from scholars in several areas, includ-

ing computer science and civil engineering, as a way

to solve complex problems, by dividing them into

smaller tasks (Labba et al., 2015) (Dorri et al., 2018).

The use of MAS is present in several applications,

such as complex systems modelling, intelligent net-

works, and computer networks (Dorri et al., 2018).

Adopting an agent-oriented world view demon-

strates that most problems demand or involve multiple

agents, because they represent multiple perspectives,

https://orcid.org/0000-0001-9445-6831

https://orcid.org/0000-0001-9694-0977

https://orcid.org/0000-0001-5457-2600

because its application deﬁnes a decentralized nature

of the problem, or because there are multiple areas of

actuation in the system (Jennings, 1999).

However, developing this kind of system brought

challenges to the software engineering. Thus, a new

area arose mixing features from both software engi-

neering and artiﬁcial intelligence areas, called AOSE

- Agent-Oriented Software Engineering. The goals of

AOSE include producing methodologies, processes,

techniques, modeling languages, and tools for MAS

development (Cervenka and Trencansky, 2007) (Sl-

houb et al., 2019), in order to increase the chances of

success in MAS development (Slhoub et al., 2019).

AOSE is also concerned in adapting requirements

engineering (RE) – an area of software engineering

focused on eliciting, analysing, specifying, and vali-

dating software requirements to ensure the correct un-

derstanding of what needs to be developed (Fuentes-

Fern

andez et al., 2009). RE performs a crucial func-

tion to the development of any software, since, if

the software needs are not correctly understood, the

project will not satisfy those for whom it is intended.

According to (Dorri et al., 2018) software engi-

neering on MAS demands the speciﬁcation of those

agent behaviors needed to provide documented re-

quirements to the project and implementation phases.

Rodriguez in (Rodriguez et al., 2011) also states that

the requirements modeling in MAS requires abstrac-

tions, techniques, and notations that had been partic-

ularly adapted for this kind of domain.

Mendonça, G., Filho, I. and Guedes, G.

A Systematic Review about Requirements Engineering Processes for Multi-Agent Systems.

DOI: 10.5220/0010240500690079

In Proceedings of the 13th International Conference on Agents and Artiﬁcial Intelligence (ICAART 2021) - Volume 1, pages 69-79

ISBN: 978-989-758-484-8

Taking this in consideration, we are concerned in

how RE was speciﬁcally adapted to the development

of this kind of system. We are particularly interested

in how the RE was adapted to focus on requirements

relative to the BDI model (Bratman et al., 1987),

a model for programming intelligent agents that is

based on the beliefs, desires, and intentions of each

agent and that, according to (Singh et al., 2016) has

been widely used in MAS development.

Thus we performed a systematic review to de-

termine the state-of-art of RE for MAS. In this

review, we retrieved studies that propose pro-

cess/methodologies (or extensions of them) for MAS

development that involves RE in somehow. We intend

with this review to understand how RE is supported

by the existing processes and to determine its gaps, in

such a way to serve as a basis for future works.

This paper is organized as follows. Section 2

contains the background. Section 3 contains related

works. The research method is described in Section

4. In Section 5, the results are presented and dis-

cussed. Threats to the validity were described in Sec-

tion 6 and, in Section 7, we present the conclusion

and future works.

2 BACKGROUND

2.1 Requirements Engineering

According to (Berenbach et al., 2009), the RE goals

are: (I) to identify software requirements, (II) to anal-

yse requirements in order to classify them and to de-

rive additional requirements, as well as to solve con-

ﬂicts among them (III) to document requirements, and

(IV) to validate the documented requirements.

In SWEBOK (Bourque et al., 2014) - a reference

book in the area - is stated that the RE process cover

four main subareas: (I) Requirements Elicitation; (II)

Requirements Analysis; (III) Requirements Speciﬁca-

tion; and (IV) Requirements Validation.

Requirements elicitation investigates how to ex-

tract requirements and which are its origins. Re-

quirements analysis aims to detect and solve con-

ﬂicts among the requirements, to discover the system

boundaries. Requirements speciﬁcation, by its turn,

produce requirements documents that can be system-

atically reviewed, evaluated, and approved. Finally,

requirements validation evaluates requirements docu-

ments to ensure that the requirements be understand-

able, consistent, and complete.

2.2 Belief-Desire-Intention Model

The Belief-Desire-Intention (BDI) model is a soft-

ware model developed to programming intelligent

agents. It includes beliefs, desires, and intentions in

the agent architecture (Bratman et al., 1987).

Beliefs represent the information state the agent

owns, i. e., what he believes to be true about the en-

vironment, about itself, and about other agents. De-

sires represent the agents motivational state. They

represent the goals or situations the agent would like

to achieve. Finally, the intentions represent desires

the agent believes he can achieve and take actions to

achieve them (Rao and Georgeff, 1995).

This model allows to the agents a more com-

plex behavior than the reactive models, without the

computational overload of the cognitive architectures.

Moreover, it is easier to specify knowledge by means

of this model (Larsen, 2018).

According (Herzig et al., 2017), concepts of belief

and goal perform a central role in the conception and

implementation of autonomous agents. The concept

of BDI, consider mental attitudes to be fundamental to

the agents, where the beliefs are adapted to the envi-

ronment truths, while in the intentions, the agents try

to make the environment to correspond to its goals.

3 RELATED WORKS

We discovered some studies that aimed to identify

and to evaluate methodologies/processes in the AOSE

area. However, these studies do not follow a system-

atic vision, they are informal literature reviews with

subjective comparison criteria.

The study of (Henderson-Sellers and Gorton,

2002) discusses the state of AOSE methodologies and

how to turn them into acceptable products for the in-

dustry. This study also present a methodology clas-

siﬁcation, dividing them in (I) independent of goal-

oriented methodologies and (II) extensions of goal-

oriented methodologies to give support to the agent

concepts. The study of (Sudeikat et al., 2004) eval-

uates agent-oriented software methodologies. The

work proposes a comparison frame with four selec-

tion groups: concepts, notations, process, and prag-

matics. This proposal was evaluated comparing the

methodology adequation and its development capac-

ity. For this comparison were used three methodolo-

gies, MaSE (an old version of O-MaSE), Tropos, and

Prometheus. Finally, the work of (Cernuzzi et al.,

2005) investigates the AOSE methodologies coverage

regarding software engineering concepts. However,

besides this work not following a systematic vision, it

ICAART 2021 - 13th International Conference on Agents and Artiﬁcial Intelligence

does not present several methodologies and does not

have a wide coverage of requirements engineering.

Regarding systematic reviews, we found the study

of (Blanes et al., 2009a) that developed a review about

requirements engineering in multi-agent systems de-

velopment. However, this study tried to verify which

modeling techniques were applied in the requirements

engineering for MAS. On the other hand, our work

has as its goal to identify the coverage of the require-

ments engineering process regarding the SWEBOK

stages and its adequation to the BDI model.

4 RESEARCH METHOD

A systematic literature review (SLR) is a research

technique whose purpose is identifying, selecting,

evaluating, interpreting, and summarizing the avail-

able studies considered relevant to the research theme

or phenomenon of interest (Kitchenham and Charters,

2007). This technique searches for primary studies

related to the theme and provides a deeper synthesis

about the data obtained from these studies (Kitchen-

ham and Brereton, 2013).

A SLR has as its basis a protocol previously de-

ﬁned, that formalizes its execution, beginning by the

stipulation of the research questions, passing by es-

tablishing the studies inclusion and exclusion criteria,

selecting the digital basis for the extraction of works

related with the keywords applied during the search in

these basis, and concluding with the deﬁnition of how

the results will be presented (Biolchini et al., 2005).

Our review had as its goal to establish the state-

of-art of the process/methodologies for MAS devel-

opment that support in somehow requirements engi-

neering for this kind of system. Our main interest

is about how these processes identify and specify the

BDI model features in the requirements engineering

phase.

4.1 The Research Questions

We deﬁned four research questions to this review. The

ﬁrst research question (RQ1) aims to identify which

methodologies/processes support RE for MAS.

The second research question (RQ2) was deﬁned

to identify the coverage of the RE by these method-

ologies. We believe that with this question we can

discover possible gaps in the area and that this will

allow for future research.

The third research question (RQ3) aims to verify

which methodologies support the BDI model. As we

stated before, this is a consolidated model in the MAS

development and we believe it aggregates better reli-

ability in using the methodologies that support it.

Finally, the fourth question (RQ4) has as its goal

to show a wider view of the area needs and to focus

on the points that can be approached in future works.

The four research questions are listed below:

• RQ1: Which methodologies for the MAS devel-

opment support a speciﬁc requirements engineer-

ing (RE) life cycle to this kind of system?

• RQ2: Which is the coverage of the requirements

engineering by these methodologies taking as a

basis the subareas deﬁned by SWEBOK (Bourque

et al., 2014)?

• RQ3: Which of these methodologies focus on the

BDI model during the requirements engineering?

• RQ4: Which are the existing gaps in the method-

ologies that support RE for MAS?

4.2 Identifying and Selecting Primary

Studies

To recover relevant works for this study, we built a

String containing a set of keywords based on the re-

search questions. This String was adapted to the par-

ticularities of each bibliographic basis.

To perform this review, we used bibliographic

bases which (I) have a search mechanism based on

web; (II) have a mechanism able to use keywords;

(III) contain documents from the computer science

area; and (IV) their data bases are updated regularly.

It is important to highlight that we do not limited the

period in which the studies were published

In addition, we have included a book (Cossentino

et al., 2014) about methodologies for MAS, as well

as other classical and known studies. These studies

were manually selected by a specialist in the area be-

cause we considered that they would not be selected

in the search String as they do not present in its title,

abstract, and keywords topics related to the require-

ments engineering, since they are not processes fo-

cused on RE, though their life cycles encompass the

RE area.

In Table 1 we show the generic String used in the

basis. In addition to the search String, we used man-

ual ﬁlters in the bibliographic bases. We considered

necessary to apply these manual ﬁlters because, in

some bases, the results obtained were high and many

of the studies returned were outside the scope.

For ACM library it was used the ﬁlter “Ti-

tle/Abstract/keywords”; for Engineering Village,

“Subject/Title/Abstract”; for IEEE Xplore, All meta-

data, ﬁlters suggested by the base software “agents

A Systematic Review about Requirements Engineering Processes for Multi-Agent Systems

and multi-agent systems”; for Science Direct, “Sub-

ject/Title/Abstract” and “Title/Abstract/keywords”

and commands “multiagent OR multi-agent OR

agent-based”; for Scopus, “Title/Abstract/keywords”;

and for Springer Link it were applied the ﬁlters “Filter

of the area: Computer science”, “Filter of the subarea:

Software Engineering and Artiﬁcial intelligence”.

Table 1: String generic.

String Conector

(“multiagent” OR “multi-agent” OR “multi agent” AND

OR “agent-based” OR “agent society”

(“methodology” OR “method” OR “process”) AND

(“requirements engineering” OR “requirements

elicitation” OR “requirements modeling” OR

“requirements analysis” OR “requirements speciﬁcation”)

4.3 Inclusion and Exclusion Criteria

The selection criteria have as its goal to identify the

primary studies that provide contents to answer the

research questions. Thus, ﬁrstly the studies were anal-

ysed with basis on the title, abstract, and keywords. If

there were still doubts about the ﬁnal classiﬁcation of

a study in relation to the inclusion or exclusion crite-

ria, a specialist would be consulted. These criteria are

described in the Table 2.

Table 2: Inclusion and Exclusion Criteria.

Criterion ID Description

Inclusion IC1 Does the study presents a methodology or an

extension of a methodology for multi-agent

systems that contemplates at least one of

the requirements engineering subareas

deﬁned in the SWEBOK?

Exclusion EC1 Studies that cover a methodology already

included in more recent work.

EC2 Studies that are not a paper or a chapter of book.

EC3 Studies with less than 6 pages.

EC4 Studies that boils down to a case study or

methodology evaluation.

EC5 Studies that boils down to a comparison of

methodologies.

EC6 Studies that do not present a methodology

(or extension of a methodology) for multi-agent

systems that contemplate at least one of

requirements engineering subareas

from SWEBOK.

EC7 Studies that concentrate in other areas of

Software Engineering.

EC8 Studies that boils down to the development of

a system.

EC9 Studies that present a methodology or extension

of a methodology created only to a kind of

speciﬁc application.

4.4 Studies Quality Assessment

We deﬁned two quality criteria to evaluate the rel-

evance of the studies to the scope of this research.

These criteria were not used to the exclusion of stud-

ies, only for the ranking of studies more relevant.

Next we described the two qualitative criteria and the

score attributed for each criterion deﬁned.

1. QC1: The work supports the BDI model?

Yes (Y): the work fully supports the BDI model;

Partly (P): the work supports at least one of the

features of the BDI model;

Not (N): the work does not support the BDI

model.

2. QC2: the work applies some empirical study (ex-

periment, case study, etc.)?

Yes (Y): the study applies some empirical study;

Not (N): the study does not apply some empirical

study.

To establish a quality general index of the selected

studies, we attributed scores to each criterion deﬁned,

where Yes (Y) corresponds to 1 score, Partly (P) 0.5

score and Not (N) 0 score.

The Table 3, shows the score of each selected

study. We noticed that only three studies ((Jo and

Einhorn, 2005), (Mylopoulos et al., 2013), and (Mor-

reale et al., 2006)) reached the maximum ranking of

2 scores.

On the other hand, some studies got 0 score

((Alonso et al., 2004), (Hajer et al., 2009), (Bokma

et al., 1994), and (Gonz

alez-Moreno et al., 2014)),

though these studies did not achieve any score, they

were kept because the qualitative criteria were used

only for ranking the studies, not for eliminate them.

4.5 Data Extraction Strategy

When the studies selection process was concluded,

the basic information of each paper was registered for

data extraction. The extraction was performed using

the Google Spreadsheet to capture all the information

of each work included, allowing the posterior synthe-

sis. The data extracted from the included works were

analysed in order to answer the research questions. In

Section 5, these results were exposed and discussed.

ICAART 2021 - 13th International Conference on Agents and Artiﬁcial Intelligence

Table 3: Quality Indexes of the Studies.

Study

QC1

QC2

Total

(Ulfat-Bunyadi et al., 2018) 0.50 0.00 0.50

(Abushark et al., 2016) 0.50 1.00 1.50

(Ribino et al., 2013) 1.00 0.00 1.00

(Liu et al., 2011) 0.00 1.00 1.00

(Argente et al., 2011) 0.50 0.00 0.50

(Sen and Hemachandran, 2010) 0.50 1.00 1.50

(Blanes et al., 2009b) 0.50 1.00 1.50

(Huiying and Zhi, 2009) 0.50 0.00 0.50

(Fuentes-Fern

andez et al., 2009) 0.50 1.00 1.50

(Rodriguez et al., 2009) 0.50 1.00 1.50

(Bryl et al., 2008) 0.50 0.00 0.50

(Lee and Lee, 2008) 0.50 1.00 1.50

(Ranjan and Misra, 2006) 0.50 0.00 0.50

(Lindoso and Girardi, 2006) 0.50 0.00 0.50

(Shen et al., 2005) 0.50 0.00 0.50

(Alonso et al., 2004) 0.00 0.00 0.00

(Jo and Einhorn, 2005) 1.00 1.00 2.00

(Mylopoulos et al., 2013) 1.00 1.00 2.00

(Marcio Cysneiros and Yu, 2003) 1.00 0.00 1.00

(Chiung-Hui Leon Lee and Liu, 2002) 0.50 0.00 0.50

(Banach, 2010) 0.50 0.00 0.50

(Morreale et al., 2006) 1.00 1.00 2.00

(Murray, 2004) 0.00 1.00 1.00

(Cossentino et al., 2010) 0.50 0.00 0.50

(Bresciani and Donzelli, 2003) 0.50 1.00 1.50

(Sen and Jain, 2007b) 0.50 1.00 1.50

(Hsieh et al., 2008) 0.00 1.00 1.00

(Sutcliffe, 2001) 0.50 1.00 0.00

(Sen and Jain, 2007a) 0.50 1.00 1.50

(Haumer et al., 1999) 0.50 1.00 1.50

(Longbing Cao et al., 2004) 0.50 1.00 1.50

(Wilmann and Sterling, 2005) 0.50 0.00 0.50

(Wu et al., 2010) 0.50 1.00 1.50

(Liu and Li, 2015) 0.00 1.00 1.00

(Hajer et al., 2009) 0.00 0.00 0.00

(Ashamalla et al., 2017) 0.50 1.00 1.50

(Bokma et al., 1994) 0.00 0.00 0.00

(Hilaire et al., 2012) 0.50 0.00 0.50

(Gaur and Soni, 2012) 0.50 1.00 1.50

(Passos et al., 2015) 0.50 1.00 1.50

(Wang et al., 2013) 0.50 1.00 1.50

(Ronald et al., 2012) 0.50 1.00 1.50

(Domann et al., 2014) 0.00 1.00 1.00

(Cernuzzi et al., 2014) 0.50 0.00 0.50

(Cossentino and Seidita, 2014) 0.00 1.00 1.00

(Gonz

alez-Moreno et al., 2014) 0.00 0.00 0.00

(Bonjean et al., 2014) 0.00 1.00 1.00

(DeLoach and Garcia-Ojeda, 2014) 0.50 0.00 0.50

(Padgham et al., 2014) 1.00 0.00 1.00

(Caire et al., 2004) 0.50 1.00 1.50

(Cao, 2015) 0.50 0.00 0.50

(Glaser, 1997) 1.00 0.00 1.00

(Iglesias et al., 1998) 1.00 0.00 1.00

(Lind, 2001) 0.00 1.00 1.00

4.6 Conducting the Review

The conduction of this systematic review was per-

formed between the months of February and May of

2020. We deﬁned four stages for the studies selec-

tion: (I) executing the search String in the biblio-

graphic bases; (II) removing the duplicated studies;

(III) applying the inclusion and exclusion criteria to

the works; and (IV) reading and extracting the infor-

mation of the remaining studies of the Stage (III). The

studies were read by two reviewers in consultation

with a specialist in the area.

In Stage 1, the search String was executed in

the bibliographic bases selected for this review. The

overview of this stage can be observed in Figure 1.

The conduction began analysing the 1060 works im-

ported from the selected bibliographic bases.

In Stage 2, a total of 247 duplicated studies were

removed. In Stage 3, there were applied the inclusion

and exclusion criteria based on the reading of the title,

abstract, and keywords, resulting in the selection of 53

studies considered promising.

Figure 1: Search process.

To avoid the selection of works that are not ﬁtted

in the scope of this review, the 53 studies, selected in

the third stage, were completely read, what resulted in

the exclusion of 10 works, totalling 43 selected works.

The rejected works in this stage were inside of two

exclusion criteria:

1. Works that concentrate in other Software Engi-

neering areas: the works excluded that were inside

on this criterion were methodologies that worked

with MAS only in posterior stages to the require-

ments engineering. The requirements engineering

was performed in a traditional way, not focusing

on any particular feature of MAS.

2. Works that cover a methodology already included

in a work: for this criterion we selected the most

recent work in such a way we can understand the

current state of the methodology.

At the end of the conduction Stage, the manually se-

lected studies ((Cernuzzi et al., 2014), (Cossentino

and Seidita, 2014), (Gonz

alez-Moreno et al., 2014),

(Bonjean et al., 2014), (DeLoach and Garcia-Ojeda,

2014), (Padgham et al., 2014), (Caire et al., 2004),

(Cao, 2015), (Glaser, 1997), (Iglesias et al., 1998),

(Lind, 2001)) were added to the set of papers searched

in the bases, according with deﬁned in Section 4.2.

This resulted in a total of 54 accepted studies.

4.7 Data Extraction

For data extracting in the accepted works, we read

them all and tried to identify which SWEBOK RE

subareas each work covers, whether the methodology

A Systematic Review about Requirements Engineering Processes for Multi-Agent Systems

proposed in the study has a well-deﬁned life cycle,

whether the RE presented in the study is adequate for

MAS, and whether the study supports the BDI model.

The conduction of this stage was performed in

pairs, where each researcher read the paper and ex-

tracted the information about the issues cited previ-

ously. The conﬂicts between the researchers were de-

cided by a specialist in the area.

5 RESULTS

The relevant information of the selected studies was

obtained using the data extraction spreadsheet. The

evidence found about each research question are dis-

cussed in the next subsections.

5.1 Analysis of the Research Questions

In this section we answered the research questions of

this study and we discussed the results achieved.

Research Question 1: To answer this question,

we found 54 methodologies which approached RE for

MAS. These studies can be observed in Table 4.

Research Question 2: From the 54 selected stud-

ies we observed that all of them present Requirements

Engineering ﬁt for multi-agent systems. Thus, we

extracted which RE sub-areas deﬁned in SWEBOK

(Bourque et al., 2014) are supported by these studies.

Table 4 shows the 54 studies and the sub-areas

that they support. Great part of these studies, 46 in

the total, support the sub-area of requirements analy-

sis. While 31 of them support the sub-area of require-

ments speciﬁcation.

The sub-area of requirements elicitation, by its

turn, is supported by 16 studies. Finally, the sub-area

of requirements validation has the lower number of

studies, with only 3 of the total.

We also observed that, from these studies, only

the ADELFE methodology (Bonjean et al., 2014)

supports the four RE sub-areas (elicitation, analysis,

speciﬁcation, and validation).Moreover, the elicita-

tion in the ADELFE methodology is not suitable for

MAS, being applied a traditional elicitation. The fea-

tures suitable for a MAS began to be presented in the

analysis stage. However, this stage does not present

the means for validating the documents speciﬁc for

MAS. ADELFE validates only documents present in

a traditional requirements engineering.

Another important fact that we noticed in the ex-

traction is that only 30 studies presented some empiri-

cal experiments for the validation of the methodology.

Research Question 3: We tried to identify which

methodologies support the BDI model. We observed

that most part of the studies, 35 in the total, support

partially the BDI model, i. e., they identify at least

one of the features of this model.

These features are: agent beliefs; agent

goals/desires; and agent intentions. However, it

is necessary to state that the majority of these works

do not cite explicitly the BDI model, most of them

are goal-oriented methodologies, i. e., they focus

on just in one feature of the BDI model and they do

not necessarily use this model, but the fact that these

studies identify one of the features is useful for our

research.

Agents goals were the most identiﬁed feature, in

most cases in isolation. There are studies that identify

intentions, however we noticed that beliefs and inten-

tions are not identiﬁed in isolation, they are always

accompanied by the identiﬁcation of their goals.

Another issue to be highlighted is that 11 stud-

ies do not present support to the BDI model and only

8 present support for all these features in at least

one stage of their requirements engineering. Table 5

presents the methodologies coverage regarding their

support to the BDI model.

Research Question 4: We noticed that only three

studies cover the validation sub-area in their RE cycle.

It demonstrates that the majority of the methodologies

do not care with this phase that is so important to the

systems quality.

We also noticed that just one study covers the four

sub-areas of RE in its cycle (Bonjean et al., 2014).

On the other hand, this study does not support the

BDI model, what demonstrates a gap and the need

of the proposition of a methodology containing a re-

quirements engineering phase that supports the BDI

model.

Regarding the BDI model coverage, we under-

stand that the support to just 8 studies from a total

of 54 is a low number. Moreover, just two method-

ologies have as their focus to cover this model ((Jo

and Einhorn, 2005), (Ribino et al., 2013)) and none of

them cover elicitation and validation, what highlights

a gap in the RE for MAS area.

Other point that we could identify as a neglect

is that, among the methodologies that support BDI,

only the Tropos methodology (Mylopoulos et al.,

2013) covers the requirements elicitation and just the

methodology proposed by Cysneiros (Marcio Cys-

neiros and Yu, 2003) includes requirements valida-

tion. It demonstrates that most of the methodologies

that support BDI focus on the requirements analysis

and speciﬁcation and that, besides these areas, there

is space to be explored in the elicitation and valida-

tion areas.

ICAART 2021 - 13th International Conference on Agents and Artiﬁcial Intelligence

Table 4: Methodologies/Processes that support RE for MAS

and its coverage with relation to the SWEBOK subareas.

Methodology

Elicitation

Analysis

Speciﬁcation

Validation

KAOS Extension (Ulfat-Bunyadi et al., 2018) X X

JAAMAS (Abushark et al., 2016) X

Patrizia Ribino (Ribino et al., 2013) X

AGSIRA (Liu et al., 2011) X X

GORMAS (Argente et al., 2011) X X

ATABGE (Sen and Hemachandran, 2010) X

RE4Gaia (Blanes et al., 2009b) X

Xu Huiying (Huiying and Zhi, 2009) X X

REG for AOSE (Fuentes-Fern

andez et al.,

2009)

X X

Extension GAIA (Rodriguez et al., 2009) X

B-Tropos (Bryl et al., 2008) X X X

JONGWON LEE (Lee and Lee, 2008) X

Prabhat Ranjan (Ranjan and Misra, 2006) X X X

SRAMO (Lindoso and Girardi, 2006) X

Zhiqi Shen (Shen et al., 2005) X

SONIA (Alonso et al., 2004) X X

BDI ASP (Jo and Einhorn, 2005) X

Tropos (Mylopoulos et al., 2013) X X X

Cysneiros (Marcio Cysneiros and Yu, 2003) X X X

Chiung-Hui (Chiung-Hui Leon Lee and Liu,

2002)

X X

KAOS (Banach, 2010) X X

PRACTIONIST (Morreale et al., 2006) X X

Murray (Murray, 2004) X

ASPECS (Cossentino et al., 2010) X X

REF (Bresciani and Donzelli, 2003) X X

Sen and Jain (Sen and Jain, 2007b) X

Hsieh et al. (Hsieh et al., 2008) X X

Sutcliffe (Sutcliffe, 2001) X X

Agile Sen and Jain (Sen and Jain, 2007a) X

CREWS-EVE (Haumer et al., 1999) X X X

Cao et al. (Longbing Cao et al., 2004) X X

HOMER (Wilmann and Sterling, 2005) X

Wu et al. (Wu et al., 2010) X

Liu and Li (Liu and Li, 2015) X

Mahmoud et al. (Hajer et al., 2009) X

Ashamalla et al. (Ashamalla et al., 2017) X X

Consensus (Bokma et al., 1994) X X

Hilaire et al. (Hilaire et al., 2012) X

Gaur and Soni (Gaur and Soni, 2012) X

Passos et al. (Passos et al., 2015) X X

PLANT (Wang et al., 2013) X X

Ronald et al. (Ronald et al., 2012) X X

aMIAC (Domann et al., 2014) X X

GAIA (Cernuzzi et al., 2014) X X

PASSI (Cossentino and Seidita, 2014) X X X

INGENIAS-SCRUM (Gonz

alez-Moreno et al.,

2014)

X X X

ADELFE (Bonjean et al., 2014) X X X X

O-MaSE (DeLoach and Garcia-Ojeda, 2014) X X

PROMETHEUS (Padgham et al., 2014) X

MESSAGE (Caire et al., 2004) X X

OSOAD (Cao, 2015) X

COMOMAS (Glaser, 1997) X

MAS-COMMONKADS (Iglesias et al., 1998) X X

MASSIVE (Lind, 2001) X X X

Total 16 46 31 3

6 THREATS TO VALIDITY

During the planning and execution of this review,

some factors were characterized as threats to the re-

search validity. The potential threats are discussed to

orient the interpretation of this work:

1. Construct Validity: The reliability of the search

string deﬁned to select relevant works can be a

threat to the construct. To minimize this threat the

string was calibrated with the execution of several

Table 5: Coverage of methodologies/Processes regarding

the BDI model support.

Methodology

Belief

Desire (Goal)

Intention

Not support

KAOS Extension (Ulfat-Bunyadi et al., 2018) X X

JAAMAS (Abushark et al., 2016) X

Patrizia Ribino (Ribino et al., 2013) X X X

AGSIRA (Liu et al., 2011) X

GORMAS (Argente et al., 2011) X

ATABGE (Sen and Hemachandran, 2010) X

RE4Gaia (Blanes et al., 2009b) X

Xu Huiying (Huiying and Zhi, 2009) X X

REG for AOSE (Fuentes-Fern

andez et al.,

2009)

Extension GAIA (Rodriguez et al., 2009) X

B-Tropos (Bryl et al., 2008) X

JONGWON LEE (Lee and Lee, 2008) X

Prabhat Ranjan (Ranjan and Misra, 2006) X

SRAMO (Lindoso and Girardi, 2006) X

Zhiqi Shen (Shen et al., 2005) X

SONIA (Alonso et al., 2004) X

BDI ASP (Jo and Einhorn, 2005) X X X

Tropos (Mylopoulos et al., 2013) X X X

Cysneiros (Marcio Cysneiros and Yu, 2003) X X X

Chiung-Hui (Chiung-Hui Leon Lee and Liu,

2002)

KAOS (Banach, 2010) X

PRACTIONIST (Morreale et al., 2006) X X X

Murray (Murray, 2004) X

ASPECS (Cossentino et al., 2010) X

REF (Bresciani and Donzelli, 2003) X

Sen and Jain (Sen and Jain, 2007b) X

Hsieh et al. (Hsieh et al., 2008) X

Sutcliffe (Sutcliffe, 2001) X

Agile Sen and Jain (Sen and Jain, 2007a) X

CREWS-EVE (Haumer et al., 1999) X

Cao et al. (Longbing Cao et al., 2004) X

HOMER (Wilmann and Sterling, 2005) X

Wu et al. (Wu et al., 2010) X

Liu and Li (Liu and Li, 2015) X

Mahmoud et al. (Hajer et al., 2009) X

Ashamalla et al. (Ashamalla et al., 2017) X

Consensus (Bokma et al., 1994) X

Hilaire et al. (Hilaire et al., 2012) X

Gaur and Soni (Gaur and Soni, 2012) X

Passos et al. (Passos et al., 2015) X

PLANT (Wang et al., 2013) X

Ronald et al. (Ronald et al., 2012) X

aMIAC (Domann et al., 2014) X

GAIA (Cernuzzi et al., 2014) X

PASSI (Cossentino and Seidita, 2014) X

INGENIAS-SCRUM (Gonz

alez-Moreno et al.,

2014)

ADELFE (Bonjean et al., 2014) X

O-MaSE (DeLoach and Garcia-Ojeda, 2014) X X

PROMETHEUS (Padgham et al., 2014) X X X

MESSAGE (Caire et al., 2004) X

OSOAD (Cao, 2015) X

COMOMAS (Glaser, 1997) X X X

MAS-COMMONKADS (Iglesias et al., 1998) X X X

MASSIVE (Lind, 2001) X

Total 8 43 11 11

tests and the area expert was consulted about the

most used terms.

2. Internal Validity: A possible threat could have

arisen from the individual interpretation of each

researcher, something that could have led to the

exclusion of relevant studies. To minimize this

threat, the protocol of this review was strictly fol-

lowed, considering mainly the inclusion and ex-

clusion criteria. When necessary, a researcher

with experience in this area was consulted to reach

a consensus about the acceptance of the identiﬁed

studies.

A Systematic Review about Requirements Engineering Processes for Multi-Agent Systems

3. External Validity: Another possible threat is

that some studies could not have been found be-

cause it does not contain keywords deﬁned in the

search string. To minimize this threat, the book

“Handbook on Agent-Oriented Design Processes”

(Cossentino et al., 2014) was used as research

source and some classical papers were manually

selected by a specialist in the area. To comple-

ment the research we performed a manual search

in the methodologies found aiming to ensure the

use of studies with the most recent version.

4. Coverage Validity:

Regarding the possible papers that were not cap-

tured by our String, we intend, as a future work,

to apply the snowballing technique trying to ﬁnd

more relevant papers. Another issue is that the

snowballing technique can allow us to ﬁnd more

papers about the analysed methodologies, since in

this analysis we focused only on the last paper of

each methodology and this practice may not fully

guarantee a complete coverage of the methodol-

ogy.

5. Conclusion Validity: In spite of following a sys-

tematic protocol, systematic reviews are subject

to human error, especially in the data extraction

from papers. To mitigate this threat, the data ex-

traction was performed by two independent re-

searchers following the strategy deﬁned in subsec-

tion 4.7 and, in case of divergences, a specialist in

the area was consulted.

7 CONCLUSIONS AND FUTURE

WORK

In this systematic review, we answered the research

questions about which methodologies for multi-agent

systems support the requirements engineering life cy-

cle, which is the coverage of requirements engineer-

ing in these methodologies, and which of them fo-

cused on the BDI model. This way, aiming to search

and categorize the studies directly related to the theme

and to support posteriorly the development of new sci-

entiﬁc researches in the area.

This revision was carried out by two reviewers.

A third specialist reviewer had as its function to de-

cide about the conﬂicts. The review phases were com-

posed by the protocol deﬁnition, conduction of the re-

view, and studies extraction.

The initial search returned 1060 studies. The ap-

plication of the inclusion and exclusion criteria re-

sulted in the selection of a total of 43 studies. Af-

ter the inclusion of the studies manually selected we

obtained a total of 54 studies.

Regarding the requirements engineering life cy-

cle, we identiﬁed 54 studies that support at least one

of the requirements engineering subareas. From these

studies, we observed that only 31 of the total present

a well-deﬁned RE life cycle.

The synthesis of the data guided us to some in-

teresting observations. Among them, we noticed that

only 3 studies support the requirements validation

subarea ((Bonjean et al., 2014), (Marcio Cysneiros

and Yu, 2003), (Haumer et al., 1999)). We also no-

ticed that only the ADELFE methodology (Bonjean

et al., 2014) covers the four requirements engineer-

ing sub-areas. And, ﬁnally, that only 8 methodologies

support the BDI model.

We identiﬁed some gaps that demonstrate the need

of speciﬁc studies to multi-agent systems. Among

them, there is the need of proposing a methodology

that covers the four requirements engineering sub-

areas, considering that ADELFE methodology (Bon-

jean et al., 2014) is speciﬁc for adaptive MAS and that

it does not support the BDI model. Another identiﬁed

gap is the weak coverage of the validation sub-area,

present in just three studies. We also noticed that none

methodology covers the four requirements engineer-

ing sub-areas with support to the BDI model.

That said, as a future work, we will propose a re-

quirements engineering process for multi-agents sys-

tems, supporting and containing guidelines for the

four requirements engineering sub-areas and with

support to the BDI model, as well. This work is al-

ready in an advanced stage. We also intend to extend

this process in such a way that it encompasses all the

development life cycle of multi-agent systems with

focus on the BDI model.

REFERENCES

Abushark, Y., Miller, T., Thangarajah, J., Winikoff, M., and

Harland, J. (2016). Requirements speciﬁcation via ac-

tivity diagrams for agent-based systems. Autonomous

Agents and Multi-Agent Systems, 31. Springer New

York LLC.

Alonso, F., Frutos, S., Mart

ınez Normand, L., and Montes,

C. (2004). Sonia: A methodology for natural agent

development. volume 3451, pages 245–260. Springer.

Argente, E., Botti, V., and Julian, V. (2011). Gormas:

An organizational-oriented methodological guideline

for open mas. In Gleizes, M.-P. and Gomez-Sanz,

J. J., editors, Agent-Oriented Software Engineering X,

pages 32–47, Berlin, Heidelberg. Springer Berlin Hei-

delberg.

Ashamalla, A., Beydoun, G., and Low, G. (2017). Model

driven approach for real-time requirement analysis of

ICAART 2021 - 13th International Conference on Agents and Artiﬁcial Intelligence

multi-agent systems. Computer Languages, Systems

& Structures, 50:127 – 139. Elsevier.

Banach, R. (2010). A deidealisation semantics for kaos. In

Proceedings of the 2010 ACM Symposium on Applied

Computing, SAC ’10, page 267–274, New York, NY,

USA. Association for Computing Machinery.

Berenbach, B., Paulish, D., Kazmeier, J., and Rudorfer,

A. (2009). Software & systems requirements engi-

neering: in practice. McGraw-Hill, Inc. 978-0-07-

160548-9.

Biolchini, J., Mian, P. G., Natali, A. C. C., and Travassos,

G. H. (2005). Systematic review in software engineer-

ing. System Engineering and Computer Science De-

partment COPPE/UFRJ.

Blanes, D., Insfran, E., and Abrah

ao, S. (2009a). Require-

ments engineering in the development of multi-agent

systems: a systematic review. In International Confer-

ence on Intelligent Data Engineering and Automated

Learning, pages 510–517. Springer.

Blanes, D., Insfran, E., and Abrah

ao, S. (2009b). Re4gaia:

A requirements modeling approach for the develop-

ment of multi-agent systems. volume 59, pages 245–

252.

Bokma, A., Slade, A., Kerridge, S., and Johnson, K.

(1994). Engineering large-scale agent-based systems

with consensus. Robotics and Computer-Integrated

Manufacturing, 11(2):81 – 90. Elsevier.

Bonjean, N., Mefteh, W., Gleizes, M. P., Maurel, C.,

and Migeon, F. (2014). ADELFE 2.0, pages 19–63.

Springer Berlin Heidelberg, Berlin, Heidelberg.

Bourque, P., Fairley, R. E., et al. (2014). Guide to the soft-

ware engineering body of knowledge (SWEBOK (R)):

Version 3.0. IEEE Computer Society Press. 978-0-

7695-5166-1.

Bratman, M. et al. (1987). Intention, plans, and practical

reason, volume 10. Harvard University Press Cam-

bridge, MA. 1-57586-192-5.

Bresciani, P. and Donzelli, P. (2003). Ref: A practical

agent-based requirement engineering framework. In

ER 2003 Workshops ECOMO, volume 2814, pages

217–228. Springer.

Bryl, V., Mello, P., Montali, M., Torroni, P., and Zannone,

N. (2008). B-tropos: Agent-oriented requirements en-

gineering meets computational logic for declarative

business process modeling and veriﬁcation. volume

5056, pages 157 – 176, Porto, Portugal. Springer Ver-

lag.

Caire, G., Coulier, W., Garijo, F., G

omez-Sanz, J., Pav

on,

J., Kearney, P., and Massonet, P. (2004). The Message

Methodology, volume 11, pages 177–194. Springer

US, Boston, MA.

Cao, L. (2015). Osoad methodology. In Metasynthetic

Computing and Engineering of Complex Systems,

pages 111–129, London. Springer London.

Cernuzzi, L., Cossentino, M., and Zambonelli, F.

(2005). Process models for agent-based develop-

ment. Engineering Applications of Artiﬁcial Intelli-

gence, 18(2):205–222.

Cernuzzi, L., Molesini, A., and Omicini, A. (2014). The

Gaia Methodology Process, pages 141–172. Springer

Berlin Heidelberg, Berlin, Heidelberg.

Cervenka, R. and Trencansky, I. (2007). The Agent Mod-

eling Language - AML: A Comprehensive Approach

to Modeling Multi-Agent Systems. Birkh

auser Basel.

978-3-7643-8396-1.

Chiung-Hui Leon Lee and Liu, A. (2002). A method for

agent-based system requirements analysis. In Fourth

International Symposium on Multimedia Software En-

gineering, 2002. Proceedings., pages 214–221. IEEE.

Cossentino, M., Galland, S., Gaud, N., Hilaire, V., and

Koukam, A. (2010). A glimpse of the aspecs process

documented with the ﬁpa dpdf template. In Proceed-

ings of The Multi-Agent Logics, Languages, and Or-

ganisations Federated Workshops (MALLOW 2010),

volume 627.

Cossentino, M., Hilaire, V., Molesini, A., and Seidita, V.

(2014). Handbook on Agent-Oriented Design Pro-

cesses. Springer-Verlag Berlin Heidelberg. 978-3-

642-39975-6.

Cossentino, M. and Seidita, V. (2014). PASSI: Process

for Agent Societies Speciﬁcation and Implementation,

pages 287–329. Springer Berlin Heidelberg, Berlin,

Heidelberg.

Deloach, S. and Wood, M. (2000). Developing Multiagent

Systems with agentTool. volume 1986, pages 46–60.

Springer Verlag.

DeLoach, S. A. and Garcia-Ojeda, J. C. (2014). The O-

MASE Methodology, pages 253–285. Springer Berlin

Heidelberg, Berlin, Heidelberg.

Domann, J., Hartmann, S., Burkhardt, M., Barge, A., and

Albayrak, S. (2014). An agile method for multiagent

software engineering. Procedia Computer Science,

32:928 – 934. The 5th International Conference on

Ambient Systems, Networks and Technologies (ANT-

2014), the 4th International Conference on Sustain-

able Energy Information Technology (SEIT-2014), El-

sevier.

Dorri, A., Kanhere, S. S., and Jurdak, R. (2018). Multi-

agent systems: A survey. IEEE Access, 6:28573–

28593. Institute of Electrical and Electronics Engi-

neers Inc.

Fuentes-Fern

andez, R., G

omez-Sanz, J., and Pav

on, J.

(2009). Requirements elicitation and analysis of mul-

tiagent systems using activity theory. Systems, Man

and Cybernetics, Part A: Systems and Humans, IEEE

Transactions on, 39:282 – 298. IEEE.

Gan, K. S., Anthony, P., Chin, K. O., and Hamdan, A. R.

(2020). Enforcing social semantic in ﬁpa-acl using

spin. In Agents and Multi-agent Systems: Technolo-

gies and Applications 2019, pages 3–13. Springer.

Gaur, V. and Soni, A. (2012). A novel approach to explore

inter agent dependencies from user requirements. Pro-

cedia Technology, 1:412–419. Elsevier.

Glaser, N. (1997). The comomas methodology and en-

vironment for multi-agent system development. In

Zhang, C. and Lukose, D., editors, Multi-Agent Sys-

tems Methodologies and Applications, volume 1286,

pages 1–16, Berlin, Heidelberg. Springer Berlin Hei-

delberg.

Gonz

alez-Moreno, J. C., G

omez-Rodr

ıguez, A., Fuentes-

Fern

andez, R., and Ramos-Valc

arcel, D. (2014).

INGENIAS-Scrum, pages 219–251. Springer Berlin

Heidelberg, Berlin, Heidelberg.

A Systematic Review about Requirements Engineering Processes for Multi-Agent Systems

Hajduk, M., Sukop, M., and Haun, M. (2018). Cognitive

Multi-agent Systems: Structures, Strategies and Ap-

plications to Mobile Robotics and Robosoccer, vol-

ume 138. Springer. 978-3-319-93687-1.

Hajer, B. M., Taieb, B. R., and Raouf, K. (2009). A new

mas based approach modeling the qms continual im-

provement. In 2009 IEEE International Conference

on Systems, Man and Cybernetics, pages 4734–4739.

IEEE.

Haumer, P., Heymans, P., Jarke, M., and Pohl, K. (1999).

Bridging the gap between past and future in re: a

scenario-based approach. In Proceedings IEEE In-

ternational Symposium on Requirements Engineering

(Cat. No.PR00188), pages 66–73. IEEE.

Henderson-Sellers, B. and Gorton, I. (2002). Agent-based

software development methodologies. In White Paper,

Summary of Workshop at the OOPSLA, volume 2003.

Herzig, A., Lorini, E., Perrussel, L., and Xiao, Z. (2017).

Bdi logics for bdi architectures: old problems, new

perspectives. KI-K

unstliche Intelligenz, 31(1):73–83.

Hilaire, V., Cossentino, M., Gechter, F., Rodriguez, S., and

Koukam, A. (2012). An approach for the integration

of swarm intelligence in mas: An engineering per-

spective. Expert Systems with Applications, 40:1–24.

Elsevier.

Hsieh, M., Hung, W., Shin, S., Lin, S., and Huang, T.

(2008). Spoken dialogue agent interface requirements

modeling based on passi methodology. In 2008 Eighth

International Conference on Intelligent Systems De-

sign and Applications, volume 1, pages 339–342.

IEEE.

Huiying, X. and Zhi, J. (2009). An agent-oriented require-

ment graphic symbol representation and formalization

modeling method. volume 4, pages 569 – 574. IEEE

Computer Society.

Iglesias, C. A., Garijo, M., Gonz

alez, J. C., and Velasco,

J. R. (1998). Analysis and design of multiagent

systems using mas-commonkads. In Singh, M. P.,

Rao, A., and Wooldridge, M. J., editors, Intelligent

Agents IV Agent Theories, Architectures, and Lan-

guages, volume 1365, pages 313–327, Berlin, Heidel-

berg. Springer Berlin Heidelberg.

Jennings, N. R. (1999). Agent-oriented software engineer-

ing. In Garijo, F. J. and Boman, M., editors, Euro-

pean Workshop on Modelling Autonomous Agents in

a Multi-Agent World, pages 1–7, Berlin, Heidelberg.

Springer Berlin Heidelberg.

Jo, C.-H. and Einhorn, J. (2005). A bdi agent-based soft-

ware process. Journal of Object Technology, 4:101–

121.

Kitchenham, B. and Brereton, P. (2013). A systematic re-

view of systematic review process research in soft-

ware engineering. Information and Software Technol-

ogy, 55(12):2049–2075. Butterworth-Heinemann.

Kitchenham, B. and Charters, S. (2007). Guidelines for per-

forming systematic literature reviews in software en-

gineering. 2.

Labba, C., Bellamine ben Saoud, N., and Dugdale, J.

(2015). Towards a conceptual framework to support

adaptative agent-based systems partitioning. IEEE.

Institute of Electrical and Electronics Engineers Inc.

Larsen, J. B. (2018). Agent Programming Languages and

Logics in Agent-Based Simulation, pages 517–526.

Springer International Publishing, Cham.

Lee, J. and Lee, H. (2008). Strategic agent based web sys-

tem development methodology. International Jour-

nal of Information Technology & Decision Making,

7:309–337. World Scientiﬁc.

Lind, J. (2001). Iterative Software Engineering for Multia-

gent Systems: The MASSIVE Method. Springer-Verlag

Berlin Heidelberg. 978-3-540-45162-4.

Lindoso, A. and Girardi, R. (2006). The sramo technique

for analysis and reuse of requirements in multi-agent

application engineering. In 9th Workshop on Require-

ments Engineering, pages 41–50. Workshop on Re-

quirements Engineering.

Liu, L., Jin, Z., Lu, R., and Yang, H. (2011). Agent-oriented

requirements analysis from scenarios. In O’Shea, J.,

Nguyen, N. T., Crockett, K., Howlett, R. J., and Jain,

L. C., editors, Agent and Multi-Agent Systems: Tech-

nologies and Applications, pages 394–405, Berlin,

Heidelberg. Springer Berlin Heidelberg.

Liu, W. and Li, M. (2015). Requirements planning with

event calculus for runtime self-adaptive system. In

2015 IEEE 39th Annual Computer Software and Ap-

plications Conference, volume 2, pages 77–82. IEEE.

Longbing Cao, Chengqi Zhang, Dan Luo, Wanli Chen,

and Zamani, N. (2004). Integrative early require-

ments analysis for agent-based systems. In Fourth In-

ternational Conference on Hybrid Intelligent Systems

(HIS’04), pages 118–123. IEEE.

Marcio Cysneiros, L. and Yu, E. (2003). Requirements en-

gineering for large-scale multi-agent systems. In Gar-

cia, A., Lucena, C., Zambonelli, F., Omicini, A., and

Castro, J., editors, Software Engineering for Large-

Scale Multi-Agent Systems, volume 2603, pages 39–

56, Berlin, Heidelberg. Springer Berlin Heidelberg.

Morreale, V., Bonura, S., Francaviglia, G., Centineo, F.,

Cossentino, M., and Gaglio, S. (2006). Goal-oriented

development of bdi agents: the practionist approach.

In Proceedings of the 2006 IEEE/WIC/ACM Inter-

national Conference on Intelligent Agent Technology,

pages 66–72. IEEE.

Murray, J. (2004). Specifying agent behaviors with uml

statecharts and statedit. In Polani, D., Browning,

B., Bonarini, A., and Yoshida, K., editors, RoboCup

2003: Robot Soccer World Cup VII, volume 3020,

pages 145–156, Berlin, Heidelberg. Springer Berlin

Heidelberg.

Mylopoulos, J., Castro, J., and Kolp, M. (2013). The evo-

lution of tropos. In Bubenko, J., Krogstie, J., Pastor,

O., Pernici, B., Rolland, C., and Sølvberg, A., editors,

Seminal Contributions to Information Systems Engi-

neering: 25 Years of CAiSE, pages 281–287. Springer

Berlin Heidelberg, Berlin, Heidelberg.

Padgham, L., Thangarajah, J., and Winikoff, M. (2014).

Prometheus Research Directions, pages 155–171.

Springer Berlin Heidelberg, Berlin, Heidelberg.

Passos, L. S., Rossetti, R. J., and Gabriel, J. (2015). An

agent methodology for processes, the environment,

and services. In Rossetti, R. J. and Liu, R., editors, Ad-

vances in Artiﬁcial Transportation Systems and Simu-

lation, pages 37 – 53. Academic Press, Boston.

ICAART 2021 - 13th International Conference on Agents and Artiﬁcial Intelligence

Ranjan, P. and Misra, A. (2006). A novel approach

of requirements analysis for agent based system.

In Seventh ACIS International Conference on Soft-

ware Engineering, Artiﬁcial Intelligence, Network-

ing, and Parallel/Distributed Computing (SNPD’06),

pages 299– 304. IEEE.

Rao, A. S. and Georgeff, M. P. (1995). Bdi agents: From

theory to practice. In Proceedings of the First Inter-

national Conference on Multi-Agent Systems (ICMAS-

95), pages 312–319.

Ribino, P., Lodato, Lopes, Sabatucci, Cossentino, M.,

Lodato, C., Lopes, S., Sabatucci, L., and Seidita, V.

(2013). Ontology and goal model in designing bdi

multi-agent systems. volume 1099. CEUR-WS.

Rodriguez, L., Hume, A., Cernuzzi, L., and Insfran, E.

(2009). Improving the quality of agent-based systems:

Integration of requirements modeling into gaia. In

2009 Ninth International Conference on Quality Soft-

ware, pages 278–283. IEEE Computer Society.

Rodriguez, L., Insfr

an, E., and Cernuzzi, L. (2011). Re-

quirements modeling for multi-agent systems. In

Multi-Agent Systems - Modeling, Control, Program-

ming, Simulations and Applications. IntechOpen.

Ronald, N., Dignum, V., Jonker, C., Arentze, T., and

Timmermans, H. (2012). On the engineering of

agent-based simulations of social activities with so-

cial networks. Information and Software Technology,

54(6):625 – 638. Special Section: Engineering Com-

plex Software Systems through Multi-Agent Systems

and Simulation, Elsevier.

Sen, A. and Hemachandran, K. (2010). Elicitation of

goals in requirements engineering using agile meth-

ods. In 2010 IEEE 34th Annual Computer Software

and Applications Conference Workshops, pages 263–

268. IEEE.

Sen, A. M. and Jain, S. K. (2007a). An agile technique for

agent based goal reﬁnement to elicit soft goals in goal

oriented requirements engineering. In 15th Interna-

tional Conference on Advanced Computing and Com-

munications (ADCOM 2007), pages 41–47. IEEE.

Sen, A. M. and Jain, S. K. (2007b). A visualization tech-

nique for agent based goal reﬁnement to elicit soft

goals in goal oriented requirements engineering. In

Second International Workshop on Requirements En-

gineering Visualization (REV 2007), pages 2–2. IEEE.

Shen, Z., Li, D., Miao, C., Gay, R., and Miao, Y. (2005).

Goal-oriented methodology for agent system develop-

ment. volume 89, pages 95–101. The Institute of Elec-

tronics, Information and Communication Engineers.

Singh, D., Padgham, L., and Logan, B. (2016). Integrat-

ing bdi agents with agent-based simulation platforms.

Autonomous Agents and Multi-Agent Systems, 30.

Slhoub, K., Carvalho, M., and Nembhard, F. (2019). Eval-

uation and comparison of agent-oriented methodolo-

gies: A software engineering viewpoint. In 2019 IEEE

International Systems Conference (SysCon), pages 1–

8. IEEE.

Sudeikat, J., Braubach, L., Pokahr, A., and Lamersdorf,

W. (2004). Evaluation of agent–oriented software

methodologies–examination of the gap between mod-

eling and platform. In International Workshop on

Agent-Oriented Software Engineering, pages 126–

141. Springer.

Sutcliffe, A. (2001). Requirements engineering for complex

collaborative systems. In Proceedings Fifth IEEE In-

ternational Symposium on Requirements Engineering,

pages 110–117. IEEE.

Ulfat-Bunyadi, N., Mohammadi, N., and Heisel, M. (2018).

Supporting the systematic goal reﬁnement in kaos us-

ing the six-variable model. In 13th International Con-

ference on Software Technologies, pages 102–111.

SciTePress.

Vicari, R. M. and Gluz, J. C. (2007). An intelligent tutor-

ing system (its) view on aose. International Journal

of Agent-Oriented Software Engineering, 1(3-4):295–

333. Inderscience Publishers.

Wang, Y., Zhao, L., Wang, X., Yang, X., and Supakkul,

S. (2013). Plant: A pattern language for transform-

ing scenarios into requirements models. International

Journal of Human-Computer Studies, 71:1026–1043.

Elsevier.

Wilmann, D. and Sterling, L. (2005). Guiding agent-

oriented requirements elicitation: Homer. In

Fifth International Conference on Quality Software

(QSIC’05), pages 419–424. IEEE.

Wooldridge, M. (2009). An introduction to multiagent sys-

tems. John Wiley & Sons. 978-0-470-51946-2.

Wooldridge, M. and Jennings, N. R. (1995). Intelligent

agents: Theory and practice. The knowledge engineer-

ing review, 10(2):115–152.

Wu, H., Liu, L., and Ma, W. (2010). Optimizing require-

ments elicitation with an i* and bayesian network

integrated modelling approach. In IEEE 34th An-

nual Computer Software and Applications Conference

Workshops, pages 182–188. IEEE.

A Systematic Review about Requirements Engineering Processes for Multi-Agent Systems