THE PROTÉGÉ - PROMETHEUS APPROACH TO SUPPORT

MULTI-AGENT SYSTEMS CREATION

Marina V. Sokolova

1,2

and Antonio Fernández-Caballero

Universidad de Castilla-La Mancha, Departamento de Sistemas Informáticos &

Instituto de Investigación en Informática de Albacete, Campus Universitario s/n, 02071-Albacete, Spain

Kursk State Technical University, ul.50 let Oktiabrya, 94, Kursk, 305040, Russia

Keywords: Ontologies, Intelligent agents, Mapping, Prometheus, Protégé.

Abstract: The integration of two existing and widely accepted tools, Protégé Ontology Editor and Knowledge-Base

Framework, and Prometheus Development Kit, into a common approach, aiming to include the principal

stages of MAS development life cycle and offering a general sequence of steps facilitating application

creation, is proposed in this paper. The approach is successfully being applied to situation assessment

issues, which has concluded in an agent-based decision-support system for environmental impact

evaluation.

1 INTRODUCTION

Creation, deployment and post implementation of a

multi-agent system (MAS) as a software product is a

complex process, which passes through a sequence

of stages forming its life cycle (Marik and

McFarlane, 2005; Vasconcelos et al., 2001). Every

step of the life cycle process has to be supported and

provided by means of program tools and

methodologies. In case of MAS development, in our

opinion there is still no solution to a unified

approach to cover all the stages. However, there are

some works dedicated to this issue (de Wolf and

Holvoet, 2005; Konichenko, 2005). For instance, de

Wolf and Holvoet have presented a methodology in

the context of standard life cycle model, with accent

to decentralization and macroscopic view of the

process. Within tools and frameworks the authors

mention Jade (Bellifemine, Poggi and Rimassa,

1999), Repast (Repast, 2003) and an environment

for coordination of situated multi-agent systems

(Schelfthout and Holvoet, 2004). The authors offer

their approach on the assumption that the research

task has already been defined, omitting the problem

definition and domain analysis stages of the MAS

development process.

The software development in case of MAS is

based on the following steps (Konichenko, 2005):

(1) domain and system requirements analysis; (2)

design; (3) implementation; (4) verification; (5)

maintenance.

System analysis and design stages are supported

by well known alternative agent-oriented software

engineering methodologies, including MaSE

(DeLoach, Wood and Sparkman 2001), Gaia

(Wooldridge, Jennings and Kinny 2000), MASDK

(Gorodetsky et al., 2005), Prometheus (Padgham and

Winikoff, 2002), Tropos (Giunchiglia, Mylopoulos

and Perini, 2002) among others.

The MAS, after being coded and tested, is ready

to be deployed and be further modified or changed,

depending from requirements. The existing

methodologies do not or only briefly extend over the

first stage, as they work under the condition that the

developer has already defined the problem and

determined the goals and the tasks of the system.

However, this stage is a crucial one and has to be

carefully examined and planned. Indeed, the whole

deployed system functionality and efficiency

depends on how precisely the problem was defined

and the domain ontology was elaborated.

In this work we introduce our approach to MAS

life-cycle support, in which we focus both on

creation ontological background and system design

and deployment.

The paper is organized as follows. In section 2

the meta-ontology creation realized in Protégé is

described and in section 3 the MAS design made in

PDT is introduced. In section 4 our intention to

442

V. Sokolova M. and Fernández-Caballero A. (2008).

THE PROTÉGÉ - PROMETHEUS APPROACH TO SUPPORT MULTI-AGENT SYSTEMS CREATION.

In Proceedings of the Tenth International Conference on Enterprise Information Systems - AIDSS, pages 442-445

DOI: 10.5220/0001685304420445

 SciTePress

implement the ideas of the integrated methodology

for practical issue is briefly explained.

2 REQUIREMENTS IN PROTÉGÉ

Ontology creation may be viewed as a crucial step in

MAS design as it determines the system knowledge

area and potential capabilities (Guarino and Giaretta,

1995). In the first part of this article a model of

distributed meta-ontology that serves as a

framework for MAS design is proposed. Its

components - private ontologies - are described in

extensive with respect to an application area and in

terms of the used semantics (Sokolova and

Fernández-Caballero, 2007).

The model of the meta-ontology that we have

created consists of five components, or private

ontologies: the “Domain Ontology”, the “Task

Ontology”, the “Ontology of MAS”, the “Interaction

Ontology” and the “Agent Ontology”.

In first place, the “Domain Ontology”, includes

the objects of the problem area, relations between

them and their properties.

OD = < C, I, P, V, Rs, Rl>

(1)

where the sets C and I are classes and their

individuals; P are class properties; V are the

properties values; Rs are some marginal levels for

values (restrictions); Rl states rules to receive new

individuals for the concrete class.

The “Task Ontology” contains information about

tasks and respective methods, about the pre-task and

post-task conditions, and informational flows for

every task. The “Task Ontology” has the following

model:

OT=<T, M, In, Ot, R>

(2)

where T is a set of tasks to be solved in the MAS,

and M is a set of methods or activities related to the

concrete task, In and Ot are input and output data

flows, R is a set of roles that use the task.

Component R is inherited from the “Ontology of

MAS” through the property belong to role. The tasks

are shared and accomplished in accordance with an

order. The “Ontology for MAS” architecture is

stated as:

OA=<L, R, IF, Or>

(3)

where L corresponds to the logical levels of the

MAS (if required), R is a set of determined roles, IF

is a set of the corresponding input and output

information represented by protocols. Lastly, the set

Or determines the sequence of execution for every

role (orders).

The interactions between the agents include an

initiator and a receiver, a scenario and the roles

taken by the interacting agents, the input and output

information and a common communication

language. They are stated in the “Interaction

Ontology” as:

OI=<In, Rc, Sc, R, In, Ot, L>

(4)

Actually, as In and Rc Initiator and Receiver,

respectively, of the interaction we use roles. The

component Sc corresponds to protocols. R is a set

of roles that the agents play during the interaction.

In and Ot are represented by informational

resources, required as input and output, respectively.

Language L determines the agent communication

language (ACL).

In our approach BDI agents, which are

represented by the “Agent Ontology”, are modelled.

Hence, every agent is described as a composition of

the following components (Georgeff et al., 1998):

Agent = <B, D, I>

(5)

Every agent has a detailed description in

accordance with the given ontology, which is

offered in a form of BDI cards, in which the pre-

conditions and post-conditions of agent execution

and the necessary conditions and resources for the

agent successful execution are stated. Evidently, B,

D and I stand for Believes, Desires and Intentions,

respectively.

Thus, the “Agent Ontology” incorporates the

following components: B is related to “Domain

Ontology”, which determines information data

resources, necessary for every agent (its believes), D

includes methods stored in the “Task Ontology”

(agent desires), and I calls for intentions necessary

for every activity or task specification. There is also

a collaborator, in case there are two or more agents

needed to solve the task.

Private ontologies mapping is made through slots

of their components. So, the “Agent Ontology” has

four properties: has intentions - which contains

individuals of the methods M class from the “Task

Ontology”; has believes - which contains individuals

from the “Domain Ontology”; has desires - which

contains individuals from the “Task Ontology”; has

type - which contains variables of String type.

There is a real connection between the “Task

Ontology” (OT) and the “Domain Ontology” (OD)

through believes. The OT, in turn, refers to the

“Ontology of MAS” (OA), which is formally

described by four components. The first two are:

THE PROTÉGÉ - PROMETHEUS APPROACH TO SUPPORT MULTI-AGENT SYSTEMS CREATION

443

level value and order; contain values of Integer type,

which determine the logical level number and the

order of execution for every role. Roles (R) are the

individuals of the named ontology. The next two

properties: has input and has output refer to

individuals of “Interaction Ontology” (OI); in

particular, to protocols, which manage

communications. Their properties are of type String:

has scenario, language, roles at scenario.

The “Interaction Ontology” slots named has

initiator and has receiver are the individuals of the

“Agent Ontology” (Agent). Thus, agents are linked

to the proper protocols within the MAS. The OD, by

means of its individuals - which contain data records

- is connected to Agent , which uses the knowledge

on the domain area as its believes. This way, the

proposed meta-ontological model realized in Protégé

covers the first steps of the software development

life cycle.

3 SYSTEM DESIGN WITH

PROMETHEUS

DEVELOPMENT TOOL

In order to validate the next step of our approach, we

introduce a running example, consisting in an agent-

based decision support system (ADSS) dedicated to

monitoring environmental pollution information,

analyzing this data, simulating with respect to health

consequences and making decisions for responsible

system users. The general view of the Classes

belonging to the domain of interest includes regions,

which are characterized with some environmental

pollution, and morbidity level.

The accent is made on regions (Sokolova and

Fernández-Caballero, 2007). The MAS is a logical

three layer architecture. The first layer is aimed for

meta-data creation, the second one is responsible for

hidden knowledge discovery, and the third level

provides real-time decision support making, data

distribution and visualization. This architecture

satisfies all the required criteria for decision support

systems.

The first level is named Information fusion and it

acquires data from diverse sources, and pre-

processes the initial information to be ready for

further analysis. The second layer is named Data

Mining and there are three roles at this level,

dedicated to knowledge recovering through

modelling, and calculation impact of various

pollutants upon human health. The third level,

Decision Making, carries out a set of procedures

including model evaluation, computer simulation,

decision making and forecasting, based on the

models created in the previous level. The main

function of this level is to provide a user - actually, a

person who makes decisions - with the possibility to

run online real-time “what - if” scenarios. The end-

user, that is to say the person making decisions,

interacts with the MAS through a System-User

Interaction protocol, which is responsible for

human-computer interaction. The user chooses the

indicator that he wants to examine and initiates a

computer simulation.

The system resembles a typical organizational

structure. The agents are strictly dedicated to work

with the stated sets of data sources. They solve the

particular tasks and are triggered when all the

necessary conditions are fulfilled, and there are

positive messages from previously executed agents

(Weiss, 2000). The system includes a set of roles,

correlated with the main system functions and a set

of agents related to each role.

4 IMPLEMENTATION IN JACK

The system design in the Prometheus design tool

ends with generation of the skeleton code for JACK

Intelligent Agents™ that facilitates further coding,

testing and deployment.

The JACK Design Tool provides a visual

interface, which supports application creation

directly created in Jack Development Environment

imported from Prometheus. This last possibility

supports our proposal. Therefore, we propose to

implement the coding, testing and deployment stages

of the MAS development process within the JACK

Development Environment.

5 CONCLUSIONS

The integration of two existing and widely accepted

tools, Protégé Ontology Editor and Knowledge-Base

Framework, and Prometheus Development Kit, into

a common methodology has been introduced in this

paper. To provide the following stages with tools,

we have tested different methodologies, and finally

decided to use the Prometheus Development Tool,

which offers a wide range of possibilities for MAS

planning and implementation, namely the system

architecture, the system entities, their internals and

communications within the system and with outer

entities.

ICEIS 2008 - International Conference on Enterprise Information Systems

444

The integrated approach covers all the stages of

MAS planning and implementation, supporting them

with tools and frameworks. The proposed fusion of

methodologies, Protégé and Prometheus, was chosen

because of the wide range of functions offered and

their conformance to international standards.

However, other combinations of agent-oriented tools

could be used, whenever it helps getting the same

result and support during MAS development,

deployment and maintenance.

ACKNOWLEDGEMENTS

Marina V. Sokolova is the recipient of Postdoctoral

Scholarship 0000253836, Program II.E (Becas

MAE) awarded by the Agencia Española de

Cooperación Internacional of the Spanish Ministerio

de Asuntos Exteriores y de Cooperación.

This work is supported in part by the Spanish

Ministerio de Educacion y Ciencia TIN2004-07661-

C02-02 and TIN2007-67586-C02-02 grants, and the

Junta de Comunidades de Castilla-La Mancha

PBI06-0099 grant.

REFERENCES

Bellifemine, F., Poggi, A., Rimassa, G. (1999). Jade - A

FIPA-compliant agent framework. Proceedings of the

Practical Applications of Intelligent Agents, pp. 97-

108.

DeLoach, S.A., Wood, M.F., and Sparkman, C.H. (2001).

Multiagent systems engineering. International Journal

of Software Engineering and Knowledge Engineering,

11, 231-258.

de Wolf, T., Holvoet, T. (2005) Towards a full life-cycle

methodology for engineering decentralised multi-agent

systems. Proceedings of the Fourth International

Workshop on Agent-Oriented Methodologies, pp. 1-

12.

Georgeff, M., Pell, B., Pollack, M., Tambe, M.,

Wooldridge, M. (1998). The Belief-Desire-Intention

model of agency. Intelligent Agents V: Agent

Theories, Architectures, and Languages. Lecture Notes

in Computer Science, 1555, pp. 1-10.

Giunchiglia, F., Mylopoulos, J., Perini, A. (2002). The

Tropos software development methodology:

Processes, models and diagrams. Third International

Workshop on Agent-Oriented Software Engineering,

pp. 162-173.

Gorodetsky, V., Karsaev, O., Konushy, V., Mirgaliev, A.,

Rodionov, I., Yustchenko, S. (2005). MASDK

software tool and technology supported. International

Conference on Integration of Knowledge Intensive

Multi-Agent Systems, pp. 528-533.

Guarino, N., Giaretta, P. (1995), Ontologies and

knowledge bases: Towards a terminological

clarification. In: Towards Very Large Knowledge

Bases, IOS Press 1995, pp. 25-32.

Konichenko A.V. (2005). Distribution Information

Systems Design Management. Rostov-Don Press,

Russia.

Marík, V., McFarlane, D. (2005). Industrial adoption of

agent-based technologies. Intelligent Systems, 20, pp.

27–35.

Padgham, L., Winikoff, M. (2002). Prometheus: A

pragmatic methodology for engineering intelligent

agents. Proceedings of the Workshop on Agent

Oriented Methodologies (Object-Oriented

Programming, Systems, Languages, and

Applications), pp. 97-108.

Repast home page. (2003). http://repast.sourceforge.net.

Schelfthout, K., Holvoet, T. (2004). ObjectPlaces: an

environment for situated multi-agent systems.

Proceedings of the Third International Joint

Conference on Autonomous Agents and Multiagent

Systems, pp. 1500-1501.

Sokolova, M.V., Fernández-Caballero, A. (2007). A multi-

agent architecture for environmental impact

assessment: Information fusion, data mining and

decision making. 9th International Conference on

Enterprise Information Systems, ICEIS 2007, vol.

AIDSS, pp. 219-224.

Sokolova, M.V., Fernández-Caballero, A. (2007). An

agent-based decision support system for ecological-

medical situation analysis. 2nd International Work-

conference on the Interplay between Natural and

Artificial Computation. Lecture Notes in Computer

Science, 4528, pp. 511-520.

Vasconcelos, W.W., Robertson, D.S., Agusti, J., Sierra,

C., Wooldridge, M., Parsons, S., Walton, C., Sabater,

J. (2001). A lifecycle for models of large multi-agent

systems. Proceedings of the Second International

Workshop on Agent-Oriented Software Engineering.

Lecture Notes in Computer Science, 2222, pp. 297-

318.

Weiss, G. (2000). Multiagent Systems: A Modern

Approach to Distributed Artificial Intelligence. The

MIT Press.

Wooldridge, M., Jennings, N.R., Kinny, D. (2000). The

Gaia methodology for agent-oriented analysis and

design. Journal of Autonomous Agents and Multi-

Agent Systems, 3, 285-312.

THE PROTÉGÉ - PROMETHEUS APPROACH TO SUPPORT MULTI-AGENT SYSTEMS CREATION

445