Use of formal languages to consolidate a Holonic MAS methodology: a specification approach for analysing Problem and Agency domains

ion step is not amenable to formal proof. However, it is possible to validate a specification by proving challenging theorems concerning properties that the specification is expected to exhibit. If correct, these theorems reinforce the specifier’s understanding of the specification and its relationship with the underlying Problem Domain. If not, the specification probably needs to be changed to better reflect the understanding domain of those involved in producing (and implementing) the specification. In the specification domain, there are several methodologies to help the modelling and analysis phase of MultiAgent Systems (MAS) and Holonic systems. Among these methodologies, we point out the well known: Tropos (Giorgini, 1995), PASSI (Azaiez, 1992) and Agent-oriented Software Process for Engineering Complex Systems (ASPECS) (Object Management Group, 2003; Gaud, 2007). ASPECS uses UML as a semi-formal modeling language and consequently it makes this meta-model ambiguous. In D’inverno and Luck (2003), a principled theory of agency is developed by describing just such a framework, called the SMART agent framework. Using Z specification formal language, a sophisticated model of agent and their relationships is built up and illustrated with some applications. They demonstrated that Z language is well suited to model data structures and functionalities in a highly abstract fashion but do not treat the behavioral aspect. In this work, we use our specification language called PNOZ, based on two formalisms: Petri Nets (PN) and Object-Z (OZ). Such a specification style facilitates the modeling of systems with functional and behavioral aspects. The objective of this work consists of consolidating the ASPECS methodology by using our formal specification and analysis and validates such a specification with the framework SAL (Symbolic Analysis *M Garoui, Department of Computer Science, Monastir 5000, Tunisia. Journal of Simulation (2013), 1–11 r 2013 Operational Research Society Ltd. All rights reserved. 1747-7778/13 www.palgrave-journals.com/jos/ Laboratory). This consolidation is done by formalizing the core concepts of the first two domains (Problem and Agency Domain) of ASPECS meta-model. After a brief presentation of the simulator for the FIRA Robot Soccer competition, a quick overview of the ASPECS process and modelling approach will be presented in Section 3. Section 4 presents formal specifications of the first two domains (Problem and Agency Domains) of ASPECS process based on composition of PN and OZ language named PNOZ. Section 5 presents the validation process of PNOZ specification with the framework SAL. Finally, Section 6 summarises the results of the paper and describes some future work directions. 2. Case study: FIRA Robot Soccer This case study intends to model a simulator for the FIRA Robot Soccer competition (FIRA Robot SoccerASPECSWiki). This competition involves two teams of five autonomous robots playing a game similar to soccer (Figure 1). This is a classical example where real-time coordination is required. It constitutes a well-known benchmark for several research fields, such as MAS, image processing and control. Robot soccer players are two wheel-driven small mobile robots, which are controlled by a host computer. A soccer team in the category of MiroSOT (Micro Robot World Cup Soccer Tournament) consists of five players, one goal keeper and four players for each team. There are several game categories. The size and form of the robot in each category is fixed by rules. Robot soccer competitions give an opportunity to foster intelligent techniques and intelligent robotics research by providing a standard problem where a wide spectrum of technologies can be developed, tested and integrated, for example, collaborative multiple agent robotics, autonomous computing, real-time reasoning and sensor fusion. From a scientific point of view, the robot soccer players are ‘intelligent, autonomous agents’. They have a global goal like ‘win the game’. Robot soccer aims at promoting the developments in small autonomous robots and intelligent system (agent) that can cooperate with each other. 3. A quick overview of the used ASPECS process As proposed by Gaud (2007) and Cossentino et al (2007), ASPECS is a step-by-step requirement to code software engineering process based on a meta-model, which defines the main concepts for the proposed Holonic Multi-Agent Systems (HMAS) analysis, design and development based on Capacities, Role, Interaction and Organization (CRIO) framework. The target scope for the proposed approach can be found in complex systems, and especially in hierarchical complex ones. The main vocation of ASPECS is towards the development of societies of holonic (as well as nonholonic) MAS. ASPECS has been built by adopting the Model-driven architecture (Object Management Group, 2003). In Cossentino et al (2010), they label the three metamodels ‘domains’ thus maintaining the link with the PASSI metamodels. The three definite fields are: The Problem Domain: It provides the organizational description of the problem independently of a specific solution. The concepts introduced in this domain are used mainly during the analysis phase and at the beginning of the design phase. The Agency Domain: It introduces agent-related concepts and provides a description of the holonic, multiagent solution resulting from a refinement of the Problem Domain elements. The Solution Domain: It is related to the implementation of the solution on a specific platform Janus Project; (Gaud et al, 2008). This domain is thus dependent on a particular implementation and deployment platform. A downfall in ASPECS is that it uses UML as a modelling language. Because of the specific needs of agents and holonic organizational design, the UML semantics and notation are used as reference points, which makes ASPECS meta-model ambiguous. Our contribution will relate to the consolidation of ASPECS process by formalizing the Problem Domain and the Agency Domain associated with the ASPECS metamodel, therefore facilitating the Solution Domain.