Ammar Lahlouhi

Ammar Lahlouhi

Systems integration is a difficult matter particularly when its components are varied. The problem becomes even more difficult when such components are heterogeneous such as humans, robots and software systems. Currently, the humans are regarded as users of artificial systems (robots and software systems). This has several disadvantages such as: (1) incoherence of artificial systems exploitation where humans' roles are not clear and (2) vain research of a user's universal model. In this paper, we adopted a cooperative approach where the system's components are regarded as being of the same level and they cooperate for the service of the global system. We concretized such approach by considering humans, robots and software systems as autonomous agents assuming roles in an organization. The latter will be implemented as a multi-agent system developed using a multi-agent development methodology.

Ammar Lahlouhi

This paper argues that modelling the development methodologies can improve the multi-agents systems software engineering. Such modelling allows applying methods, techniques and practices used in the software development to the methodologies themselves. The paper discusses then the advantages of the modelling of development methodologies. It describes a model of development methodologies, uses such a model to develop a system of their partial validation, and applies such a system to multi-agent methodologies. Several benefits can be gained from such modelling, such as the improvement of the works on the development, evaluation and comparison of multi-agent development methodologies.

Ammar Lahlouhi

The evolution of self-adaptive systems poses the problems of their coherence and the resume of the systems' functioning taking into account the accomplished work. While they are the base of the self-adaptive systems, these two aspects are not considered in the related works. In this paper, we propose a methodological based approach. In such approach, the adaptive system's evolution is thought at its model level where its execution is made on the system by exploiting a methodological process. For its concretization, we use colored Petri nets to describe the agents' individual tasks. To handle the system's functioning resume, we exploit the property of Petri nets on which the control flow depends on last marking only.