Samir Aknine

A multi-agent method for forming and dynamic restructuring of pareto optimal coalitions

The first part of this paper presents a coalition formation method for multi-agent systems which finds a Pareto optimal solution without aggregating the preferences of the agents. This protocol is adapted to problems requiring coordination by coalition formation, where it is undesirable, or not possible, to aggregate the preferences of the agents. The second part proposes an extension of this method enabling dynamic restructuring of coalitions when changes occur in the system.

Coalition Formation Strategies for Self-Interested Agents

Coalition formation is a major research issue in multiagent systems in which the agents are self-interested. In these systems, agents have to form groups in order to achieve common goals, which they are not able to achieve individually. A coalition formation mechanism requires two definition levels: firstly agents need a common protocol to reach an agreement and secondly individual strategies are required to make efficient proposals. Both issues are addressed in this paper. First, we propose a two-phase decentralized protocol that allows agents to interact directly through message passing. Secondly we propose some strategies which allow agents to make clever proposals using the information that has already been collected from other agents. The experimental evaluation shows that the proposed mechanism allows agents to efficiently form coalitions and that the strategies make real improvements for the coalition search process.

Agent oriented conceptual modeling of parallel workflow systems

This research shows how to use ontology and conceptual models for the design of parallel workflow systems in a multi-agent perspective. To illustrate this work, we propose the example of cooperative writing of technical specifications in the telecommunications domain.

Managing Distributed Parallel Workflow Systems Using a Multi- Agent Method

In this chapter, we introduce new research directions for agent knowledge representation and multi-agent research and we enhance distributed parallel workflow systems (DPWS) with theory and tools developed in Artificial Intelligence. This article shows the limits of the traditional workflow systems and proposes a way of overcoming them through a new workflow model which parallelizes its activities. The first section defines the concepts of task, task execution validity, dependence and consistency between tasks. The second presents a multi-agent model which is used to manage the interactions between human agents, called actors, and artificial agents. It goes on to present a multiagent software agent architecture which addresses the needs of these systems.

Exploiting Knowledge Ontology for Managing Parallel WorkFlow Systems

In this article, we propose a framework for cooperation among human actors of a cooperative work system supported by software agents based on a common knowledge representation ontology. Ontology gives a background of knowledge to share among autonomous agents of a cooperative system and solves indirect conflicts between actors’ activities. We realised and tested our mediating system by an experimentation on cooperative writing process in the domain of telecommunications.

Strategies and Behaviours of Agents in Multi-phased Negotiations

This paper presents new multi-agent negotiation models for electronic commerce. These models address M-N-P negotiation problems, i.e. negotiations between m buyers and n sellers for buying p dependent products or services. These products or services are not necessarily provided by the same seller. We propose two new negotiation protocols as well as algorithms describing the behaviors of seller and buyer agents.

Coordination of Supply Chain Activities: A Coalition-Based Approach

Companies operate in an environment increasingly demanding in terms of flexibility and reactivity. The introduction of the entities resulting from Distributed Artificial Intelligence (DAI) and Multi-Agent Systems (MAS) in the management of enterprises prove to be an interesting technology to simulate and reproduce the collaborative and adaptive behaviors of enterprises. This article models the coordination of the various collaborative parties both inside and outside a supply chain using coordination methods of MAS mainly coalition formation mechanisms. In this paper, we present our agent modeling of supply chains, and then we detail the coalition formation algorithm. Lastly, we illustrate our approach with an example chosen in the industrial domain.

Gael: a generic multi-agent meta-programming method

We propose a generic multi-agent collective metaprogramming language and method. Nowadays, a fair amount of research work is learning oriented, but there is none which defines an operational collective (multi-agent) meta-programming method and language. Our method (Gael) is an extension of the explanation-based learning method (EBL) suggested by (Minton et al., 1989).

Exploiting Social Reasoning of Open Multi-agent Systems to Enhance Cooperation in Hospitals

The Virtual Hospital Patient Scheduling System (VHPSS) focuses on building multi-agent cooperative systems. We have chosen to build Intelligent agents that perform coordination tasks for the users, i.e. the medical staff. An agent in VHPSS system has a limited information. To solve some problems, the agent has to cooperate with the other agents of the surrounding environment. This article presents this scheduling system based on our principal contributions in research on multi-agent systems. (1) At the architectural level, the system is based on a multi-agent software agent architecture which has several advantages: parallelization of agents tasks, reusability of agents components, partial mobility and partial cloning of agents code. (2) At the cooperative level, the system uses a new agent negotiation protocol making it possible to accelerate the process of task allocation.