Personal tools
Laboratoire d'InfoRmatique en Image et Systèmes d'information

Skip to content. | Skip to navigation

Laboratoire d'InfoRmatique en Image et Systèmes d'information
UMR 5205 CNRS / INSA Lyon / Université Claude Bernard Lyon 1 / Université Lumière Lyon 2 / École Centrale de Lyon
You are here: Home > membres

Ndeye-Arame Diago

photo

PhD student

Team Multi-Agent Systems
 
Institution Claude Bernard University of Lyon 1
Location Nautibus (Université Lyon1)
 
E-mail diago.ndeye-arame at liris.cnrs.fr
Contact details Publications Thesis
 
Subject Coordination models in multi agent systems
Abstract Advancements in technology have enabled the emergence of intelligent systems for the benefit of the human. Intelligent systems are well known due to their ability to emulate a specific human behavior. For example, there is a wide range of applications for smart homes, health-care and in transport systems. However, artificial intelligence presents some limitations in the coordination of expertise from different systems in order to emulate behaviors which are more complex. Thus, it relies on fields such as distributed artificial intelligence, distributed systems and software engineering. This constitutes a novel research field called multi-agent systems.
Research in multi-agent systems can be oriented in different areas such as the simulation or modeling of a system, complex problem solving and applications design. Computer simulation is nowadays well developed and it is used in any discipline of science. It enables to observe and analyze the dynamic interaction of the different entities of a complex system. Multi-agent systems are used to model the exact behavior of each of these entities and its interactions. Its purpose is to study the dynamic evolution of the system in order to predict its final state.
Multi-agent approach takes all its interest in modeling and studying the decision-making processes that are at the heart of our daily activities, of any government, non-governmental organization and companies. Today, all of these entities encounter decision-making problems related to the complexity of their organization, the conflicts of interest from decision-makers, the presence of several alternative solutions and the difficulties to consider all consequences of their choices. Today, many companies and research organizations are involved in the study and development of strategies and decision-making techniques by calling on different fields such as economics, sociology, technology, etc. Multi-agent systems enable the modeling and simulation of decision-making mechanisms. They make it possible to understand the complexity of these mechanisms and aim to provide decision support tools. In the context of this thesis, we address the key challenges of multilateral negotiation for collective decision-making.
The objective of this thesis is to propose multilateral negotiation mechanisms in a context where the agents are selfish but need to negotiate to achieve a collective solution. Agents’ interaction is fully decentralized and occurs simultaneously. In other words, there is no mediator. The agents make their decisions according to their own beliefs and desires. They are also free to participate in the negotiation at any time. Devising such negotiation mechanism requires specific protocols that can guide the agents in their interaction process. So we propose negotiation mechanisms based on an incremental approach in order to limit agents reasoning complexity and thus to facilitate the convergence of the negotiation and to limit the time.
Advisor Samir Aknine

Last update : 2018-03-02 09:02:36