Distributed Decision-Making and TaskCoordination in Dynamic, Uncertain andReal-Time Multiagent Environments

Decision-making in uncertainty and coordination are at the heart of multiagent systems. In this kind of systems, agents have to be able to perceive their environment and take decisions while considering the other agents. When the environment is partially observable, agents have to be able to manage this uncertainty in order to take the most enlightened decisions they can based on the incomplete information they have acquired. Moreover, in the context of cooperative multiagent environments, agents have to coordinate their actions in order to accomplish complex tasks requiring more then one agent. In this thesis, we consider complex cooperative multiagent environments (dynamic, uncertain and real-time). In this kind of environments, we propose an approach of decision-making in uncertainty that enable the agents to flexibly coordinate themselves. More precisely, we present an online algorithm for partially observable Markov decision processes (POMDPs). Furthermore, in such complex environments, agent’s tasks can also become quite complex. In this context, it could be complicated for the agents to determine the required number of resources to accomplish each task. To address this problem, we propose a learning algorithm to learn the number of resources necessary to accomplish a task based on the characteristics of this task. In a similar manner, we propose a scheduling approach enabling the agents to schedule their tasks in order to maximize the number of tasks that could be accomplish in a limited time. All these approaches have been developed to enable the agents to efficiently coordinate all their complex tasks in a partially observable, dynamic and uncertain multiagent environment. All these approaches have demonstrated their effectiveness in tests done in the RoboCupRescue simulation environment.