Planning in multiagent systems by Mathijs Weerdt and Cees Witteveen

Course description:

By definition agents are autonomous entities that are able to act.  Hence, the process of determining which actions to execute and the order in  which these actions have to be executed is considered as an essential property of  agents. Within the AI-community such processes have been studied in the  context of planning problems. The major issue in multi-agent  planning is the coordination of single-agent planners.

In this tutorial we start with an overview of single-agent AI planning techniques in the framework of refinement planning. After that, we  present some multi-agent planning problems and discuss the properties of these  problems that have the most influence on the applicability of certain existing  solutions. We will give an overview of these approaches by distinguishing their  role in the phases of the planning and coordination process for a multi-agent  system. Finally, some of these approaches will be discussed in more detail.

Tutor information:

Mathijs de Weerdt completed his Master's in computer science at  the Utrecht  University supervised by prof.dr. JJ.Ch. Meyer, prof.dr. W. van  der Hoek, and dr. F. de Boer in 1998 (cum laude). Subsequently, he started  studying multi-agent systems and inter-organizational coordination as a PhD  student at the Delft University of Technology supervised by dr. C.  Witteveen and
prof.dr. H.J. Sips. In 2003, Mathijs successfully defended his PhD  thesis on "Plan Merging in Multi-Agent Systems". He worked at the research  company Almende for a couple of months and currently he is appointed as  assistant professor at the Delft University of Technology.

Cees Witteveen studied psychology and mathematics at Utrecht  University. He completed his PhD-thesis "Programmed Production  Systems" supervised by prof.dr. P. van Emde Boas and prof.dr.ir. R.  Stobberingh at Utrecht University and then joined the department of  Computer Science at Delft University of Technology. Since 1998, he is  an associate professor heading the CABS (Collective Agent Based  Systems) group within the department of Software Technology.
His current interests include planning (theories and applications),  algorithms for coordinating autonomous systems, distributed diagnosis  and complexity theory.