Plan Merging in Multi-Agent Systems
Mathijs M. De Weerdt · 2003
This volume deals with multi-agent based systems in transport planning. Existing methods are described that are helpful in goal attainment in planning, and how these methods have been extended and improved upon to deal with multiple agents is detailed. A basic multi-agent planning problem is defined and results are analyzed. A formal framework for dealing with both resources and multiple agents is presented, and the semantics and properties of this framework are studied. An algorithm, called Plan Merging, is developed that analyzes potential inefficiencies that can be resolved by combining the plans of the agents. The effect of taxi sharing for a taxi company is examined as a case example.