Complexity of Single-Agent and Equilibrium-Based Multiagent Planning and Plan Checking: Approach When We Have a List of Valid States

Authors

Chitta Baral
Vladik Kreinovich, The University of Texas at El PasoFollow

Publication Date

1-2003

Comments

Technical Report: UTEP-CS-03-05

Abstract

In a recent paper M. Bowling, R. Jensen, and M. Veloso proposed a new formalization of the problem of multiagent planning - a formalization that is based on the (intuitively natural) notion of an equilibrium. In this paper we analyze the computational complexity of the corresponding equilibrium-based multiagent planning and plan checking. Within the traditional approach in which states are described by fluents, the computational complexity of planning under incompleteness is PSPACE-hard already for a single agent; therefore, to make a meaningful comparison between the complexity of single-agent and multi-agent planning, we analyze complexity in a different approach, in which a list of valid states is assumed to be explicitly given.