Complexity results for blocks-world planning

  • Authors:

  • Naresh Gupta;Dana S. Nau

  • Affiliations:

  • Computer Science Department, University of Maryland, College Park, MD;Computer Science Department and Systems Research Center, University of Maryland, College Park, MD

  • Venue:
  • AAAI'91 Proceedings of the ninth National conference on Artificial intelligence - Volume 2
  • Year:
  • 1991

Quantified Score

Hi-index 0.00

Visualization

Abstract

Although blocks-world planning is well-known, its complexity has not previously been analyzed, and different planning researchers have expressed conflicting opinions about its difficulty. In this paper, we present the following results: 1. Finding optimal plans in a well-known formulation of the blocks-world planning domain is NP-hard, even if the goal state is completely specified. 2. Classical examples of deleted-condition interactions such as Sussman's anomaly and creative destruction are not difficult to handle in this domain, provided that the right planning algorithm is used. Instead, the NP-hardness of the problem results from difficulties in determining which of several different actions will best help to achieve multiple goals.