Cost Based Satisficing Search Considered Harmful

Recently, several researchers have found that cost-based satisficing search with A* often runs into problems. Although some "work arounds" have been proposed to ameliorate the problem, there has not been any concerted effort to pinpoint its origin. In this paper, we argue that the origins can be traced back to the wide variance in action costs that is observed in most planning domains. We show that such cost variance misleads A* search, and that this is no trifling detail or accidental phenomenon, but a systemic weakness of the very concept of "cost-based evaluation functions + systematic search + combinatorial graphs". We show that satisficing search with sized-based evaluation functions is largely immune to this problem.

[1]  Nils J. Nilsson,et al.  A Formal Basis for the Heuristic Determination of Minimum Cost Paths , 1968, IEEE Trans. Syst. Sci. Cybern..

[2]  Ira Pohl,et al.  The Avoidance of (Relative) Catastrophe, Heuristic Competence, Genuine Dynamic Weighting and Computational Issues in Heuristic Problem Solving , 1973, IJCAI.

[3]  Bart Selman,et al.  Unifying SAT-based and Graph-based Planning , 1999, IJCAI.

[4]  Geoffrey J. Gordon,et al.  Bounded real-time dynamic programming: RTDP with monotone upper bounds and performance guarantees , 2005, ICML.

[5]  Malte Helmert,et al.  How Good is Almost Perfect? , 2008, AAAI.

[6]  Rina Dechter,et al.  Generalized best-first search strategies and the optimality of A* , 1985, JACM.

[7]  Avrim Blum,et al.  Fast Planning Through Planning Graph Analysis , 1995, IJCAI.

[8]  Scott Sanner,et al.  Bayesian Real-Time Dynamic Programming , 2009, IJCAI.

[9]  Subbarao Kambhampati,et al.  G-Value Plateaus: A Challenge for Planning , 2010, ICAPS.

[10]  Wheeler Ruml,et al.  Using Distance Estimates in Heuristic Search , 2009, ICAPS.

[11]  Edsger W. Dijkstra,et al.  Go to Statement Considered Harmful (Reprint) , 2002, Software Pioneers.

[12]  Stefan Edelkamp,et al.  Optimal Symbolic Planning with Action Costs and Preferences , 2009, IJCAI.

[13]  Nils J. Nilsson,et al.  Correction to "A Formal Basis for the Heuristic Determination of Minimum Cost Paths" , 1972, SGAR.

[14]  Wheeler Ruml,et al.  Finding Acceptable Solutions Faster Using Inadmissible Information , 2010, SOCS.

[15]  Richard E. Korf,et al.  Depth-First Iterative-Deepening: An Optimal Admissible Tree Search , 1985, Artif. Intell..

[16]  Andrew G. Barto,et al.  Learning to Act Using Real-Time Dynamic Programming , 1995, Artif. Intell..

[17]  Stephen F. Smith,et al.  New Techniques for Algorithm Portfolio Design , 2008, UAI.

[18]  Silvia Richter,et al.  The LAMA Planner: Guiding Cost-Based Anytime Planning with Landmarks , 2010, J. Artif. Intell. Res..

[19]  Bernhard Nebel,et al.  The FF Planning System: Fast Plan Generation Through Heuristic Search , 2011, J. Artif. Intell. Res..

[20]  Blai Bonet,et al.  Labeled RTDP: Improving the Convergence of Real-Time Dynamic Programming , 2003, ICAPS.

[21]  Michael Kearns,et al.  Near-Optimal Reinforcement Learning in Polynomial Time , 1998, Machine Learning.

[22]  Edsger W. Dijkstra,et al.  A note on two problems in connexion with graphs , 1959, Numerische Mathematik.

[23]  Judea Pearl,et al.  Heuristics : intelligent search strategies for computer problem solving , 1984 .