论文信息 - Soft Arc Consistency Applied to Optimal Planning

Soft Arc Consistency Applied to Optimal Planning

We show in this article how the Weighted CSP framework can be used to solve an optimisation version of numerical planning. The WCSP finds an optimal plan in the planning graph containing all solution plans of minimum length. Experimental trials were performed to study the impact of soft arc consistency techniques (FDAC and EDAC) on the efficiency of the search for an optimal plan in this graph. We conclude by giving a possible theoretical explanation for the fact that we were able to solve optimisation problems involving several hundred variables.

Martin C. Cooper | Sylvain Cussat-Blanc | Pierre Régnier | Marie de Roquemaurel

[1] Thomas Schiex,et al. On the Complexity of Compact Coalitional Games , 2009, IJCAI.

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

[3] Jörg Hoffmann,et al. The Metric-FF Planning System: Translating ''Ignoring Delete Lists'' to Numeric State Variables , 2003, J. Artif. Intell. Res..

[4] Martin C. Cooper. Reduction operations in fuzzy or valued constraint satisfaction , 2003, Fuzzy Sets Syst..

[5] Simon de Givry,et al. Existential arc consistency: Getting closer to full arc consistency in weighted CSPs , 2005, IJCAI.

[6] Subbarao Kambhampati,et al. Planning as constraint satisfaction: Solving the planning graph by compiling it into CSP , 2001, Artif. Intell..

[7] Hector Geffner,et al. Heuristic Planning with Time and Resources , 2014 .