Improving Plan Quality in SAT-Based Planning

Planning as Satisfiability (SAT) is the best approach for optimally (wrt makespan) solving classical planning problems. SAT-based planners, like satplan , can thus return plans having minimal makespan guaranteed. However, the returned plan does not take into account plan quality issues introduced in the last two International Planning Competitions (IPCs): such issues include minimal-actions plans and plans with "soft" goals, where a metric has to be optimized over actions/goals. Recently, an approach to address such issues has been presented, in the framework of planning as satisfiability with preferences: by modifying the heuristic of the underlying SAT solver, the related system (called satplan(P) ) is guaranteed to return plans with minimal number of actions, or with maximal number of soft goals satisfied. But, besides such feature, it is well-known that introducing ordering in SAT heuristics can lead to significant degradation in performances. In this paper, we present a generate-and-test approach to tackle the problem of dealing with such optimization issues: without imposing any ordering, a (candidate optimal) plan is first generated, and then a constraint is added imposing that the new plan (if any) has to be "better" than the last computed, i.e., the plan quality is increased at each iteration. We implemented this idea in satplan , and compared the resulting systems wrt satplan(P) and SGPlan on planning problems coming from IPCs. The analysis shows performance benefits for the new approach, in particular on planning problems with many preferences.

[1]  Bart Selman,et al.  Planning as Satisfiability , 1992, ECAI.

[2]  Enrico Giunchiglia,et al.  SAT-Based Planning with Minimal-#actions Plans and "soft" Goals , 2007, AI*IA.

[3]  Yixin Chen,et al.  MaxPlan : Optimal Planning by Decomposed Satisfiability andBackward Reduction , 2006 .

[4]  Joost P. Warners,et al.  A Linear-Time Transformation of Linear Inequalities into Conjunctive Normal Form , 1998, Inf. Process. Lett..

[5]  Yixin Chen,et al.  Plan-A : A Cost Optimal Planner Based on SAT-Constrained Optimization ∗ , 2008 .

[6]  Patrik Haslum,et al.  Deterministic planning in the fifth international planning competition: PDDL3 and experimental evaluation of the planners , 2009, Artif. Intell..

[7]  Craig Boutilier,et al.  CP-nets: a tool for represent-ing and reasoning with conditional ceteris paribus state-ments , 2004 .

[8]  S. Edelkamp,et al.  The Deterministic Part of IPC-4: An Overview , 2005, J. Artif. Intell. Res..

[9]  Hector Geffner,et al.  Heuristics for Planning with Action Costs , 2007, CAEPIA.

[10]  Ronen I. Brafman,et al.  Planning with Goal Preferences and Constraints , 2005, ICAPS.

[11]  Marco Gavanelli,et al.  An Algorithm for Multi-Criteria Optimization in CSPs , 2002, ECAI.

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

[13]  Jussi Rintanen,et al.  Satisfiability Planning with Constraints on the Number of Actions , 2005, ICAPS.

[14]  Claudette Cayrol,et al.  Using the Davis and Putnam Procedure for an Efficient Computation of Preferred Models , 1996, ECAI.

[15]  Enrico Giunchiglia,et al.  Planning as Satisfiability with Preferences , 2007, AAAI.

[16]  Enrico Giunchiglia,et al.  A new Approach for Solving Satisfiability Problems with Qualitative Preferences , 2008, ECAI.

[17]  Ilkka Niemelä,et al.  Unrestricted vs restricted cut in a tableau method for Boolean circuits , 2005, Annals of Mathematics and Artificial Intelligence.