Diversify Intensification Phases in Local Search for SAT with a New Probability Distribution

A key challenge in developing efficient local search solvers is to intelligently balance diversification and intensification. This study proposes a heuristic that integrates a new dynamic scoring function and two different diversification criteria: variable weights and stagnation weights. Our new dynamic scoring function is formulated to enhance the diversification capability in intensification phases using a user-defined diversification parameter. The formulation of the new scoring function is based on a probability distribution to adjust the selecting priorities of the selection between greediness on scores and diversification on variable properties. The probability distribution of variables on greediness is constructed to guarantee the synchronization between the probability distribution functions and score values. Additionally, the new dynamic scoring function is integrated with the two diversification criteria. The experiments show that the new heuristic is efficient on verification benchmark, crafted and random instances.

[1]  Marijn Heule,et al.  EagleUP: Solving Random 3-SAT Using SLS with Unit Propagation , 2011, SAT.

[2]  Dongmo Zhang,et al.  AI 2012: Advances in Artificial Intelligence , 2012, Lecture Notes in Computer Science.

[3]  Bart Selman,et al.  Evidence for Invariants in Local Search , 1997, AAAI/IAAI.

[4]  Adrian Balint,et al.  Improving Stochastic Local Search for SAT with a New Probability Distribution , 2010, SAT.

[5]  Hector J. Levesque,et al.  A New Method for Solving Hard Satisfiability Problems , 1992, AAAI.

[6]  Holger H. Hoos,et al.  An adaptive noise mechanism for walkSAT , 2002, AAAI/IAAI.

[7]  Abdul Sattar,et al.  Weight-Enhanced Diversification in Stochastic Local Search for Satisfiability , 2013, IJCAI.

[8]  Abdul Sattar,et al.  Combining Adaptive and Dynamic Local Search for Satisfiability , 2008, J. Satisf. Boolean Model. Comput..

[9]  Abdul Sattar,et al.  A Method to Avoid Duplicative Flipping in Local Search for SAT , 2012, Australasian Conference on Artificial Intelligence.

[10]  Abdul Sattar,et al.  A Study of Local Minimum Avoidance Heuristics for SAT , 2012, ECAI.

[11]  John Thornton,et al.  Additive versus Multiplicative Clause Weighting for SAT , 2004, AAAI.

[12]  Harry Zhang,et al.  A Switching Criterion for Intensification and Diversification in Local Search for SAT , 2008, J. Satisf. Boolean Model. Comput..

[13]  Ofer Strichman,et al.  Theory and Applications of Satisfiability Testing – SAT 2010 , 2010, Lecture Notes in Computer Science.

[14]  Kaile Su,et al.  Configuration Checking with Aspiration in Local Search for SAT , 2012, AAAI.

[15]  Steven David Prestwich,et al.  Random Walk with Continuously Smoothed Variable Weights , 2005, SAT.

[16]  Karem A. Sakallah,et al.  Theory and Applications of Satisfiability Testing - SAT 2011 - 14th International Conference, SAT 2011, Ann Arbor, MI, USA, June 19-22, 2011. Proceedings , 2011, SAT.

[17]  Bart Selman,et al.  Domain-Independent Extensions to GSAT : Solving Large StructuredSatis ability , 1993 .

[18]  Abdul Sattar,et al.  Trap Avoidance in Local Search Using Pseudo-Conflict Learning , 2012, AAAI.