A Comparison of Hill-Climbing Methods for Solving Static and Recurrent Dynamic Constraint Satisfaction Problems

In this paper we compare the performance of Minton’s et al. min-conflicts hill-climbing (MCHC) algorithm [8], Wallace and Freuder’s MCHC algorithm [18], Morris’ breakout algorithm (BA) [9, 10], and modified variants of the BA that we developed on static constraint satisfaction problems and recurrent dynamic constraint satisfaction problems (CSPs & rDCSPs). In this study, our results show that the BA and our modified variants of the BA are the most efficient algorithms in terms of average number of candidate solutions and average clock cylces on static CSPs. Our modified variants of the BA are the most efficient algorithms in terms of average solution instability, average number of candidate solutions, and average run-time in seconds on rDCSPs.