A Linear Weight Transfer Rule for Local Search

The Divide and Distribute Fixed Weights algorithm (ddfw) is a dynamic local search SAT-solving that transfers weight from satisfied to falsified clauses in minima. ddfw remarkably effective on several hard combinatorial instances. Yet, despite its success, it has received little study since debut 2005. In this paper, we propose three modifications the base algorithm: linear transfer method moves amount of between minima, an adjustment how are chosen minima give weight, weighted-random selecting variables flip. We implemented our top solver yalsat. Our experiments show boost performance compared original multiple benchmarks, including those past years SAT competitions. Moreover, improved exclusively solves instances refute conjecture lower bound two Van der Waerden numbers set forth by Ahmed et al. (2014), performs well graph-coloring instance been open for over decades.