Acceleration of genetic algorithms for Sudoku solution on many-core processors

Abstract

In this chapter, we use the problem of solving Sudoku puzzles to demonstrate the possibility of achieving practical processing time through the use of many-core processors for parallel processing in the application of evolutionary computation. To increase accuracy, we propose a genetic operation that takes building-block linkage into account. As a parallel processing model for higher performance, we use a multiple-population coarse-grained genetic algorithm (GA) model to counter initial value dependence under the condition of a limited number of individuals. The genetic manipulation is also accelerated by the parallel processing of threads. In an evaluation using even very difficult problems, we show that execution times of several tens of seconds and several seconds can be obtained by parallel processing with the Intel Core i7 and NVIDIA GTX 460, respectively, and that a correct solution rate of 100% can be achieved in either case. In addition, genetic operations that take linkage into account are suited to fine-grained parallelization and thus may result in an even higher performance.