A Variable Neighborhood Search for Solving Sudoku Puzzles

摘要

Sudoku is a well-known puzzle that has achieved international popularity in the latest decade. Recently, there are explosive growths in the application of metaheuristic algorithms for solving Sudoku puzzles. In this paper, an algorithm based on Variable Neighborhood Search (VNS) is proposed to solve the Sudoku problem and the details of the implementation such as problem representation, neighbourhood structures are explained. The proposed algorithm is tested with Sudoku benchmarks which have been used in previous studies. The experimental results indicate that VNS is able to produce competitive results in easy level puzzles and promising results in medium and hard level puzzles.