Development and analysis of genetic algorithms: Sudoku case study

Abstract

This paper discusses the design and subsequent analysis of software implementing a configurable genetic algorithm. The genetic algorithm is primarily targeted towards the solving of Sudoku puzzles. Sudoku is regarded as an ideal test-bed for algorithm development due to the fact that it is a constrained optimisation problem that belongs to the NP-complete class of computational problems. The aim of this paper is to outline the various features currently implemented in the software, and to present preliminary results of an analysis of various aspects of the underlying genetic algorithm.