Efficient solutions for Mastermind using genetic algorithms

We present a new genetic algorithm for playing the game of Master-mind. The algorithm requires low run-times and results in a low ex-pected number of guesses. Its performance is comparable to that of other meta-heuristics for the standard setting with four positions and six colors, while it outperforms the existing algorithms when more col-ors and positions are examined. The central idea underlying the algo-rithm is the creation of a large set of eligible guesses collected through-out the different generations of the genetic algorithm, the quality of each of which is subsequently determined based on a comparison with a selection of elements of the set