Genetic algorithm crossover operators for ordering applications

In this paper, we compare the performance of several crossover operators, including two new operators and a new faster formulation of a previously published operator. This new formulation performs better than the other operators we have tested while taking no more computation time. In addition, with practical applications in mind, we show how the use of problem specific information can improve the performance of the Genetic Algorithm and we describe a method for designing problem specific crossover incorporating a novel tie-breaking algorithm.