The effects of two new crossover operators on genetic algorithm performance

In this study, two new crossover operators in genetic algorithm are proposed; sequential and random mixed crossover. In the first stage, existing and developed crossover operators were applied to two benchmark problems, namely the reinforced concrete beam problem and the space truss problem. In the second stage, the developed crossover operators were applied to a deep beam problem and, a concrete mix design problem. The fittest values obtained using developed crossover operators were higher than those were obtained with existing crossover operator after 30,000 generations. Moreover, in developed crossover operators, the random mixed crossover yields higher fitness values than the existing crossover operators.