Crossover operator of continuous GA with cost information

Genetic algorithm (GA) is the most famous kind of the evolutionary algorithms (EA). Similar to other EAs, it uses population to search for the global minimum on the optimal plate. It has three main operators: selection, reproduction and mutation. Fathers and mothers are selected from previous generation by the selection operator to breed the new individuals by the reproduction operator. Then, mutation operates and produces new attributes on offspring. In GA, reproduction operator is known by as the crossover operator. Many kinds of crossover operators have been introduced up to now. Almost all of them use coordinate of parents to determine the location of new individuals, but the cost information of parents has not been considered yet. By adding cost information of parents, the algorithm will be able to produce better points. Parent with low cost tell us that its district is not near to the global minimum, so offspring must be prevented from getting close to that locations. Inversely, locations of the parents who have good costs are probably nearer to the destination. Therefore, algorithms must steer offspring toward parents with suitable cost and prevent them from getting close to other parent´s locations. This is what has been supposed and implemented in this paper. In this paper, a new crossover method is proposed and it is compared with other introduced crossover methods on some well-known cost functions. The results show capability of new method.