A crossover in simulated annealing for population initialization of genetic algorithm to optimize the distribution cost

Solving distribution problems have been an alluring topic for some academician. The determination of proper distribution network to provide a minimal cost is still difficult to resolve. This is because there are some difficult constraints to be addressed. As an algorithm, which typically offers a set of solutions in solving the problems, genetic algorithms (GA) has verified its power in solving complex combinatorial problems. The generation of a set of initial solutions (population) generally performed randomly in GA. In the large cases, it is becoming one of the drawbacks since the search space becomes too wide, so the probability to get stuck in a local optimum solution is also high. Therefore, simulated annealing (SA) is employed to generate the initial population for the GA. SA has been selected since it is able to avoid a local optimum solution. In this study, the process of finding new solutions using SA is improved by using the crossover process, which is commonly used in GA. This method has become novel because the crossover has the same principle of providing varied new solutions that still retain some of the properties of the parent solution. The result of the modification SA-GA proven to provide superior results than the existing algorithms