Genetic algorithms for optimization of multi-level product distribution

The process of wide product distribution to some areas consumes a very high cost. By minimizing the cost of the distribution process, companies can increase their profits. As algorithms that stochastically provide multiple diverse solutions, Genetic Algorithms (GAs) were proposed to solve the complex problems of multi-level distribution. To prevent premature convergence, preliminary numerical experiments were conducted to obtain the best parameters of GAs. Computational results of GAs that were adjusted by parameter testing were compared with the results of Random Search (RS) computation. The results indicated that there was an enormous difference in costs that resulted from computing GAs compared to RS in multi-level distribution problems.