Maryam Morady - Msc. Graduated Financial management, Ershad Damavand University, Tehran, Iran.

The model presented in this research is multi-product and multi-rating, which simultaneously covers transportation costs and facility construction. The model is NP-Hard, considering the minimization of costs (costs of deployment of facilities and shipping costs) as well as considering the uncertainty in the demand for returned products, in which the problem solving time is exponential and increases due to the dimensions of the problem; thus, in this research, the particle mass and genetic algorithm are used for solving the model. The results of the model solving showed that, in all cases of small size, the value of the target function achieved by the particle mass algorithm is better than the target function value obtained by the genetic algorithm. Also, the gap between the values of the target function is negative due to the goodness of the target function derived from the PSO algorithm. On the other hand, the gap between the values obtained from the two algorithms is very high, which indicates the high ability and power of the mass particle algorithm compared to the genetic algorithm in achieving the near optimal solution. The results from the implementation of problems of medium and large size showed that in all cases of medium and large size, such as issues of small size, the target function achieved by the particle mass algorithm is better than the target function obtained by the genetic algorithm.

Reverse Logistics, Mixed Integer Planning, Genetic Algorithm, Particle Mass Algorithm.