Reverse Logistics Optimization Based on Parallel Genetic Algorithm

This paper is concerned with the efficient design of a reverse logistics network. The mixed integer linear models is formulated and determine how to optimize collection cost, transportation cost, fixed cost, variable cost, disposal cost, the sale revenue of reclaimed materials and revenue of selling two-hands products. GI/G/m queuing model presents uncertainty inherent to reverse logistics. In order to search for the optimal solution of this model, the paper proposes an algorithm based on the technique of Parallel Genetic Algorithm (PGA). To speed up the processing of generations of populations, parallel genetic algorithm splits the population into several sub-populations and run them in the parallel way. An example is solved by PGA and the result shows the algorithm has a rapid convergence rate.