Evolutionary Algorithm of Port Based Location Routing Problem

Location routing problem (LRP) is a significant subject in logistics systems, and genetic algorithm can obtain the near optimum solutions of large scale nonlinear mixed integer programming models which are NP-hard in nature. In the process of algorithm design, two-dimensional chromosomes that satisfy constraints automatically and their corresponding genetic arithmetic operators are designed, such as selection, crossover, mutation of seeds, re-insert and so on. Three-layer process of genetic evolution is conducted and examined by a group of random experiments. This paper compares the precision, effectiveness and applicable scope between the proposed algorithm and current optimizing software. With increase number of variables and constrains, computing time and iterations of genetic algorithm increase almost linearly, while that of the current software LINGO present NP-hard. Results show the proposed genetic algorithm is effective and efficient in solving the LRP problem especially of large scale.