Fuzzy-GA and multi-objective transportation optimization

Abstract

The spanning tree-based genetic algorithm (ST-GA) is an effective method to solve the multi-objective transportation optimization problem. The performance of this algorithm is definitely superior to the typical matrix-based GA. A new fuzzy-GA is proposed in this paper. At first, it adopts the Pareto method, so that the problem that the non-convex solutions are difficult to be found through common aggregation function method could be avoided; Moreover, the theory of fuzzy rules which can easily express explicit knowledge is introduced and merged in the ST-GA. Better Pareto front and Pareto optimal solutions are found in applications. Therefore, it displays that the new algorithm not only has better practicability but also has stronger intelligence. From the example, It can be found that the new algorithm is better then the ST-GA in performance.