Fuzzy multi-population cooperative GA and MOT optimization

In the multi-objective transportation (MOT) optimization problems, it is quite necessary to consider the tradeoff between all conflictive sub-objectives, consequently it leads to difficulties in solving. So this paper proposed a new Fuzzy Multi-Population Cooperative Genetic Algorithm, called fmc-GA. We firstly infuse the combination of fuzzy rule, which is convenient to express the explicit knowledge, and the greedy algorithm into the spanning tree decoding, sequentially enhancing the intelligent learning ability of chromosomes; And with the introducing of concepts of fitness vector functions and Pareto approach, the problem that the non-convex solutions are difficult to be found through common aggregation function method could be avoided; Moreover the introduction of master-slave multi-population coordination, balanced the relationship between the global exploration and the local development and enhanced the optimization ability of the algorithm. From the examples we can see that compared to existing algorithms fuzzy-GA, st-GA and m-GA, fmc-GA has not only obtained a better Pareto boundary and Pareto optimal solution, but also demonstrated better practicability and stronger intelligence from both the convergence and distribution. In addition, fmc-GA has the mode of master-slave multi-population collaboration and the mechanism of protecting the non-dominate solution, which can prevent the phenomenon of premature and degradation.