Application of multi population genetic algorithm in traffic assignment problem

In view of the problem of premature convergence of simple genetic algorithm, a multi population genetic algorithm for traffic assignment problems was proposed, multiple populations are introduced and search simultaneous, coevolution through populations are implemented, use immigration operator to exchange information, artificial selection operator to keep best individuals of every generation, and as a judge of the convergence of the algorithm as well. Finally, an example is tested to verify that the algorithm can effectively overcome the premature convergence of simple genetic algorithm, and has strong robustness.