In this paper an improved genetic algorithm based on the simplex self-mapping fixed point algorithm is proposed. With this algorithm, the optimal problem of n-dimensional closure function will be transformed as the solution of approximate fixed point problem of n-dimensional standard simplexes by homeomorphism mapping. The genetic operators relying on the integer labels are designed. In this case, whether every individual loading simplex of the population is a completely labeled simplex can be used as an objective convergence criterion. The simulation results demonstrate that the proposed algorithm is valid and effective.