Applied Mathematics & Information Sciences

Author Country (or Countries)

P. R. China


A discrete hybrid evolutionary algorithm is developed to solve global numerical optimization problems with discrete variables. In this algorithm, the orthogonal experimental design acts as the crossover operator to achieve crossover, and the migration operator is employed to keep the population’s diversity. In addition, the simplified quadratic interpolation method is taken as a local search operator, which is adopted to improve the algorithm’s local search ability. Moreover, a few of foreign chromosomes, which are generated via randomly perturbing the best candidate chromosome in the current population, are introduced into the next generation to avoid most of chromosomes gradually clustering around the best candidate chromosome in some subsequent generations. A rounding and truncation procedures is incorporated in the operations of the algorithm to ensure that the integer restrictions and box constraints are satisfied. Numerical experiments on 22 test problems have demonstrated the efficiency of the proposed method.