Abstract: Congruent circle packing problem is one of the complex combination optimization problems with strong constraint. Aiming at the characteristics and difficulties of strong constraints, this paper proposes a method to solve the problem by improving the traditional differential evolution algorithm. The improved algorithm uses the effective solution space to improve the mutation of differential evolution, and uses the stochastic ranking mechanism to improve the selection of differential evolution. A lot of experiments have been done and prove that the result of the proposed algorithm is about less than 1% with the current optimal value, so the evolutionary computation is feasible in solving the congruent circle packing problem. At the same time, evolution algorithm is of good convergence, thus there will be a very good application in other strong constraint complex optimization problems.

Key words: improved differential evolution, congruent circle packing problem, strong constrain, stochastic selecting, effective solution space