Abstract: Evolutionary algorithm is a new kind of efficient methods for complex nonlinear programming, however, the amount of their computation is usually very large, and the constraints can not be handled efficiently. In this paper, firstly, the constrained problem is transformed into an unconstrained one so as to reduce the difficulty of problem solving. Secondly, to reduce the number of local optimal solutions, a smoothing technique is adopted. It can eliminate all local optimal solutions which are not better than the current best solution found so far, and keep all the local optimal solutions which is better than the current best solution. Furthermore, a new crossover operator is designed. Based on all these, an improved evolutionary algorithm is proposed and experimental results show the efficiency of the proposed algorithm with less computation, higher convergent speed for all test problems.

Key words: constraint handling, smoothing techniques, evolutionary algorithm, evolutionary operators, global optimization