A Hybrid Adaptive Evolutionary Algorithm for Constrained Optimization

In this paper a hybrid adaptive genetic algorithm is proposed for solving constrained optimization problems. Genetic algorithm proposed here combines adaptive penalty method and smoothing technique in order to make the algorithm not needing parameters tuning and easily escaping from the local optimal solutions. Meanwhile, local line search technique is introduced and a new crossover operator is designed for getting much faster algorithm convergence. If there is no feasible solutions in the current population, finding feasible solutions is prior to finding optimal solution, otherwise the exploitation for global optimal solution based on a certain smoothing function at the best feasible solution in the current population and the exploration for whole search space are processing at the same time. The performance of the algorithm is tested on thirteen benchmark functions in the literature and the results indicate that the algorithm proposed here is robust and effective.