求解约束优化问题的一种新方法——基于量子粒子群优化算法

A New Method for Solving Constrained Optimization Problems Based on Particle Swarm Optimization with Quantum Behavior

在用粒子群优化(PSO)算法求解约束优化问题时,处理好约束条件是取得良好的优化效果的关键。针对群体智能和约束优化问题的特点,提出了一种在每次迭代中有选择地保留一定数量不可行解的方法——DCFI(DirectChooseFixedInfeasiblesolutions)法,并把它结合到最近提出的量子粒子群优化(QDPSO)算法中。该算法可以利用保留下来的不可行解来帮助搜索靠近边界的最优解,同时又可以避免罚因子的选择问题。数值实验显示了该算法的有效性。

A good constraints handling mothed is very significant in constrained optimization problems.According to the characters of swarm Intelligence and cotnsrained optimization,we propose an mothod to handling constraints called Direct Choose Fixed Infeasible solutions（DCFI）.In this algorithm,we search better feasible solutions closed to constraints through preserving a certain number of classy infeasible individuals.We also associate the method with Quantum Delta- Potential-Well-based Particle Swarm Optimization （QDPSO）algorithm recently proposed.Results of experiments de monstrate that the algorithm we proposed is efficient.