传统的粒子群优化算法(PSO)基于粒子的搜索向最优方向逼近,在解决复杂的多峰函数的优化问题时容易陷入局部极值点,得不到正确的优化结果。为此研究能够优化多峰函数的新型粒子群优化算法。将混沌的概念和子种群的概念引入粒子群优化算法,从而形成一种新型的基于子种群的混沌粒子群优化算法,用以解决多峰值优化问题。这种新型的优化算法应用于PCB电路板的模块匹配问题,通过实验研究验证了它的有效性。算法能够识别并准确找到电阻元件,并检测出焊错的原件。该优化方法适用于过程优化、模式识剐、图像检测等复杂的工程优化问题。
The usual particle swarm optimization (PSO) algorithm tends to the optimal direction based upon the search of the particles. When solving the optimal problems with multi - peaks function, it is easy to fall into a local optimal point and cannot obtain the desired optimizing results. This paper studies a novel PSO algorithm which can solve the multi -peaks optimizing problem. A novel species - based chaotic particle swarm optimization method is proposed by introducing the notion of chaos and species into PSO to solve multi - peaks optimal problems. This novel algorithm is then used to solve the template matching problems in PCB industry. By experiment tests, its effectiveness is fully illustrated. This optimal method is suitable to some complex optimal problems such as processes optimiza- tion, mode reorganization and image detection.