CAJ | 학술논문

针对粒子群优化算法易出现早熟收敛、陷入局部最优的问题,提出了在粒子群搜索解的过程中监控粒子健康度的方法,对健康度低的粒子进行交叉操作。该方法既保证了健康粒子继续搜索最优解,又有效地改变了非健康粒子的状态,提高了粒子群的寻优能力以及跳出局部最优解的能力。最后通过实验数据集验证了新算法,实验结果表明与标准粒子群算法相比新算法在探索潜在最优解、保持种群多样性方面具有良好的效果。
침대입자군우화산법역출현조숙수렴、함입국부최우적문제,제출료재입자군수색해적과정중감공입자건강도적방법,대건강도저적입자진행교차조작。해방법기보증료건강입자계속수색최우해,우유효지개변료비건강입자적상태,제고료입자군적심우능력이급도출국부최우해적능력。최후통과실험수거집험증료신산법,실험결과표명여표준입자군산법상비신산법재탐색잠재최우해、보지충군다양성방면구유량호적효과。
For the premature convergence which is easily falling into local optimum on the particle swarm optimization searching process,this paper proposed a crossover operation to the particle with low health degree.This method not only effectively improved the unhealthy particles and let them jump out of local optimum,but also ensured the healthy particles to continue searching for optimal solutions.Finally,the new algorithm is verified by the Benchmark problem.The experimental results show that the new algorithm proposed is competitive to solve vehicle routing problem with time window.