CAJ | 학술논문

针对粒子群算法后期存在寻优效率降低、收敛缓慢等问题,提出了多级优化算法.该算法具有局部快速收敛特性,通过对粒子群所生成的最优粗略解进行局部最优处理,从而能够快速地从粗略解中提取出全局最优信息,将粗略解变为最优解.仿真结果显示,该组合算法能将粒子群算法的全局搜索特性和多级优化算法的局部优化特性有机结合起来,达到了准确而快速生成路径的目的.
침대입자군산법후기존재심우효솔강저、수렴완만등문제,제출료다급우화산법.해산법구유국부쾌속수렴특성,통과대입자군소생성적최우조략해진행국부최우처리,종이능구쾌속지종조략해중제취출전국최우신식,장조략해변위최우해.방진결과현시,해조합산법능장입자군산법적전국수색특성화다급우화산법적국부우화특성유궤결합기래,체도료준학이쾌속생성로경적목적.
For some deficiency in PSO anaphase convergence’s efficiency,a new method called multi-step opti?mization was proposed. As its high efficiency in searching the local-best route result,it can be used it to get the global-best information from the cursory route created by PSO algorithm. Then the cursory route would be?come the global-best route. The simulation result indicated that the method could combine the specialty of PSO with multi-step optimization effectively. The process was fast and the route was preise.