CAJ | 학술논문

　　提出一种双种群粒子群算法，在粒子进化过程中，具有当前最优位置的种群侧重于局部搜索，而不具有当前最优位置的种群侧重于全局搜索。两个种群在进化过程中受共同的群体最优位置影响进行进化，从而实现信息共享，协调进化。利用几个测试函数对算法性能进行分析验证，并与其他改进算法进行比较，结果表明算法在搜索精度、稳定性以及搜索速度上均优于改进算法。将双种群粒子群算法用于UUV三维空间轨迹规划问题，获得了满意的规划效果。
　　제출일충쌍충군입자군산법，재입자진화과정중，구유당전최우위치적충군측중우국부수색，이불구유당전최우위치적충군측중우전국수색。량개충군재진화과정중수공동적군체최우위치영향진행진화，종이실현신식공향，협조진화。이용궤개측시함수대산법성능진행분석험증，병여기타개진산법진행비교，결과표명산법재수색정도、은정성이급수색속도상균우우개진산법。장쌍충군입자군산법용우UUV삼유공간궤적규화문제，획득료만의적규화효과。
Two-Subpopulation Particle Swarm Optimization(TSPSO)is proposed. The subpopulation which has the optimal location of the current iterative tends to local exploration, while the other subpopulation tends to global exploration. Both sub-populations are influenced by the group optimal location of the current iterative, so they can fully share information. The perfor-mance of the Particle Swarm Optimization is tested by several test functions. It is turned out that the TSPSO is better than other algorithms in search accuracy, stability and search speed. TSPSO is used to solve UUV 3D path planning problem, and obtains satisfactory performance.