CAJ | 학술논문

为了提高粒子群算法中粒子搜索全局最优解的准确度,确保粒子的收敛性,提出了基于知识空间的分组式粒子群算法(KGPSO).该算法使用K-means算法对粒子群进行分组,利用较小的最大飞行速度(Vmax)加强粒子在组内的局部搜索能力,并将"知识空间"的概念带入到分组中,由知识空间中的粒子来引导群中粒子前往更好的解空间搜索.实验结果表明,KGPSO算法在测试函数的表现整体优于过去学者提出的标准PSO,HPSO、FPSO.
위료제고입자군산법중입자수색전국최우해적준학도,학보입자적수렴성,제출료기우지식공간적분조식입자군산법(KGPSO).해산법사용K-means산법대입자군진행분조,이용교소적최대비행속도(Vmax)가강입자재조내적국부수색능력,병장"지식공간"적개념대입도분조중,유지식공간중적입자래인도군중입자전왕경호적해공간수색.실험결과표명,KGPSO산법재측시함수적표현정체우우과거학자제출적표준PSO,HPSO、FPSO.
To improve the accuracy of the global optimal solution and ensure the convergence of particle, an algorithm of grouping particle swarm optimization based on knowledge space (KGPSO) is presented. Firstly, the initial particles are divided into several groups by K-means algorithm and the smaller Vmax is gotten to enhance searching ability. Then, the concept of "knowledge space" is taken into the particle groups. All particle groups are guided to search by the particles that in knowledge space. Finally, performances of the proposed algorithm are demonstrated by the application and the results shown that KGPSO is effective and gains better performance than SPSO (standard particle swarm optimization), HPSO (hybrid particle swarm optimization) and FPSO (fuzzy adaptive particle swarm optimization).