计算机工程与应用
計算機工程與應用
계산궤공정여응용
COMPUTER ENGINEERING AND APPLICATIONS
2014年
16期
35-38,47
,共5页
过滤机制%适应度缩放%惯性权重%非线性粒子群算法
過濾機製%適應度縮放%慣性權重%非線性粒子群算法
과려궤제%괄응도축방%관성권중%비선성입자군산법
filtering mechanism%fitness scaling%inertia weight%nonlinear particle swarm algorithm
为改进非线性惯性权重粒子群算法,提出了一种带过滤机制的非线性惯性权重粒子群算法。由于原算法存在粒子易陷入局部最优解与搜索效率较低的缺点,将适应度缩放函数引入到非线性惯性动态调整的粒子群算法中,剔除适应度过高与过低的粒子,再对剩余种群部分优良个体进行复制,并随机产生一些新粒子,然后进行交叉操作,种群数量保持不变,减少了粒子陷入局部极值的概率,使结果收敛于全局最优解。通过低维度与高维度函数的对比测试,表明新算法具有较为理想的效果。
為改進非線性慣性權重粒子群算法,提齣瞭一種帶過濾機製的非線性慣性權重粒子群算法。由于原算法存在粒子易陷入跼部最優解與搜索效率較低的缺點,將適應度縮放函數引入到非線性慣性動態調整的粒子群算法中,剔除適應度過高與過低的粒子,再對剩餘種群部分優良箇體進行複製,併隨機產生一些新粒子,然後進行交扠操作,種群數量保持不變,減少瞭粒子陷入跼部極值的概率,使結果收斂于全跼最優解。通過低維度與高維度函數的對比測試,錶明新算法具有較為理想的效果。
위개진비선성관성권중입자군산법,제출료일충대과려궤제적비선성관성권중입자군산법。유우원산법존재입자역함입국부최우해여수색효솔교저적결점,장괄응도축방함수인입도비선성관성동태조정적입자군산법중,척제괄응도과고여과저적입자,재대잉여충군부분우량개체진행복제,병수궤산생일사신입자,연후진행교차조작,충군수량보지불변,감소료입자함입국부겁치적개솔,사결과수렴우전국최우해。통과저유도여고유도함수적대비측시,표명신산법구유교위이상적효과。
This paper proposes nonlinear inertia weight particle swarm optimization with a filtering mechanism to improve the non-linear inertia weight particle swarm algorithm. Due to the original algorithm exsists two shortcomings of particles falling into the local optimal solution and lower search efficiency, introduces fitness scaling function to the nonlinear inertia dynamically for the particle swarm optimization, fitness of excellent and poor particle are removed, then copy some excellent individual of remaining population, meanwhile randomly generated new particles, and crossover operation to them, popu-lations remain unchanged, the methed reduces the opportunity that particulates fall into the localmaximum and make the results converge to the global optimum. In order to verify the effectiveness of the algorithm. In this paper, low dimensions and high dimensional function are compared with each other. The result shows that this method achieves good effects.