DNAPSO算法在连续空间优化问题中的应用  

Application of DNAPSO Algorithm for Continuous Space Optimization

作　　者：马翠[1] 周先东[1] 

机构地区：[1]第三军医大学数学与生物数学教研室,重庆400038

出　　处：《重庆理工大学学报（自然科学）》2015年第5期87-92,共6页Journal of Chongqing University of Technology：Natural Science

基　　金：重庆市自然科学基金资助项目(CSTC2013jcyj A10041)

摘　　要：针对DNA计算方法中的个体在进化过程中具有多样性、容易导致局部最优的问题,提出了一种新的DNAPSO算法。该算法利用PSO算法中个体依据全局最优解和局部最优解决定的进化方向原理,设计了向特定方向变异的多点变异算子,同时保留了DNA计算中复制、交叉重组等算子,使新算法既具有了个体多样性特点,又具备了向最优解快速收敛的能力。多维连续空间优化问题中4个典型函数的仿真测试结果表明:所提出的DNAPSO算法在收敛精度、收敛速度和鲁棒性方面较之DNA计算方法和标准PSO算法都有明显提高,丰富了连续空间优化问题的求解方法。In order to overcome the defects of local optimum which are generated by the individual di- versity in the evolutionary process, a new DNAPSO algorithm based on DNA structure and the evolu- tion process of particle swarm optimization was proposed. The DNAPSO algorithm used the evolution- ary principle of the individual＇ s gradually flying to the optimal solution in the PSO. The new algo- rithm retained the advantages of DNA algorithm, multipoint mutation operator which can mutate to the particular direction was designed. Therefore, the proposed algorithm both had the feature of individual diversity and the ability of fast convergence to the optimal solution. And then, results of four typical functions in the continuous space optimization show that the DNAPSO algorithm has better stability and convergence compared with DNA algorithm and standard PSO algorithm, and a new way is found to solve the continuous space optimization.

关 键 词：DNAPSO算法 粒子群优化算法 连续空间优化问题 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]