摘要
针对差分进化算法存在易早熟、收敛精度低等缺陷,提出一种自适应二次变异的改进差分进化算法(Modified differential evolution algorithm based on adaptive secondary variation,ASVDE)。采用多变异策略,并加入动态调节因子平衡不同变异策略的权重;当适应值不更新的代数达到设定值时,利用全局最优信息和柯西分布对当前种群进行二次变异优化,使算法及时跳出停滞状态,最终在反向个体与试验个体间获得最优结果。仿真结果表明,相比于其他3种算法,ASVDE算法的精度更高,应用于电力系统经济调度问题所得结果也更优。
In order to improve the shortcomings of differential evolution algorithm,such as premature convergence and low convergence precision,a modified differential evolution algorithm based on adaptive secondary variation(ASVDE)is proposed.It adopted the multi-variation strategy,and added a dynamic adjustment factor to balance the weights of different mutation strategies.When the algebra whose adaptive value was not updated reached the set threshold,the global optimal information and Cauchy distribution were used to execute the second variation of the current population,so that the algorithm can jump out of the stagnant state in time.Finally,the best result was obtained between the opposite individual and the experimental individual.The simulation results show that the accuracy of ASVDE algorithm is higher than other three algorithms.The algorithm is applied to the economic load dispatch problem of power system,and the result is better than other algorithms.
作者
胡福年
董倩男
吕璐
Hu Funian;Dong Qiannan;LüLu(School of Electrical Engineering and Automation,Jiangsu Normal University,Xuzhou 221116,Jiangsu,China)
出处
《计算机应用与软件》
北大核心
2021年第7期271-280,共10页
Computer Applications and Software
基金
江苏省普通高校研究生科研创新计划项目(2018YXJ077)。
关键词
差分进化算法
变异策略
反向个体
电力系统经济调度
Differential evolution algorithm
Mutation strategy
Opposite individual
Economic load dispatch