针对经典蜘蛛蜂优化算法初始种群分布不合理、搜索与开发之间的转换不平衡、易陷入局部最优等问题,提出了一种融合对偶学习的动态蜘蛛蜂优化算法(dynamic spider wasp optimizer combined with duality learning,CLDSWO).首先,结合Tent和Sinusoidal映射,设计了TS(Tent-Sinusoidal)映射,并采用TS映射生成分布更广泛且均匀的初始蜘蛛蜂种群.其次,设计了一个动态权衡因子,自适应地调整狩猎和交配行为之间的转换,实现全局搜索和局部优化之间的平衡.引入了基于对偶学习的变异机制,在对偶学习的过程中,引入逐维变异机制,加速算法的收敛,增强逃离局部最优的能力.为了验证CLDSWO算法的有效性,利用10个基准函数和CEC2017函数进行实验,并通过 Wilcoxon检验证实仿真结果的显著性,实验结果表明,CLDSWO在平衡收敛精度和速度方面更具竞争力.将CLDSWO算法应用至压力容器设计问题和无源时差定位问题中,结果表明CLDSWO的精度分别提升了1.28%和36.67%,验证了CLDSWO算法在求解实际工程应用问题中的有效性.
Dynamic spider wasp optimizer incorporating duality learning and its application
The spider wasp optimizer has problems such as irrational initial population distribution,unbalanced transition between search and exploitation,and a tendency to fall into local optimization.Therefore,a dynamic spider wasp optimizer combined with duality learning(CLDSWO)is proposed to solve these problems.Firstly,the Tent-Sinusoidal(TS)mapping which combines the Tent and Sinusoidal mapping is designed to generate the initial spider-wasp population with a wider and uniform distribution.Secondly,a dynamic tradeoff factor is developed to adaptively adjust the tradeoff between hunting and mating behaviors to achieve a balance between global search and local optimization.Finally,a mutation mechanism based on duality learning is introduced to accelerate the convergence and enhance the ability to escape from the local optimum.To verify the effectiveness of CLDSWO,10 benchmark functions,CEC2017 functions,and Wilcoxon tests are carried out.The results show that CLDSWO is more competitive in balancing convergence accuracy and speed.The CLDSWO algorithm is applied to the pressure vessel design problem and the time difference of arrival localization problem.The results show that the accuracy of CLDSWO was improved by 1.28%and 36.67%,respectively,validating the effectiveness of CLDSWO in solving practical engineering applications.
spider wasp optimizerdynamic tradeoff factorduality learningdimensional mutationengineering applica-tions
万方数据
维普
