教与学信息交互粒子群优化算法

针对单一种群在解决高维问题中收敛速度较慢和多样性缺失的问题,提出了一种教与学信息交互粒子群优化(PSO)算法.根据进化过程将种群动态地划分为两个子种群,分别采用粒子群优化算法和教与学优化算法,同时粒子利用学习者阶段进行子种群之间信息交互,并通过评价收敛性和多样性指标让粒子的收敛能力和多样性在进化过程中得到平衡.与粒子群...     

基于邻域差分和协方差信息的单目标进化算法

复杂的单目标优化问题是进化计算领域的一个研究热点问题.已有差分进化和协方差进化被认为是处理该问题的较有效的方法,其中差分信息类似于梯度可以有效的指导算法朝着最优解方向搜索,而协方差则是基于统计的方式来生成较优的子代种群.本文引入了协方差信息对差分算子进行改进,提出了一种基于邻域差分和协方差信息的进化算法（DEA/NC）来处理复杂的单目标优化问题.算法对现有差分算子中通常采用的随机选点或结合当前最优解进行差分的方式进行了分析,当随机选择的差分个体间的差异较大时,差分信息不能作为一种局部的梯度信息来指导算法的搜索;而结合最优解的差分信息又会使得种群朝着当前最优解的方向搜索,导致种群快速的陷入局部最优.基于此,本文采用了邻域差分的方式来提高差分算子的有效性,同时避免种群的多样性丢失.另外,引入了协方差来度量个体变量间的相关度,并利用相关度来优化差分算子.最后,算法对cec2014中的单目标优化问题进行了测试,并将实验结果与已有的较好的差分进化算法进行了比较,实验结果表明了本算法的有效性.     

多源基因学习的微种群教与学优化及应用

由于微种群教与学优化算法的种群规模较小, 故其种群多样性很难维持. 为提高微种群教与学优化算法的搜索性能, 提出了一种基于多源基因学习的微种群教与学优化算法(micro-population teaching-learning-based optimization based on multi-source gene learning, MTLBO-MGL). 在MTLBO-MGL算法中, 将教阶段和学阶段根据随机选择策略来对个体进行基因水平上的进化操作; 并从基因层面上对种群多样性进行检测和使用稀疏谱聚类方法对种群的每个维度进行聚类. 然后, 根据多样性检测和聚类结果, 选择不同的进化策略来提高所提算法的搜索性能. 在28个测试函数上, 通过将所提算法与其他4种微种群进化算法作对比, 证明了所提算法的整体性能要显著好于所对比的4种算法. 本文还将所提算法应用于无人机三维路径规划问题, 结果表明MTLBO-MGL算法能够在该问题上取得较好结果.     

Biogeography-based optimization with improved migration operator and self-adaptive clear duplicate operator

Biogeography-based optimization (BBO) is a new emerging population-based algorithm that has been shown to be competitive with other evolutionary algorithms. However, there are some insufficiencies in solving complex problems, such as poor population diversity and slow convergence speed in the later stage. To overcome these shortcomings, we propose an improved BBO (IBBO) algorithm integrating a new improved migration operator, Gaussian mutation operator, and self-adaptive clear duplicate operator. The improved migration operator simultaneously adopts more information from other habitats, maintains population diversity, and preserves exploitation ability. The self-adaptive clear duplicate operator can clear duplicate or almost identical habitats, while also preserving population diversity through a self-adaptation threshold within the evolution process. Simulation results and comparisons from the experimental tests conducted on 23 benchmark functions show that IBBO achieves excellent performance in solving complex problems compared with other variants of the BBO algorithm and other evolutionary algorithms. The performance of the improved migration operator is also discussed.     

双群分段交换的改进微粒群优化算法研究

针对微粒优化算法在高维复杂函数寻优上容易陷入局部极值的问题,提出了一种双群分段交换的改进微粒群优化算法（TSME-PSO）。算法将群体分成规模相同的两个种群,两分群采用不同的进化模型更新微粒的位置与速度。算法搜索的不同阶段,交换不同数目的微粒,且数量是不断减少的。通过这些方法,可以有效地提高种群多样性,增强微粒寻优活力。仿真实验表明,TSME-PSO算法可以有效逃离局部极值,整体寻优性能良好,优于其他算法。     

一种基于自适应模糊支配的高维多目标粒子群算法

高维多目标优化问题由于具有巨大的目标空间使得一些经典的多目标优化算法面临挑战.提出一种基于自适应模糊支配的高维多目标粒子群算法MAPSOAF,该算法定义了一种自适应的模糊支配关系,通过对模糊支配的阈值自适应变化若干步长,在加强个体间支配能力的同时实现对种群选择压力的精细化控制,以改善算法的收敛性;其次,通过从外部档案集中选取扰动粒子,并在粒子速度更新公式中新增一扰动项以克服粒子群早熟收敛并改善个体分布的均匀性;另外,算法利用简化的Harmonic归一化距离评估个体的密度,在改善种群分布性的同时降低算法的计算代价.该算法与另外五种高性能的多目标进化算法在标准测试函数集DTLZ{1,2,4,5}上进行对比实验,结果表明该算法在收敛性和多样性方面总体上具有较显著的性能优势.     

一种基于种群多样度的实数编码并行遗传算法

为了改善遗传算法的收敛性能,提出了一种基于个体适应度的种群多样性度量函数,恰当地反映了遗传算法的进化阶段,预报了早熟收敛的趋势. 设计了基于种群多样度函数的迁移算子和交叉算子,并对交叉、变异概率等进行了动态调整,构成了具有多层迁移特点的实数编码并行遗传算法. 通过和其他优秀遗传算法对测试函数的验证比较,结果表明,该算法对于解决遗传算法中早熟、收敛速度慢等问题具有优越的性能.     

Genetic diversity as an objective in multi-objective evolutionary algorithms

A key feature of an efficient and reliable multi-objective evolutionary algorithm is the ability to maintain genetic diversity within a population of solutions. In this paper, we present a new diversity-preserving mechanism, the Genetic Diversity Evaluation Method (GeDEM), which considers a distance-based measure of genetic diversity as a real objective in fitness assignment. This provides a dual selection pressure towards the exploitation of current non-dominated solutions and the exploration of the search space. We also introduce a new multi-objective evolutionary algorithm, the Genetic Diversity Evolutionary Algorithm (GDEA), strictly designed around GeDEM and then we compare it with other state-of-the-art algorithms on a well-established suite of test problems. Experimental results clearly indicate that the performance of GDEA is top-level.     

协同演化算法研究进展

协同演化算法(coevolutionary algorithms,CEA)是当前国际上计算智能研究的一个热点,它运用生物协同演化的思想,是针对演化算法的不足而兴起的,通过构造两个或多个种群,建立它们之间的竞争或合作关系,多个种群通过相互作用来提高各自性能,适应复杂系统的动态演化环境,以达到种群优化的目的.介绍了协同演化算法的研究状况以及目前的研究进展,概述了它的基本算法、主要特点、理论与技术,同时介绍了一些主要的应用领域,指出了协同演化算法的研究方向.     

多样性指导进化算法及其在机器人路径规划中的应用

通过分析及结合机器人路径规划的进化编程仿真实验发现，保存最优个体或淘汰最差个体都会引起进化算法早熟现象，而种群多样性无疑在进化算法中扮演着关键角色。虽然多样性已经用于分析算法中，但是很少用于指导搜索。多样性指导进化算法使用了众所周知的到平均点距离法使变异期与杂交期交替出现。多样性指导进化算法在机器人路径规划问题中展现出显著的结果，与用适应值比较的简单进化算法有着重大的区别。     

一种基于邻域的多目标进化算法

种群维护是多目标进化算法的重要组成部分。针对维护方法和运行效率的矛盾,提出一种基于邻域的多目标进化算法（NMOEA）。定义了一个反映个体之间邻近程度的指标--邻域包含关系,利用此关系对个体进行分布适应度分级的赋值,并用动态方法快速地对种群进行维护。通过7个测试问题和3个方面的测试标准,结果表明新算法在较快速地接近真实的最优面的同时,拥有良好的分布性。     

A radial space division based evolutionary algorithm for many-objective optimization

In evolutionary many-objective optimization, diversity maintenance plays an important role in pushing the population towards the Pareto optimal front. Existing many-objective evolutionary algorithms mainly focus on convergence enhancement, but pay less attention to diversity enhancement, which may fail to obtain uniformly distributed solutions or fall into local optima. This paper proposes a radial space division based evolutionary algorithm for many-objective optimization, where the solutions in high-dimensional objective space are projected into the grid divided 2-dimensional radial space for diversity maintenance and convergence enhancement. Specifically, the diversity of the population is emphasized by selecting solutions from different grids, where an adaptive penalty based approach is proposed to select a better converged solution from the grid with multiple solutions for convergence enhancement. The proposed algorithm is compared with five state-of-the-art many-objective evolutionary algorithms on a variety of benchmark test problems. Experimental results demonstrate the competitiveness of the proposed algorithm in terms of both convergence enhancement and diversity maintenance.     

MaOEA/d2:一种基于双距离构造的高维多目标进化算法

传统的基于Pareto支配关系的多目标进化算法(MOEA)难以有效求解高维多目标优化问题(MaOP). 提出一种利用PBI效用函数的双距离构造的支配关系, 且无需引入额外的参数. 其次, 利用双距离定义了一种多样性保持方法, 该方法不仅考虑了解个体的双距离, 而且还可以根据优化问题的目标数目自适应地调整多样性占比, 以较好地平衡高维目标解群的收敛性和多样性. 最后, 将基于双距离构造的支配关系和多样性保持方法嵌入到NSGA-II算法框架中, 设计了一种基于双距离的高维多目标进化算法MaOEA/d². 该算法与其他5种代表性的高维多目标进化算法一同在5-、10-、15-和20-目标的DTLZ和WFG基准测试问题上进行了IGD和HV性能测试, 结果表明, MaOEA/d²算法具有较好的收敛性和多样性. 由此表明, MaOEA/d²算法是一种颇具前景的高维多目标进化算法.     

基于基因水平多样性的微种群教与学优化算法

种群多样性对微种群教与学优化算法的性能有极大影响。为进一步提高其性能,提出一种基于基因水平多样性的微种群教与学优化算法(MTLBO-GLD)。该算法从基因水平上对种群多样性进行监测;并使用混沌搜索和余弦函数分阶段进行扰动以增加种群多样性。所提算法与八种元启发式算法(四种微种群算法和四种非微种群算法)在13个测试函数上进行性能比较。实验结果表明,MTLBO-GLD的整体性能要显著好于其他八种对比算法。     

Cellular differential evolutionary algorithm with double-stage external population-leading and its application

This paper aims to address the problem of poor diversity and convergence for traditional evolutionary algorithms in solving multi-objective optimization problems and proposes a cellular differential evolutionary algorithm with double-stage external population-leading. This algorithm divides the maintenance of external population diversity into double stages and introduces new disturbance in the mutation operation to avoid the algorithm falling into a local optimum. In the first stage, the external population is maintained according to the rank and k-nearest neighbor distance. The second stage adopts the external population retention strategy of multi-objective cellular differential algorithm that only retains non-dominated solutions. In both stages, the external population has complete feedback, which means the solutions from the external population randomly replace existing individuals in the two-dimensional grid population after every iteration. Tests on fifteen benchmark functions show that the new algorithm can obtain more uniform Pareto front and competitive convergence results than the other four typical algorithms. Finally, the feasibility and effectiveness of this algorithm are verified by the case study of cycloid speed reducer.

     

一种新型的差分演化算法及其应用研究

提出了一种新的基于简单多样性规则的改进差分演化算法,并把它运用于约束全局最优化问题的求解中。新算法的特征是: 1)提出一种新的混合自适应交叉变异算子,以增强算法的搜索能力; 2)采用具有保持群体多样性的约束函数处理技术; 3)简化基本差分演化算法的缩放因子,尽量减少算法的控制参数,方便工程人员的使用。通过对13个标准测试函数进行测试,并与其他演化算法结果进行比较。实验结果表明,新算法在求解精度和稳定性具有很好的性能,而且其函数平均评价次数要低于所比较的其他演化算法。     

基于分级策略的演化算法

为了避免演化算法过早收敛,保持种群多样性,增加算法的搜索能力,本文提出基于分级策略的演化算法．即通过对种群进行分级,来度量种群的多样性,衡量算法是否陷入局部最优,协调种群多样性和精英策略之间的矛盾,再根据种群分布的多样性设计一种有效的半一致交叉算子与单重均匀变异算子。     

基于聚类的快速多目标遗传算法

多目标遗传算法非常适合于求解多目标优化问题．讨论了进化个体之间的支配关系及有关性质，论证了可以用快速排序的方法对进化群体中的个体进行分类，同时探讨了用聚类方法来保持群体的多样性，具体讨论了基于层次凝聚距离的聚类，在此基础上提出了用分类和聚类的方法构造新的进化群体．理论分析与实验结果表明，所讨论的方法比较国际上已有的方法具有更快的收敛速度．     

基于多策略的多目标粒子群优化算法

针对目前多目标粒子群优化算法的收敛性能和非劣解的多样性不能同时得到满足等缺陷,提出一种基于多策略的多目标粒子群优化算法（Multi-Objective Particle Swarm Optimization algorithm for Multi-Strategy,MS-MOPSO）。采用非支配排序和拥挤距离排序相结合策略,重新划分外部种群和进化种群;采用小生境选择策略,在外部种群中选择最佳粒子作为领导粒子,用于领导进化种群中粒子的进化;在进化种群中利用多尺度高斯变异策略,平衡算法的全局搜索和局部精确搜索;采用邻域认知个体极值更新策略,不断更新个体极值。将该算法应用到典型的多目标测试函数,并与其他多目标优化算法进行对比分析,测试结果表明该算法中四个策略的有效性和互补性,同时验证了该算法不但具有较好的收敛性和收敛速度,而且该算法最优解的分布具有良好的均匀性和多样性。     

A diversity control mechanism in many objective optimizations

Multi-objective evolutionary optimization algorithms are among the best optimizers for solving problems in control systems, engineering and industrial planning. The performance of these algorithms degrades severely due to the loss of selection pressure exerted by the Pareto dominance relation which will cause the algorithm to act randomly. Various recent methods tried to provide more selection pressure but this would cause the population to converge to a specific region which is not desirable. Diversity reduction in high dimensional problems which decreases the capabilities of these approaches is a decisive factor in the overall performance of these algorithms. The novelty of this paper is to propose a new diversity measure and a diversity control mechanism which can be used in combination to remedy the mentioned problem. This measure is based on shortest Hamiltonian path for capturing an order of the population in any dimension. In order to control the diversity of population, we designed an adaptive framework which adjusts the selection operator according to diversity variation in the population using different diversity measures as well as our proposed one. This study incorporates the proposed framework in MOEA/D, an efficient widely used evolutionary algorithm. The obtained results validate the motivation on the basis of diversity and performance measures in comparison with the state-of-the-art algorithms and demonstrate the applicability of our algorithm/method in handling many-objective problems. Moreover, an extensive comparison with several diversity measure algorithms reveals the competitiveness of our proposed measure.