顺序相依拆卸线平衡多目标优化模型及算法研究

拆卸是产品回收过程最关键环节之一,拆卸效率直接影响再制造成本。本文在分析现有模型不足基础上,考虑最小化总拆卸时间,建立多目标顺序相依拆卸线平衡问题优化模型,并提出了一种自适应进化变邻域搜索算法。所提算法引入种群进化机制,并采用一种组合策略构建初始种群,通过锦标赛法选择个体进化;在局部搜索时,设计了邻域结构自适应选择策略,并采用基于交叉的全局学习机制加速跳出局部最优,以提高算法寻优能力。对比实验结果,证实了所提模型的合理性以及算法的高效性。     

改进的布谷鸟算法求解物流配送中心选址问题

针对基本布谷鸟算法求解物流配送中心选址问题时存在搜索精度低、易陷入局部最优值的缺陷,提出一种改进的布谷鸟算法.算法采用基于寄生巢适应度值排序的自适应方法改进基本布谷鸟算法的惯性权重,以平衡算法的全局开发能力和局部探索能力;利用NEH领域搜索以提高算法的搜索精度和收敛速度;引入停止阻止策略对全局最优寄生巢位置进行变异避免算法陷入局部最优值、增加种群的多样性.通过实验仿真表明,改进的布谷鸟算法在求解物流配送中心选址问题上要优与基本布谷鸟算法以及其它智群算法,是一种有效的算法.     

基于正余双弦自适应灰狼优化算法的医药物流配送路径规划

针对传统灰狼优化算法易早熟收敛陷入局部最优和收敛速度慢的缺陷,提出一种正余双弦自适应灰狼优化算法.首先,在灰狼捕食阶段引入正弦搜索,增强算法的全局勘探能力,减少算法的搜索盲点,提高算法的搜索精度.在引入正弦搜索的同时,引入余弦搜索,增强算法的局部开发能力,提高算法的收敛速度.其次,在搜索过程中加入自适应交叉变异机制,通过适应度值的大小自适应选取交叉变异概率,有效的提高了粒子跳出局部最优的概率.通过数值对比试验,验证了改进算法具有较强的收敛精度和收敛速度.     

基于混沌优化和自适应反向学习的秃鹰搜索算法

针对秃鹰搜索算法求解精度低、收敛速度较慢、容易陷入局部最优的问题,提出一种基于混沌优化和自适应反向学习的秃鹰搜索算法.首先,在选择搜索空间阶段引入正弦混沌映射更新秃鹰群位置,增加随机性,优化全局搜索能力;其次,在俯冲捕获猎物阶段加入指数自适应,平衡了全局搜索和局部搜索,同时加快收敛速度;最后,对更新后的最优秃鹰位置使用反向学习策略,提高跳出局部最优的可能性.选取12个测试函数对算法的性能进行了测试,结果表明本文改进的秃鹰搜索算法具有更优的求解精度和收敛速度.     

基于动态自适应t分布变异的人群搜索算法

针对人群搜索算法在进化后期大量个体聚集局部最优时,易陷入局部最优,搜索精度低的缺陷,提出一种基于t分布变异的人群搜索算法.算法使用动态自适应方式确定变异步长,引入t分布变异算子以融合柯西变异和高斯变异的优点,促进算法在进化早期具备良好的全局探索能力,在进化后期收获较强的局部开发能力,增加种群的多样性;采用边界缓冲墙策略处理越界问题,避免越界个体聚集在边界值上的缺陷.实验结果表明,算法比基本人群搜索算法具有更高的寻优精度和收敛速度,是一种有效的算法.     

一种改进灰狼优化算法研究及应用

针对灰狼算法易陷入局部最优、收敛精度不高、收敛速度慢等缺点,提出一种改进的灰狼算法.引入莱维飞行,扩大搜索范围,增强全局搜索能力,避免陷入局部最优;引入贪婪原理,提升种群优良性以提高算法收敛精度;引入自适应收敛因子,加快收敛速度;引入动态权重策略,制约全局搜索与局部搜索的相互影响.将改进算法与其他四种算法作对比,实验表明,改进算法在收敛速度与收敛精度上都有更好的性能.最后,应用于图像多阈值分割中,采用GWO-Otsu法可以克服传统Otsu法在多阈值分割时计算量大,实时性差的特点,不但能够取得最优解,且明显缩减计算时间.     

基于梯度的自适应快速布谷鸟搜索算法

针对标准布谷鸟搜索(CS)算法存在全局搜索和局部搜索能力不平衡的缺点, 提出一种基于梯度的自适应快速布谷鸟搜索(GBAQCS)算法. 在改进的算法中, 针对偏好随机游动的步长, 在利用目标函数的梯度决定步长方向的基础上, 首先提出自适应搜索机制平衡了算法的全局搜索和局部搜索能力; 其次提出快速 搜索策略, 充分利用当前鸟巢信息进行精细化搜索, 从而提高算法的搜索精度和收敛速度. 实验结果表明, 相比其他算法, 所提出的改进策略使算法的全局搜索和局部搜索能力保持了相对的平衡, 并提高了算法的收敛性能.     

求解带时间窗车辆路径问题的混合Memetic算法

为提高带时间窗车辆路径问题的求解精度和求解效率,设计了一种混合Memetic算法。采用基于时间窗升序排列的混合插入法构造初始种群,提高解质量的同时兼顾多样性,扩大搜索空间;任意选择组成父代种群,以维持搜索空间;运用简化的变邻域搜索进行局部开发,引入邻域半径减少策略提高开发效率,约束放松机制开放局部空间;以弧为对象,增加种群向当前最优解和全局最优解的后学习过程。实验结果表明,所提出的算法具有较好的寻优精度和稳定性,能搜索到更好的路径长度结果,更新了现有研究在最短路径长度的目标函数上的下限。     

基于正弦控制因子的Lateral变异鲨鱼优化算法

针对传统鲨鱼优化算法在求解高维目标函数时,易早熟收敛,陷入局部最优的缺陷.提出一种基于正弦控制因子的Lateral变异鲨鱼优化算法.通过正弦曲线的特性和自适应惯性权重,改善了传统鲨鱼优化算法中由于随机选取控制因子数值大小可能导致算法在迭代后期全局搜索能力降低的问题,提高了算法在迭代后期的全局收敛能力,并对最佳鲨鱼位置引入Lateral变异策略,加强了算法跳出局部最优的可能性.改进后的算法对多个shifted单峰,多峰以及固定维测试函数进行求解,实验结果表明,对比多种不同优化算法而言,本文所提LSSO算法具有更高的收敛精度和搜索速度.     

基于Cubic混沌模型的自适应布谷鸟优化算法

针对基本布谷鸟算法(CS)求解精度有限、收敛速度慢,易陷入局部最优的不足,提出一种基于Cubic混沌模型的自适应布谷鸟优化算法.算法在迭代时,自动调整Lévy flights随机搜索的步长因子,提高算法的收敛速度;将Cubic混沌映射模型嵌入布谷鸟算法,产生混沌扰动信号,对鸟巢位置进行更新,扩大种群多样性,提高全局最优值的搜索能力.通过10个标准测试函数的实验及结果分析,表明算法在寻优精度和收敛速度两方面的改进是有效的.     

Adaptive brainstorm optimisation with multiple strategies

Brainstorm optimisation (BSO) algorithm is a recently developed swarm intelligence algorithm inspired by the human problem-solving process. BSO has been shown to be an efficient method for creating better ideas to deal with complex problems. The original BSO suffers from low convergence and is easily trapped in local optima due to the improper balance between global exploration and local exploitation. Motivated by the memetic framework, an adaptive BSO with two complementary strategies (AMBSO) is proposed in this study. In AMBSO, a differential-based mutation technique is designed for global exploration improvement and a sub-gradient strategy is integrated for local exploitation enhancement. To dynamically trigger the appropriate strategy, an adaptive selection mechanism based on historical effectiveness is developed. The proposed algorithm is tested on 30 benchmark functions with various properties, such as unimodal, multimodal, shifted and rotated problems, in dimensions of 10, 30 and 50 to verify their scalable performance. Six state-of-the-art optimisation algorithms are included for comparison. Experimental results indicate the effectiveness of AMBSO in terms of solution quality and convergence speed.     

A quantum-behaved simulated annealing algorithm-based moth-flame optimization method

This study develops an improved moth-flame optimization algorithm, which is a recently proposed optimizer based on moth behavior in nature. It has achieved favorable results in medical science, educational evaluation, and other fields. However, the convergence rate of the original moth-flame optimization algorithm is too fast in the running process, and it is prone to fall into local optimum, which leads to the failure to produce the high-quality optimal result. Accordingly, this paper proposes a reinforced technique for the moth-flame optimization algorithm. Firstly, the simulated annealing strategy is introduced into the moth-flame optimization algorithm to boost the advantage of the algorithm in the local exploitation process. Then, the idea of the quantum rotation gate is integrated to enhance the global exploration ability of the algorithm and ameliorate the diversity of the moth. These two steps maintain the relationship between exploitation and exploration as well as strengthen the performance of the algorithm in both phases. After that, the method is compared with ten well-regarded and ten alternative algorithms on benchmark functions to verify the effectiveness of the approach. Also, the Wilcoxon signed rank and Friedman assessment were performed to verify the significance of the proposed method against other counterparts. The simulation results reveal that the two introduced strategies significantly improve the exploration and exploitation capacity of moth-flame optimization algorithm. Finally, the algorithm is utilized to feature selection and two engineering problems, including pressure vessel design and multiple disk clutch brake problems. In these practical applications, the novel algorithm also achieves particularly notable results, which also illustrates that the algorithm is qualified is an effective auxiliary appliance in solving complex optimization problems.     

Particle swarm optimization with chaotic opposition-based population initialization and stochastic search technique

Particle swarm optimization (PSO) is a relatively new optimization algorithm that has been applied to a variety of problems. However, it may easily get trapped in a local optima when solving complex multimodal problems. To address this concerning issue, we propose a novel PSO called as CSPSO to improve the performance of PSO on complex multimodal problems in the paper. Specifically, a stochastic search technique is used to execute the exploration in PSO, so as to help the algorithm to jump out of the likely local optima. In addition, to enhance the global convergence, when producing the initial population, both opposition-based learning method and chaotic maps are employed. Moreover, numerical simulation and comparisons with some typical existing algorithms demonstrate the superiority of the proposed algorithm.     

Adaptive mesh selection asymptotically guarantees a prescribed local error for systems of initial value problems

We study potential advantages of adaptive mesh point selection for the solution of systems of initial value problems. For an optimal order discretization method, we propose an algorithm for successive selection of the mesh points, which only requires evaluations of the right-hand side function. The selection (asymptotically) guarantees that the maximum local error of the method does not exceed a prescribed level. The usage of the algorithm is not restricted to the chosen method; it can also be applied with any method from a general class. We provide a rigorous analysis of the cost of the proposed algorithm. It is shown that the cost is almost minimal, up to absolute constants, among all mesh selection algorithms. For illustration, we specify the advantage of the adaptive mesh over the uniform one. Efficiency of the adaptive algorithm results from automatic adjustment of the successive mesh points to the local behavior of the solution. Some numerical results illustrating theoretical findings are reported.     

An evolutionary artificial immune system for multi-objective optimization

In this paper, an evolutionary artificial immune system for multi-objective optimization which combines the global search ability of evolutionary algorithms and immune learning of artificial immune systems is proposed. A new selection strategy is developed based upon the concept of clonal selection principle to maintain the balance between exploration and exploitation. In order to maintain a diverse repertoire of antibodies, an information-theoretic based density preservation mechanism is also presented. In addition, the performances of various multi-objective evolutionary algorithms as well as the effectiveness of the proposed features are examined based upon seven benchmark problems characterized by different difficulties in local optimality, non-uniformity, discontinuity, non-convexity, high-dimensionality and constraints. The comparative study shows the effectiveness of the proposed algorithm, which produces solution sets that are highly competitive in terms of convergence, diversity and distribution. Investigations also demonstrate the contribution and robustness of the proposed features.     

Adjacency based method for generating maximal efficient faces in multiobjective linear programming

Multiobjective linear optimization problems (MOLPs) arise when several linear objective functions have to be optimized over a convex polyhedron. In this paper, we propose a new method for generating the entire efficient set for MOLPs in the outcome space. This method is based on the concept of adjacencies between efficient extreme points. It uses a local exploration approach to generate simultaneously efficient extreme points and maximal efficient faces. We therefore define an efficient face as the combination of adjacent efficient extreme points that define its border. We propose to use an iterative simplex pivoting algorithm to find adjacent efficient extreme points. Concurrently, maximal efficient faces are generated by testing relative interior points. The proposed method is constructive such that each extreme point, while searching for incident faces, can transmit some local informations to its adjacent efficient extreme points in order to complete the faces’ construction. The performance of our method is reported and the computational results based on randomly generated MOLPs are discussed.     

FEMOEA: a fast and efficient multi-objective evolutionary algorithm

A multi-objective evolutionary algorithm which can be applied to many nonlinear multi-objective optimization problems is proposed. Its aim is to quickly obtain a fixed size Pareto-front approximation. It adapts ideas from different multi-objective evolutionary algorithms, but also incorporates new devices. In particular, the search in the feasible region is carried out on promising areas (hyperspheres) determined by a radius value, which decreases as the optimization procedure evolves. This mechanism helps to maintain a balance between exploration and exploitation of the search space. Additionally, a new local search method which accelerates the convergence of the population towards the Pareto-front, has been incorporated. It is an extension of the local optimizer SASS and improves a given solution along a search direction (no gradient information is used). Finally, a termination criterion has also been proposed, which stops the algorithm if the distances between the Pareto-front approximations provided by the algorithm in three consecutive iterations are smaller than a given tolerance. To know how far two of those sets are from each other, a modification of the well-known Hausdorff distance is proposed. In order to analyze the algorithm performance, it has been compared to the reference algorithms NSGA-II and SPEA2 and the state-of-the-art algorithms MOEA/D and SMS-EMOA. Several quality indicators have been considered, namely, hypervolume, average distance, additive epsilon indicator, spread and spacing. According to the computational tests performed, the new algorithm, named FEMOEA, outperforms the other algorithms.     

A hybrid ACO-GRASP algorithm for clustering analysis

Cluster analysis is an important tool for data exploration and it has been applied in a wide variety of fields like engineering, economics, computer sciences, life and medical sciences, earth sciences and social sciences. The typical cluster analysis consists of four steps (i.e. feature selection or extraction, clustering algorithm design or selection, cluster validation and results interpretation) with feedback pathway. These steps are closely related to each other and affect the derived clusters. In this paper, a new metaheuristic algorithm is proposed for cluster analysis. This algorithm uses an Ant Colony Optimization to feature selection step and a Greedy Randomized Adaptive Search Procedure to clustering algorithm design step. The proposed algorithm has been applied with very good results to many data sets.     

Ensemble feature selection for high dimensional data: a new method and a comparative study

The curse of dimensionality is based on the fact that high dimensional data is often difficult to work with. A large number of features can increase the noise of the data and thus the error of a learning algorithm. Feature selection is a solution for such problems where there is a need to reduce the data dimensionality. Different feature selection algorithms may yield feature subsets that can be considered local optima in the space of feature subsets. Ensemble feature selection combines independent feature subsets and might give a better approximation to the optimal subset of features. We propose an ensemble feature selection approach based on feature selectors’ reliability assessment. It aims at providing a unique and stable feature selection without ignoring the predictive accuracy aspect. A classification algorithm is used as an evaluator to assign a confidence to features selected by ensemble members based on their associated classification performance. We compare our proposed approach to several existing techniques and to individual feature selection algorithms. Results show that our approach often improves classification performance and feature selection stability for high dimensional data sets.