利用冲突信息降维的进化高维目标优化算法

进化多目标优化算法求解高维目标优化问题面临收敛能力、计算复杂度、决策以及Pareto前沿的可视化等困难，其根本原因是目标空间维数高。目标降维通过丢弃冗余目标，为缓解高维目标优化求解困难提供一种新思路。本文提出利用冲突信息降维的分解进化高维目标优化算法（CIOR-MOEA/D）。该方法通过衡量目标在近似解集上体现的冲突性，构造问题的冲突信息矩阵，对该矩阵进行特征分析，确定目标的重要性程度，实现维数约简，并利用分解进化多目标优化算法（MOEA/D）对重要子目标集合进行分解进化,从而得到问题的近似解集。实验结果表明,本文提出的目标降维算法在降维的准确性与鲁棒性上均表现突出，能够有效地处理冗余高维目标优化问题。      

基于混合进化算法的特征选择方法研究

特征选择（Feature Selection,FS）是一种有效的数据预处理方法，它可以通过选择高维数据中一组具有高相关性和低冗余性的特征，从而解决数据冗余引起的维数灾难.目前许多计算方法已经被应用于求解FS问题，其中基于教与学优化（Teaching and Learning-based Optimization Algorithm,TLBO）的特征选择模型由于其高效的全局搜索能力受到越来越多学者的关注.然而，随着数据规模的不断扩大，这些算法所具有的模型不稳定、模型精确度低和局部搜索能力差等局限性，使算法的研究逐步陷入困境.为解决上述问题，本文提出了融合教与学优化算法与局部搜索方法（Local Search,LS）的混合进化Wrapper算法模型（Teaching and Learning-based Optimization-Local Search Algorithm,TLBOLS）.首先，由于传统的教与学优化算法不能直接用于求解特征选择问题，算法在初始化阶段将实数型编码转为二进制编码，然后为保证种群的多样性，在教阶段引入最差个体重启机制，并针对进化班级过程中学习者与教学者两种身份采...     

基于仿生形象思维方法的图像检索算法

在人类对图像的理解过程中,认知结果是由人脑形象思维活动得到的.模仿人脑从形象思维角度,本文提出了一种图像特征提取方法,设计了一种图像检索算法.把每幅图像都映射为高维特征矢量（每个矢量都被看作是高维特征空间中的一个点）,计算高维特征空间中点与点的距离判别函数即可得到图像之间的关系.与其他检索算法比较,实验结果表明该方法在检索效率和检索速度具有优越性.     

量子禁忌搜索算法的研究

针对量子智能算法对高维函数的优化时存在容易陷入局部最优的问题,提出了量子禁忌搜索算法.在量子比特相位增量空间方面,提出了一种按指数级别下降并可动态循环调整的策略;在候选解相位邻域空间方面,提出了一种与禁忌表中最优解有关的可动态调整的划分方法,并增加了候选解局部优化处理方法.为了验证算法的有效性,在高维函数极值问题和多维背包问题进行了仿真,结果表明本文算法收敛速度快,求解精度高.     

基于模糊C均值聚类的锦标赛选择机制与多目标优化研究

本文提出了一种用于多目标优化的进化算法--基于模糊C均值聚类的进化算法（A Fuzzy C-Means Clustering Based Evolutionary Algorithm,FCEA）.在算法的迭代过程中,先利用模糊C均值聚类算法寻找种群的分布结构,通过对每一代种群进行模糊划分,获得每个个体隶属于每一类的隶属度,然后本文设计了一种基于隶属度的锦标赛选择算子,用于从整个种群中选择相似个体进行重组,引导算法进行搜索.实验结果表明,基于隶属度的锦标赛选择算子的应用能够提升算法的性能,与MOEA/D-DE、NSGAⅡ、SPEA2、SMS-EMOA等先进的优化算法进行比较的结果表明,FCEA在求解具有复杂Pareto前沿的多目标优化问题（GLT系列）时具有一定的竞争力.     

多种群协方差学习差分进化算法

种群多样性与交叉算子在差分进化(DE)算法求解全局优化问题中具有重要作用,该文提出一种多种群协方差学习差分进化(MCDE)算法。首先,采用多种群机制的种群结构,利用每一子种群结合相应的变异策略保证进化过程个体多样性。然后,通过种群间的协方差学习,为交叉操作建立一个适当旋转的坐标系统;同时,使用自适应控制参数来平衡种群的勘测与收敛能力。最后,在单峰函数、多峰函数、偏移函数和高维函数的25个基准测试函数上进行测试,并同其他先进的进化算法对比,实验结果表明该文算法相较于其他算法在求解全局优化问题上达到最优效果。     

核不相关空间算法及其在人脸识别中的应用

不相关空间算法是求解不相关鉴别矢量集的快速算法,但是将其应用在人脸识别中将遇到小样本问题,并且算法只是一种线性的特征提取方法。该文提出一种核不相关空间算法,该方法的关键是高维特征空间中不相关空间的计算,对此提出一种简单的计算方法,即根据eigenface中将高阶矩阵计算转化成低阶矩阵计算的思想,将高维特征空间中不相关空间的计算仍归结为标准的特征值分解问题。所提出的算法能够有效地解决小样本问题。在ORL人脸库上的实验结果验证了所提出的算法的可行性和有效性。     

梯度策略自适应差分进化算法

差分进化算法是一种有效求解全局优化问题的方法,为进一步提高求解精度,加快求解过程,文中提出一种梯度策略自适应差分进化算法。该算法是在差分进化算法中加入梯度下降法,使其不仅有较好的全局搜索能力,且具有传统优化方法的快速局部搜索能力,因此具有较高搜索精度和较快的搜索过程。通过对CEC2005测试集中的1～14号测试函数进行仿真实验,并与SaDE,NSDE以及CMAES等算法实验结果进行了对比,结果表明了该算法的有效性。     

一种基于Kriging模型和受限差分进化的电磁结构快速优化算法

进化算法在各类电磁结构优化设计中有着广泛的应用，但由于需要在参数空间中进行随机搜索并仿真试探，优化效率普遍较低.针对这一问题，提出受限差分进化（Differential Evolution，DE）算法与Kriging代理模型相结合的电磁结构快速优化算法.算法根据参考设计结果建立圆柱管道空间，通过参数变换将进化区域限制在管道内部.Kriging模型学习管道内样本及其仿真数据，代替电磁仿真快速预测进化产生下一代种群的响应.相比整个参数空间，该算法DE寻优和Kriging学习的区域被显著减小，优化效率得到提升.通过一个波导双孔定向耦合器的优化设计，表明该方法的求解质量和收敛速度优于现有算法.     

一种基于差分进化的教与学优化算法

教与学优化算法(TLBO)是一种基于教学过程现象的启发式算法。针对求解高维复杂优化问题时容易陷入局部最优的不足,文章提出了一种基于差分进化的TLBO,采用自适应教学因素和基于差分进化的学习过程来提高基本的TLBO性能。9个复杂的测试函数被用来验证所提出的方法的有效性和准确性。实验结果验证了算法的有效性,表明所提出的算法是一种具有优势的优化算法。     

毫微微小区中一种基于分组的资源分配方法

为有效解决毫微微小区间( Femtocell)干扰,采用分布式方式对毫微微小区进行资源管理。首先,对毫微微接入点( FAPs)进行分组。基于Lingo数学建模的思想,提出了一种解决分组优化问题的算法。该算法在使用分支定界算法寻找最优解的同时,通过建立单纯形表剪去偏离最优解方向的分支;其次,每组选择一个簇头为本组内FAPs分配资源,为此,提出了新的子信道分配方法,该方法根据干扰指示矩阵修正子信道分配的情况。仿真结果表明：和其他算法相比,提出的算法不仅能找到分组优化问题的最优解,并且效率更高;另外,提出的资源分配算法不仅减小了用户间干扰,而且提高了户间速率公平。     

Quantum-inspired evolutionary algorithm for analog test point selection

An important problem that arises in fault diagnosis of analog circuit for fault dictionary technique is the test point selection, which is known to be NP-hard. This paper develops a mathematical optimization model for analog test point selection (ATPS) problem and proposes a novel method to solve it based on quantum-inspired evolutionary algorithm (QEA). The proposed method uses the solution produced by the inclusive algorithm to initialize Q-bit individuals and presents a new fitness function to search the global minimum test point set. In addition, an approach for dynamically determining the magnitude of rotation angle is introduced to accelerate the convergent speed. The efficiency of the proposed algorithm is proven by one practical analog circuit example and a group of statistical experiments. Results show that the proposed algorithm, compared with other methods, finds the global minimum set of test points more efficiently and more accurately.     

基于凸优化方法的认知雷达波形设计

针对认知雷达扩展目标检测的问题,提出了一种与目标散射特性相关的相位编码信号设计方法,利用半正定松弛将输出信噪比的优化问题松弛为一个凸优化问题,并利用一维交互迭代搜索逼近原问题的全局最优解。该方法具有收敛速度块、运算量小等优点,能够准确逼近全局最优解。     

Application of Multi-agent Particle Swarm Algorithm in Distribution Network Reconfiguration

An improved algorithm based on Multiagent particle swarm (MAS) is proposed to solve the distribution network reconflguration problem in this paper.The approach is a combination of the learning,competition and cooperation mechanism of multi-agent technology and the strategies of Particle swarm optimization (PSO) algorithm.Using the Von Neumann topology structure in PSO algorithm,each particle represents an agent;each agent not only competes and cooperates with its neighborhood,but also absorbs the evolutionary mechanism of PSO algorithm,so as to share the information with the agent of global optimal.The rules of particle renovating reduce unfeasible solution in the process of particle renovating,and it is able to converge to global optimal accurately and quickly.Test on the IEEE 16-node,32-node and 69-node system shows both a rapid convergence and a good robustness of this proposed approach.     

Evolutionary Multiobjective Optimization for the Pickup and Delivery Problem with Time Windows and Demands

This paper studies an evolutionary algorithm to solve a new multiobjective optimization problem, the Pickup and Delivery Problem with Time Windows and Demands (PDP-TW-D), which is applicable to operational optimization in various mobile network systems. With respect to multiple optimization objectives, PDP-TW-D is to find a set of Pareto-optimal routes for a fleet of vehicles (e.g., mobile robots, drones and autonomous heavy-haulage trucks) in order to serve given transportation requests. The proposed algorithm uses a population of individuals, each of which represents a solution candidate, and evolves them through generations to seek the Pareto-optimal solutions. In addition to the evolution-based global search process, the proposed algorithm allows individuals to improve their optimality in each generation with a local search process, which is designed based on iterative neighborhood search. Experimental results demonstrate that the integration of global and local search processes improves the optimality of individuals and expedites convergence speed. The proposed algorithm outperforms two well-known existing EMOAs, NSGA-II and MOEA/D, in relatively large-scale problems that have up to 400 pickup and delivery locations.     

Self-adaptive mechanism based genetic algorithms for combinatorial optimization problems

To improve the evolutionary algorithm performance, especially in convergence speed and global optimization ability, a self-adaptive mechanism is designed both for the conventional genetic algorithm (CGA) and the quantum inspired genetic algorithm (QIGA). For the self-adaptive mechanism, each individual was assigned with suitable evolutionary parameter according to its current evolutionary state. Therefore, each individual can evolve toward to the currently best solution. Moreover, to reduce the running time of the proposed self-adaptive mechanism based QIGA (SAM-QIGA), a multi-universe parallel structure was employed in the paper. Simulation results show that the proposed SAM-QIGA have better performances both in convergence and global optimization ability.     

高维多目标优化问题融入决策者偏好的集合进化优化方法

高维多目标优化问题普遍存在且非常重要,但是,已有的解决方法却很少.本文提出一种有效解决该问题的融入决策者偏好的集合进化优化方法,该方法首先基于决策者给出的每个目标的偏好区域,将原优化问题的目标函数转化为期望函数;然后,以原优化问题的多个解形成的集合为新的决策变量,以超体积和决策者期望满足度为新的目标函数,将优化问题转化为2目标优化问题;最后,采用多目标集合进化优化方法求解,得到满足决策者偏好且收敛性和分布性均衡的Pareto优化解集.将所提方法应用于4个基准高维多目标优化问题,并与其他2种方法比较,实验结果验证了所提方法的优越性.     

An evolutive framework for server placement optimization to digital twin networks

Digital twin network (DTN) is a foremost enabler for efficient optimization in modern networks, as it owns massive real-time data and requires interaction with the physical network in real-time. When constructing a DTN, it is necessary to deploy many servers in the physical network for digital models' storage, calculation, and communication. Evolutionary algorithms show outstanding global optimization capabilities compared to the constructive heuristic method in such an optimization problem. However, due to the high dimensionality of the problem and the complicated evaluation of the deployment plan, evolutionary algorithms easily fall into the optimum local at a high computational cost, given that the server placement problem is an NP-hard combinatorial optimization problem. In this research, we propose an evolutionary framework for server layout optimization that significantly improves the optimization efficiency of evolutionary algorithms and reduces the algorithm's computational cost. An offline-learning-based approach is used to reduce the search space, and a self-examining guided local search method is proposed to improve the search efficiency. Additionally, a look-up table-based hybrid approach is used for solution evaluation, reducing computational overhead. Experimental results show that the proposed framework and optimization strategy can significantly improve the evolutionary algorithm search efficiency and achieve excellent convergence performance.     

A boolean spider monkey optimization based energy efficient clustering approach for WSNs

Wireless sensor network (WSN) consists of densely distributed nodes that are deployed to observe and react to events within the sensor field. In WSNs, energy management and network lifetime optimization are major issues in the designing of cluster-based routing protocols. Clustering is an efficient data gathering technique that effectively reduces the energy consumption by organizing nodes into groups. However, in clustering protocols, cluster heads (CHs) bear additional load for coordinating various activities within the cluster. Improper selection of CHs causes increased energy consumption and also degrades the performance of WSN. Therefore, proper CH selection and their load balancing using efficient routing protocol is a critical aspect for long run operation of WSN. Clustering a network with proper load balancing is an NP-hard problem. To solve such problems having vast search area, optimization algorithm is the preeminent possible solution. Spider monkey optimization (SMO) is a relatively new nature inspired evolutionary algorithm based on the foraging behaviour of spider monkeys. It has proved its worth for benchmark functions optimization and antenna design problems. In this paper, SMO based threshold-sensitive energy-efficient clustering protocol is proposed to prolong network lifetime with an intend to extend the stability period of the network. Dual-hop communication between CHs and BS is utilized to achieve load balancing of distant CHs and energy minimization. The results demonstrate that the proposed protocol significantly outperforms existing protocols in terms of energy consumption, system lifetime and stability period.