基于非线性递减选择策略的人工蜂群算法

针对基本人工蜂群算法种群多样性难以保持,进化速度慢等问题,提出了一种基于非线性递减选择策略的人工蜂群算法.算法在雇佣蜂阶段采用非线性递减选择策略以提高种群的多样性,进而改善种群的全局勘探能力;在跟随蜂阶段由全局最优解引导搜寻新解,以提高种群的局部开发能力;侦察蜂采用贴近最优解的策略以提高生成新解的质量,加速种群进化.改进的三个阶段改善了算法的寻优性能,最后通过实验对比与分析,验证了该算法的有效性.     

交叉型状态空间模型进化算法的全局收敛性分析

基于状态空间模型的进化算法(SEA)是一种新颖的实数编码进化算法,在实际工程优化问题中取得了良好的优化效果。为促进SEA的理论及应用研究,对交叉型SEA(SCEA)的全局收敛性进行了研究,得出SCEA不是全局收敛的结论。通过改变状态进化矩阵的构造方式和提出弹力搜索操作,得到改进交叉型SEA(SMCEA),利用齐次有限Markov链对SMCEA的全局收敛性进行了证明。最后利用两个测试函数对算法进行实验分析,结果表明,SMCEA在收敛速度、最优解搜索能力和运算时间等方面都有较大改善,验证了SMCEA的有效性,得到了SMCEA优于遗传算法(GA)和SCEA的结论。     

基于差异化邻域策略的分解多目标进化算法*

子问题邻域对基于分解的多目标进化算法性能影响较大.当邻域过大时,种群繁殖产生的新解偏离Pareto解集,在更新子问题时,新解与邻域内旧解的比较次数增多,算法的计算复杂度增加;当邻域过小时,算法容易陷入局部最优.为了解决上述问题,文中提出基于差异化邻域策略的分解多目标进化算法(MOEA/D-DN),通过分析不同大小的邻域对算法性能的影响,选择合适的参数.并根据每个子问题的权重向量与中心向量的偏角,为各子问题设置不同大小的邻域,合理分配算法资源,提高算法搜索全局最优解的速率.在2维ZDT系列和3维、5维DTLZ系列测试函数上的实验表明,MOEA/D-DN 的收敛速度与收敛性能均有明显提高,算法的计算资源分配更合理,所获解集整体质量更优.     

多策略协方差矩阵学习差分进化算法

针对差分进化算法（DE）存在的早熟收敛和搜索停滞的问题,提出了多策略协方差矩阵学习的差分进化算法。通过协方差矩阵建立特征坐标系,通过在特征坐标系中执行变异和交叉操作,来充分利用当前种群的分布信息以及各变量之间的关系,保证种群能朝着全局最优解的方向进化;根据历史进化信息来选择变异策略的方式使得个体能选择当前最合适的变异策略,提高找到最优解的概率;交叉概率的自适应也一定程度上平衡算法的全局探索能力和局部探索能力。对算法的收敛性进行了证明,同时将算法在CEC2017测试集上进行了仿真实验,并将实验结果跟其他优秀的差分进化算法进行了对比,对比结果表明了该算法的有效性。     

基于改进全局人工蜂群算法的WSN节点定位研究

无线传感器网络(Wireless Sensor Network,WSN)系统性能的提高,离不开对WSN中每一个传感器节点地理位置的精准定位。全局人工蜂群算法在基本人工蜂群算法的基础上,在邻域搜索后将迭代最优解添加到新解的更新公式中,提高了算法的开发能力。但将其应用于WSN节点位置求解时,存在计算时间长、收敛不稳定的问题。提出一种改进的全局人工蜂群算法,在邻域搜索后对新解进行衡量,若新解适应值在可接受的范围内,与迭代最优解进行交叉操作;若新解适应值较好,不与迭代最优解进行交叉操作;若新解适应值较差,舍弃新解。这较好地平衡了算法的探索和开发能力。求解WSN节点位置时,证明了该算法有更快的收敛速度和更好的收敛效果。     

从动态适值空间补偿信息: 一种抗病态合作协同进化算法

针对合作协同进化算法(CCEA) 动态适值空间的特点, 研究信息补偿方法以消除由问题分解所导致的病态现象, 并提出基于动态多种群进化策略的抗病态CCEA. 每个协进化种群可动态分离出多个变化的子种群, 利用它们同时获取多个全局或局部最优解作为交互信息, 以实现信息补偿. 针对引发病态行为的标准测试函数, 与3 种典型CCEA 进行比较分析, 实验结果表明所提出算法能有效克服病态现象, 具有良好的全局优化能力.     

基于文化算法的符号网络全局不平衡度计算

针对已有符号网络不平衡度计算方法大都只关注局部网络单元的平衡信息,没有考虑网络更大范围乃至全局角度的平衡,无法揭示网络中的不平衡区域这一问题,提出基于文化算法的符号网络全局不平衡度计算方法。该方法利用伊辛自旋玻璃模型描述符号网络的全局状态,将不平衡度的计算转换为一个优化问题,并设计一种具有双层进化结构的文化算法——CA-SNB进行求解。首先,该算法采用遗传算法进行种群空间进化;其次,在信度空间中记录较优个体,并采用贪婪算法提取状况知识;最后,利用状况知识引导种群空间的进化,在保证种群多样性的基础上提高了收敛速度。实验表明,与遗传算法和矩阵变换算法相比,CA-SNB能较快地收敛到最优解,具有较高鲁棒性,在计算全局不平衡度的同时识别不平衡区域。     

求解job shop调度问题的量子粒子群优化算法*

针对粒子群优化算法搜索空间有限、容易出现早熟现象的缺陷,提出将量子粒子群优化算法用于求解作业车间调度问题。求解时,将每个调度按照一定的规则编码为一个矩阵,并以此矩阵作为算法中的粒子;然后根据调度目标确定目标函数,并按照量子粒子群优化算法的进化规则在调度空间内搜索最优解。仿真实例结果证明,该算法具有良好的全局收敛性能和快捷的收敛速度,调度效果优于遗传算法和粒子群优化算法。     

集成协方差矩阵自适应进化策略与差分进化的优化算法

不同智能优化算法在求解优化问题时通常表现出显著的性能差异.差分进化(DE)算法具备较好的全局搜索能力,但存在收敛慢、效率低的不足,协方差矩阵自适应进化策略(CMA–ES)局部搜索能力强,具备旋转不变性,但容易陷入局部最优,因此, DE和CMA–ES之间具有潜在的协同互补能力.针对上述问题,提出了一种集成协方差矩阵自适应进化策略与差分进化的优化算法(CMADE).在CMADE框架中, DE算法负责全局搜索, CMA–ES算法进行局部搜索.通过周期性解交换机制实现CMA–ES和DE两个算法间协同交互和反馈控制.在解交换时,从DE种群中选择优秀个体,利用CMA–ES算法在优秀个体周围进行局部搜索.同时在DE和CMA–ES的混合种群中,综合考虑解的多样性和最优性,选取一定比例的解作为DE算法的新种群进行全局搜索,实现全局搜索与局部搜索的动态平衡.将CMADE算法与CMA–ES, DE, SaDE, jDE, EPSDE, ACODE和SHADE算法在CEC2014标准测试集上进行比较实验.结果表明, CMADE整体性能显著优于其它比较算法.     

求解job-shop调度问题的量子粒子群优化算法

针对粒子群优化算法搜索空间有限、容易出现早熟现象的缺陷,提出将量子粒子群优化算法用于求解作业车间调度问题.求解时,将每个调度按照一定的规则编码为一个矩阵,并以此矩阵作为算法中的粒子;然后根据调度目标确定目标函数,并按照量子粒子群优化算法的进化规则在调度空间内搜索最优解.仿真实例结果证明,该算法具有良好的全局收敛性能和快捷的收敛速度,调度效果优于遗传算法和粒子群优化算法.     

Improved spotted hyena optimizer with space transformational search for training pi-sigma higher order neural network

Spotted hyena optimizer (SHO) is a recently developed swarm-based algorithm in the field of metaheuristic research, for solving realistic engineering design constraint and unconstrained difficulties. To resolve complicated nonlinear physical world tasks, at times, SHO reveals deprived performance concerning to explorative strength. So, to enhance the explorative strength along with exploitation in the search region, an attempt has been made by proposing the enhanced version of classical SHO. The suggested method is designated as space transformation search (STS)-SHO. In STS-SHO, a new evolutionary technique named as STS technique has been incorporated with original SHO. The suggested method has been assessed by IEEE CEC 2017 benchmark problems. The efficacy of the said method has been proven by using standard measures such as given performance metrics in CEC 2017, complexity analysis, convergence analysis, and statistical implications. Further as real-world application, the said algorithm has been applied to train pi-sigma neural network by means of 13 benchmark datasets considered from UCI depository. From the article it can be concluded that the suggested method STS-SHO is an effective and trustworthy algorithm, which has the ability to resolve real-life optimization complications.     

An Admission Control-based Benefit Optimization Model for Mobile Communications: the Effect of a Decision Time Budget

In mobile communication systems, a soft handoff (SHO) technique is used to optimize the quality and capacity of communications. However, because the handoff process incurs a high overhead there must be a tradeoff between the system capacity and the handoff overhead. In this paper, we propose a benefit optimization model for mobile communications. The model tries to maximize the overall system capacity by considering SHO process overhead and quality of service requirements jointly. We first construct a framework of admission policies and devise an appropriate admission control policy, which is then used to analyze the system benefit. The service rate is defined by three measures: the call blocking ratio, system load, and admit-to-existence ratio; while the solution quality is defined by the gap between the upper bound and lower bound of the objective function value. By applying iteration-based Lagrangian relaxation as a solution approach, a time budget is allocated to each iteration so that admission control can be implemented. To fulfill the continuous admission process requirements in the long-term, users’ demands are randomly distributed via a simulation process. The goal of this paper is to investigate the effect of the admission control policy on the system benefit, service rate and solution quality. Experiment results are presented to demonstrate the efficacy of both the proposed model and the solution approach.     

Spotted hyena optimizer with lateral inhibition for image matching

Multimedia Tools and Applications - A hybrid spotted hyena optimizer (SHO) based on lateral inhibition (LI) is proposed, it has been applied to solve complication image matching problems. Lateral...     

Proposing two new metaheuristic algorithms of ALO-MLP and SHO-MLP in predicting bearing capacity of circular footing located on horizontal multilayer soil

In this study, for the issue of shallow circular footing’s bearing capacity (also shown as F_ult), we used the merits of artificial neural network (ANN), while optimized it by two metaheuristic algorithms (i.e., ant lion optimization (ALO) and the spotted hyena optimizer (SHO)). Several studies demonstrated that ANNs have significant results in terms of predicting the soil’s bearing capacity. Nevertheless, most models of ANN learning consist of different disadvantages. Accordantly, we focused on the application of two hybrid models of ALO–MLP and SHO–MLP for predicting the F_ult placed in layered soils. Moreover, we performed an Extensive Finite Element (FE) modeling on 16 sets of soil layer (soft soil placed onto stronger soil and vice versa) considering a database that consists of 703 testing and 2810 training datasets for preparing the training and testing datasets. The independent variables in terms of ALO and SHO algorithms have been optimized by taking into account a trial and error process. The input data layers consisted of (i) upper layer foundation/thickness width (h/B) ratio, (ii) bottom and topsoil layer properties (for example, six of the most important properties of soil), (iii) vertical settlement (s), (iv) footing width (B), where the main target was taken F_ult. According to RMSE and R², values of (0.996 and 0.034) and (0.994 and 0.044) are obtained for training dataset and values of (0.994 and 0.040) and (0.991 and 0.050) are found for the testing dataset of proposed SHO–MLP and ALO–MLP best-fit prediction network structures, respectively. This proves higher reliability of the proposed hybrid model of SHO–MLP in approximating shallow circular footing bearing capacity.

     

Artificial memory optimization

To solve some complicated optimization problems, an artificial memory optimization (AMO) is constructed based on the human memory mechanism. In AMO, a memory cell is used to trace an alternative solution of a problem to be solved; memorizing and forgetting rules of the human memory mechanism are used to control state transition of each memory cell; the state of a memory cell consists of two components, one is the solution state which associates with an alternative solution being traced; another is the memory state which associates with the memory information resulting from tracing results, where the memory residual value (MRV) is stored; the states of memory cells are divided into three types: instantaneous, short- and long-term memory state, each of which can be strengthened or weakened by accepted stimulus strength. If the solution state of a memory cell has transferred to a good position, its MRV will increase, and then the memory cell is not easily to be forgotten; when the solution state of a memory cell is at sticky state, its MRV will decrease until the memory cell is forgotten; this will effectively prevent invalid iteration. In the course of evolution, a memory cell may strive to evolve from the instantaneous, short-term memory state to long-term memory state, it makes search to be various. Because AMO has 6 operators at the curent version, it has wider adaptability to solve different types of optimization problems. Besides, these operators are automatically dispatched according to their executing efficiency. Results show that AMO possesses of strong search capability and high convergence speed when solving some complicated function optimization problems.     

Pareto-MEC算法及其收敛性分析

介绍了一种新的多目标进化算法——Pareto-MEC。将基本MEC和Pareto思想结合起来处理多目标问题。提出了局部Pareto最优解集与局部Pareto最优态集概念，并利用概率论的基本理论证明了趋同过程产生的序列强收敛于局部Pareto最优态集。数值试验验证了Pareto-MEC算法的有效性。     

A Lyapunov approach to analysis of discrete singular systems

In this paper, a new type generalized Lyapunov equation for discrete singular systems is proposed. Then it is applied to study problems such as pole clustering, controllability and observability for discrete singular systems. First, some necessary and sufficient conditions for pole clustering are derived via the solution of this new type Lyapunov equation. Further, the relationship between the solution of the Lyapunov equation and structure properties of discrete singular systems will be investigated based on these results. Finally, a type of generalized Riccati equation is proposed and its solution is used to design state feedback law for discrete singular systems such that all the finite poles of the closed-loop systems are clustered into a specified disk.     

一种协调勘探和开采的遗传算法：收敛性及性能分析

提出了一种新的遗传算法结构。在该结构中,每一代的新种群由保留种 群、繁殖种群的随机种群三部分组成,而它们的相对数量则由不同的参数进行控制,这体现了该算法在运行过程中对搜索空间勘探和开采操作的协调和权衡。通过把该算法建模为齐次的有限Markov链,该文证明了该算法具有全局收敛性。对试验数据的分析表明,该算法能够有效协调算法对问题解空间的勘探和开采操作,因而在处理复杂问题时表现出较高的性能。     

A complete analytical solution to the equationTA - FT = LCand its applications

In this note, a complete, analytical, and restriction-free solution with complete and explicit freedom of the matrix equationTA - FT = LCis proposed. Here(A, C)is given and is observable, andFis in the Jordan form with arbitrary given eigenvalues. This solution appears to be new because it can be applied directly to obtain significantly better solutions to the following three basic design problems: 1) 2-D system eigenvalue assignment; 2) function observer design; and 3) state feedback eigenstructure design, as shown in this note.     

基于增量式有限混合模型的多目标状态极大似然估计

提出了增量式有限混合模型来提取概率假设密度滤波器序贯蒙特卡罗实现方式中的多目标状态. 该模型以增量方式构建, 其混合分量采用逐个方式插入其中. 采用极大似然准则来估计多目标状态. 对于给定分量数目的混合模型, 应用期望极大化算法来获得参数的极大似然解. 在新分量插入混合模型时, 保持已有混合模型的参数不变, 仍旧采用极大似然准则从候选新分量集合中选择新插入分量. 新分量插入混合步和期望极大化算法拟合混合参数步交替应用直到混合分量数目达到概率假设密度滤波器的目标数目估计值. 利用k-d树生成插入到混合模型的新分量候选集合. 增量式有限混合模型统一了分量数目变化趋势和粒子集合似然函数的变化趋势, 有助于一步一步地搜寻混合模型的极大似然解. 仿真结果表明, 基于增量式有限混合模型的概率假设密度滤波器状态提取算法在多目标跟踪的应用中优于已有的状态提取算法.