约束优化问题的改进遗传算法设计

遗传算子是影响遗传算法优化效果的重要因素,针对目前遗传算法研究中对约束优化问题求解的不足,提出基于退火思想的退火选择算子和加权适应度算子,并给出了退火选择算子和加权适应度算子设计方法及其计算过程.在此基础上与现有的遗传算子结合,提出一种新的改进遗传算法,分析了改进遗传算法与基于罚函数遗传算法之间在原理上的区别.最后以两个测试函数为算例对算法进行了性能测试,结果表明改进的遗传算法具有良好的优化性能,能获得更好的优化结果.     

基于混合智能学习算法的函数优化研究

为了改善人工神经网络在优化计算中的一些缺陷和提高遗传算法的局部搜索能力及收敛性能,提出了一种混合智能学习算法,采用遗传算法和误差反向传播算法(BP算法)相结合,将BP算法以一个算子的形式插入到遗传算法中,以提高利用人工神经网络和遗传算法进行优化计算的搜索能力和收敛性能;通过对实例函数的优化计算,对插入BP算子的遗传算法和传统遗传算法的优化结果进行了比较分析,结果表明BP算子的插入对遗传算法的优化性能、收敛速度和收敛精度有较大改善.     

适应噪声环境的改进遗传算法

遗传算法领域考虑噪声存在的情况非常复杂,设计合理的改进方法提高遗传算法的性能非常必要.从改进算法设计策略的角度,基于引进平滑滤波的方法提出了改进方法,并利用计算机仿真实验,与现有方法比较,结果表明改进的遗传算法不仅提高了算法的收敛性能,并且提高了算法的计算速度.归纳并通过实验表明了现有方法的有效性.最后,从信噪比这样一个新的角度对噪声环境下的遗传算法进行分析,得出了有效的改进方案.     

基于递阶和小生境的离散分段遗传算法

为了解决简单遗传算法过早收敛的问题,并进一步改善简单遗传算法的寻优质量,在分析递阶遗传算法和小生境遗传算法的基础上,提出了离散分段遗传算法.该方法在微观上,采用了递阶遗传算法的递阶编码方式和小生境的选择思想.宏观上,通过分层多级寻优操作来适当加快遗传算法的寻优速度.该算法非常适合解决多峰值优化问题,同时也能够有效地修复早熟现象的影响,加快收敛速度.实验表明该方法在性能方面明显优于简单遗传算法.     

基于BP算子的自适应混合遗传算法研究

将误差反向传播算法(BP算法)以一个算子的形式融入到遗传算法中,以提高遗传算法的优化性能.其基本思路是:在遗传算法收敛速度放慢时启用BP算子,把新一代群体作为BP算子的初始值再用BP算法训练网络,这样交替运行BP算法和遗传算法,直到达到问题要求的精度.通过对4例实验函数的优化,证明了混合遗传算法具有良好的收敛性和稳定性.实验对插入BP算子的遗传算法和传统遗传算法的优化结果进行了比较分析,结果表明BP算子的插入对遗传算法的优化性能、收敛速度和收敛精度方面都有了很大的改进.     

基于预测的遗传算法

遗传算法被广泛应用于解决各类优化问题.常规的遗传算法易于陷入局部最优,其收敛速度也较慢.为了提高常规遗传算法的优化性能,将预测的概念引入遗传算法的循环过程,提出基于预测的遗传算法框架;并以人工神经网络算法作为预测算法,提出了一种基于神经网络预测的遗传算法.通过优化8个典型的函数优化问题,将该算法与常规遗传算法的性能进行了比较;结果显示该算法具有很强的全局优化能力,能有效地增强种群的多样性和进化速度,明显优于常规遗传算法.     

自适应变异遗传算法及其应用

为了改善变异操作在遗传算法中的作用,提出自适应变异遗传算法,其变异操作能根据种群进化代数和个体的适应度值自适应地确定每个个体的变异概率,从而在保留遗传算法当前最优解的同时,维持了群体的多样性,提高了算法的全局搜索能力.与传统遗传算法相比,自适应变异遗传算法的离线性能和在线性能都有较大的改善.本文在实际应用中,将自适应变异遗传算法应用于估计动力学参数取得了较好的结果.     

基于遗传算法的多连接表达式并行查询优化

多连接表达式的并行查询优化是提高数据库性能的关键问题之一.提出了使用遗传算法来解决多连接表达式的并行查询优化问题.为了提高查询处理器的执行效率,采用启发式规则来搜索最优的多连接表达式并行调度执行计划.文中给出了详细的测试结果和性能分析.实验结果表明,结合启发式知识的遗传算法是解决多连并行查询优化的有效途径,对提高数据库的性能起到重要作用.     

一种新的模糊自适应模拟退火遗传算法

针对遗传算法收敛速度慢、容易"早熟"等缺点,结合模糊推理、模拟退火算法和自适应机制,提出一种改进的遗传算法--模糊自适应模拟退火遗传算法(FASAGA),并分析了该算法的性能和特点,实验研究表明,该算法比标准的遗传算法(SGA)具有更快的收敛速度和寻优效果.     

元胞遗传算法的收敛性分析和收敛速度估计

元胞遗传算法是空间结构化种群的遗传算法,将遗传操作限制在相邻个体之间进行,限制优势基因的扩散速度,保持种群的多样性,改善遗传算法的性能.但是,目前有关元胞遗传算法收敛性的分析还较缺乏.文中根据元胞遗传算法的特性,建立元胞遗传算法的吸收态Markov链模型,证明元胞遗传算法的收敛性.提出元胞遗传算法的首达最优解期望时间的估算方法,并估计标准同步元胞遗传算法首达最优解期望时间的上下界.     

模糊神经网络的DNA算法训练

提出了基于DNA计算和遗传算法的DNA遗传算法,给出了DNA遗传算法的结构，讨论了遗传操作算子,利用DNA遗传算法对FNN进行学习，比采用梯度型算法和遗传算法有更高的学习精度和更快的收敛速度,该算法有全局收敛性避免了采用梯度型学习算法训练FNN时固有的局部收敛问题，同样，该算法加速了FNN的训练，能够在线应用．     

Testing the performance of a 2D nearest point algorithm with genetic algorithm generated Gaussian distributions

Genetic algorithms have successfully been used in automatic software testing. Particularly programming errors and inputs that conflict with time constraints can be found. In this paper, the idea of genetic algorithm based software testing is broadened to algorithm performance testing. It is shown how the best and worst case performance of the algorithms can be found effectively. This information can be further utilized when comparing and improving algorithms. In this paper, the proposed test method is introduced and the advantages of using genetic algorithms are discussed. Furthermore, the proposed method is applied to a 2D nearest point algorithm, which is tested by optimizing the parameters of 2D Gaussian distributions using genetic algorithms in order to find the best and worst case distributions and the corresponding performances.     

A parallel evolutionary algorithm for the vehicle routing problem with heterogeneous fleet

Nowadays genetic algorithms stand as a trend to solve NP-complete and NP-hard problems. In this paper, we present a new hybrid metaheuristic which uses parallel genetic algorithms and scatter search coupled with a decomposition-into-petals procedure for solving a class of vehicle routing and scheduling problems. The parallel genetic algorithm presented is based on the island model and its performance is evaluated for a heterogeneous fleet problem, which is considered a problem much harder to solve than the homogeneous vehicle routing problem.     

粗粒度并行遗传算法性能分析

依据实验来分析影响并行遗传算法性能的因素得到的结论缺乏理论上的说服力．通过对粗粒度并行遗传算法加速比公式的分析，提出了影响并行遗传算法性能的关键因素，同时否定了以迁移率作为评价并行遗传算法性能指标的合理性，并通过实难进一步验证结论的正确性．得到的结论为提高遗传算法的并行化效率提供了可靠的依据。     

基于生态种群捕获竞争模型的进化遗传算法

将协同进化的思想运用到遗传算法，是对遗传算法的一大改进和拓展，借鉴此思想，提出了一种生态种群捕获竞争的协同进化模型和基于此模型的改进的进化遗传算法(PCGA)。实验结果表明，该算法在改善未成熟收敛和提高收敛速度方面都具有良好的性能。     

用改进的遗传算法解决ALB问题

针对生产装配线平衡问题,提出一种改进的遗传算法.算法采用缩放适应度法、随机普遍取样的选择策略、线性可变的杂交和变异算子.使用PB语言实现了这一应用平台,给出了系统的功能结构图和主要的数据结构,并结合实例给出了ALB-2问题的解决方案.实例对比证明,改进的算法很好地解决了简单遗传算法易早熟的问题,大大改善了简单算法的性能.     

Steady-state genetic algorithms for discrete optimization of trusses

This paper presents the applications of steady-state genetic algorithms to discrete optimization of trusses. It is mathematically formulated as a constrained nonlinear optimization problem with discrete design variables. Discrete design variables are treated by a two-stage mapping process which is constructed by the mapping relationships between unsigned decimal integers and discrete values. With small generation gap and careful modification, steady-state genetic algorithms can significantly reduce the computational effort and promote the computational efficiency. The effectiveness, robustness and fast convergence of steady-state genetic algorithms are demonstrated through several examples. The performance of four crossover operators is also compared.     

遗传算法在肉类近红外光谱分析中波长选择的应用

论文旨在研究遗传算法波长选择对近红外光谱数据建立的模型的预测性能的影响.文中选用遗传算法对光谱数据进行波长选择,选择后建立偏最小二乘法（PLS）模型,再将得到的模型和没有经过波长选择的全波长建立的PLS模型进行对比.结果表明,经过遗传算法波长选择后建立的模型预测性能优越,预测误差明显减少.所以,将遗传算法应用到波长选择中是可行的,可以继续推广.     

Genetic algorithms for flowshop scheduling problems

In this paper, we apply a genetic algorithm to flowshop scheduling problems and examine two hybridizations of the genetic algorithm with other search algorithms. First we examine various genetic operators to design a genetic algorithm for the flowshop scheduling problem with an objective of minimizing the makespan. By computer simulations, we show that the two-point crossover and the shift change mutation are effective for this problem. Next we compare the genetic algorithm with other search algorithms such as local search, taboo search and simulated annealing. Computer simulations show that the genetic algorithm is a bit inferior to the others. In order to improve the performance of the genetic algorithm, we examine the hybridization of the genetic algorithms. We show two hybrid genetic algorithms: genetic local search and genetic simulated annealing. Their high performance is demonstrated by computer simulations.     

A novel hybrid genetic algorithm using local optimizer based on heuristic pattern move

A new hybrid genetic algorithm with the significant improvement of convergence performance is proposed in this study. This algorithm comes from the incorporation of a modified microgenetic algorithm with a local optimizer based on the heuristic pattern move. The hybridization process is implemented by replacing the two worst individuals in the offspring obtained from the conventional genetic operations with two new individuals generated from the local optimizer in each generation. Some implementation-related problems such as the selection of control parameters in the local optimizer are addressed in detail. This new algorithm has been examined using six benchmarking functions, and is compared with the conventional genetic algorithms without the local optimizer incorporated, as well as the hybrid algorithms incorporated with the hill-climbing method in terms of convergence performance. The results show that the proposed hybrid algorithm is more effective and efficient to obtain the global optimum. It takes about 6.4%-74.4% of the number of generations normally required by the conventional genetic algorithms to obtain the global optimum, while the computation cost for reproducing each new generation has hardly increased compared to the conventional genetic algorithms. Another advantage of this new algorithm is the implementation process is very simple and straightforward. There are no extra function evaluations and other complex calculations involved in the added local optimizer as well as in the hybridization process. This makes the new algorithm easy to be incorporated with the existing software packages of genetic algorithms so as to further improve their performance. As an engineering example, this new algorithm is applied for the detection of a crack in a composite plate, which demonstrates its effectiveness in solving engineering practical problems.