二部图的(g,f)匹配

本文证明了二部图存在(g,f)匹配和f 因子的充要条件以及有关的几个结果,并且给出了求二部图的最大(g,f)匹配、最小(g,f)匹配和最小权最大f 匹配、最小权(g,f)匹配、最大权(g,f)匹配的算法。     

Minimizing makespan in a multimode multiprocessor shop scheduling problem

We study the problem of multimode scheduling tasks on dedicated processors, with the objective of minimizing the maximum completion time. Each task can be undertaken in one among a set of predefined alternative modes, where each mode specifies a required set of dedicated processors and a processing time. At any time each processor can be used by a single task at most. General precedence constraints exist among tasks, and task preemption is not allowed. The problem consists of assigning a mode and a starting time to each task, respecting processor and precedence constraints, to minimize the time required to complete all tasks. The problem is NP-hard in several particular cases. In previous works, we studied algorithms in which a solution was obtained by means of an iterative procedure that combines mode assignment and sequencing phases separately. In this paper, we present some new heuristics where the decision on the mode assignment is taken on the basis of a partial schedule. Then, for each task, the mode selection and the starting time are chosen simultaneously considering the current processor usage. Different lower bounds are derived from a mathematical formulation of the problem and from a graph representation of a particular relaxed version of the problem. Heuristic solutions and lower bounds are evaluated on randomly generated test problems. © 1999 John Wiley & Sons, Inc. Naval Research Logistics 46: 893–911, 1999     

Two exact algorithms for the vehicle routing problem on trees

This article describes a heuristic and two exact algorithms for several classes of vehicle routing problems defined on tree networks. These include capacitated and time‐constrained vehicle routing problems. One of the exact algorithms is based on the computation of bin packing lower bounds. The other uses column generation. The first algorithm performs better on problems containing small customer demands and in which all vehicles are identical. Otherwise, the second algorithm is more powerful and more versatile. © 1999 John Wiley & Sons, Inc. Naval Research Logistics 46: 75–89, 1999     

整数线性规划问题的一个新算法

本文给出求解整数线性规划问题的一个算法。基本思想是通过求出其伴随线性规划问题的最优单纯形表,把整数线性规划化成正整数系数的不定方程,然后从不定方程的非负整数解集中选取一组满足整数线性规划的约束条件的解,作为整数线性规划的最优解。     

基于遗传BP神经网络的装备研制风险分析方法

简要介绍了GA和BP算法,利用遗传算法全局性搜索的特点,改变BP算法依赖梯度信息的指导来调整网络权值的方法,寻找最为合适的网络连接权和网络结构,提出了遗传算法优化BP神经网络的思路及其数学模型.最后,结合对某型航空装备的风险源的分析,利用此优化模型进行了验证.结果表明了该方法的可行性,为装备研制风险分析提供了一种新思路.     

果蝇算法和5种群智能算法的寻优性能研究

截止到目前为止进化式算法主要有遗传算法、蚁群算法、鱼群算法、免疫算法、粒子群算法.这些算法已经被广泛地用于寻优,但都有各自的缺点,导致其不易被用于解决实际问题.某学者提出了一种新群智能算法——果蝇算法.对该算法的起源进行分析,并将该算法与其他算法对比,通过仿真分析各个算法寻优性能.重点分析果蝇算法的寻优性能,得出果蝇算法简单、参数少、易调节、计算量小、寻优精度较高,从而较容易被用于解决实际问题,对于复杂问题算法可能不稳定.指出该算法的缺点,提出应改进的地方,对其应用前景作了概括.     

舰船固定磁场分量分离的一种新方法

为了准确了解舰船的固定磁场分布,进而对舰船磁性防护作出指导,用等效磁偶极子的方法对舰船固定磁场分量分离的方法进行了研究,通过对固定磁场垂直分量总和的曲线进行分析,根据磁偶极子组合产生磁场规律的先验知识,选择最佳的磁偶极子的排列方式,然后对磁偶极子磁矩范围进行估计,最后利用遗传算法这种全局优化算法求出磁源的参数,进而分离出固定磁场各个分量。仿真计算结果表明,该方法能够比较好地搜索到磁偶极子的最佳数目、位置和磁矩,可以用于舰船固定磁场的分离。     

Interval scheduling: A survey

In interval scheduling, not only the processing times of the jobs but also their starting times are given. This article surveys the area of interval scheduling and presents proofs of results that have been known within the community for some time. We first review the complexity and approximability of different variants of interval scheduling problems. Next, we motivate the relevance of interval scheduling problems by providing an overview of applications that have appeared in literature. Finally, we focus on algorithmic results for two important variants of interval scheduling problems. In one variant we deal with nonidentical machines: instead of each machine being continuously available, there is a given interval for each machine in which it is available. In another variant, the machines are continuously available but they are ordered, and each job has a given “maximal” machine on which it can be processed. We investigate the complexity of these problems and describe algorithms for their solution. © 2007 Wiley Periodicals, Inc. Naval Research Logistics, 2007     

多核数字信号处理器矩阵乘卷积算法性能评测

矩阵乘卷积算法能够为各种卷积配置提供高性能基础实现,是面向给定芯片进行卷积性能优化的首要选择。针对国防科技大学自主研制的飞腾异构多核数字信号处理器(digital signal processor, DSP)芯片的特征以及矩阵乘卷积算法自身的特点,提出了一种面向多核DSP架构的高性能并行矩阵乘卷积实现算法ftmEConv。该算法由输入特征图转换、卷积核转换、矩阵乘以及输出特征图转换这四个均运行在通用多核DSP上的并行化部分构成,通过有效挖掘通用DSP核中功能单元的潜力来提升各个部分的性能。实验结果表明,ftmEConv实现了高达42.90%的计算效率,与芯片上的其他矩阵乘卷积算法实现相比,获得了高达7.79倍的性能加速。     

基于遗传算法的嵌入式软件测试用例生成系统研究

综合考虑了嵌入式软件的特点，将遗传算法应用于搜索全面覆盖软件特殊路径的测试用例生成的问题，合理构建了嵌入式软件测试用例自动生成系统，给出了系统流程图，改进了部分算子，使搜索效果得以加强，并分析了交叉率、变异率等参数的取值范围，通过改变不同的参数值得到对比图，以确定适合该系统的最佳参数值，并针对实际汇编语言编写的嵌入式软件指定路径进行仿真测试，从理论和实践上证明了该思路的正确性和可行性。