动圈式永磁直线直流电机的精英保留多种群遗传优化算法

为了实现陆基无人机电磁弹射器高机动性及其直线弹射电机的高功率密度,针对动圈式永磁直线直流电机,提出精英保留的多种群遗传算法(Multi-Population Genetic Algorithm with Elite Retention, MPGAER)的电机最大功率密度优化方法。以磁通密度和电流密度为约束条件,利用其搜索能力强、收敛速度快的特点优化电机的结构参数,并与磁路法初始设计结果和传统遗传算法优化结果进行比较。结果表明:与磁路法初始设计相比,MPGAER能使电机质量减少6.25%,功率密度提高10%,电机动态性能得到提高;MPGAER优化设计的电机功率密度高于遗传算法设计结果,所提方法有效地解决了在优化过程中出现易收敛于局部最优点和寻优效果差的问题。     

An optimal recursive method for various inventory replenishment models with increasing demand and shortages

We establish various inventory replenishment policies to solve the problem of determining the timing and number of replenishments. We then analytically compare various models, and identify the best alternative among them based on minimizing total relevant costs. Furthermore, we propose a simple and computationally efficient optimal method in a recursive fashion, and provide two examples for illustration. © 1997 John Wiley & Sons, Inc. Naval Research Logistics 44: 791–806, 1997     

引入区域量化的相控阵天线维修优化模型

针对大型相控阵天线维修成本高、多部件维修时机难以确定、模型仿真计算困难等问题,提出一种引入区域量化的维修优化模型。建立天线阵列的区域量化模型,将天线阵列维修问题转化成一个由不同区域阵元组成的k/n系统维修优化问题。以使用可用度为约束条件,以单位时间维修费用最小为优化目标,建立引入区域量化的大型相控阵天线多部件视情维修决策优化模型,并通过算例仿真与分析进行了模型验证。结果表明:提出的维修优化模型真实有效,能够很好地解决分布不均匀的相控阵天线维修优化问题。     

预报基座结构噪声的有限元方法研究

有限元法适用的频率范围与单元的大小有关,为了用有限元准确地描述高频振动(结构噪声),一个振动波长中至少划分4个单元.由此,根据研究问题的频率范围更合理地确定单元的大小,并适当选取计算的规模,预报了某基座模型2000Hz内的结构噪声,此研究具有工程应用意义.     

A bidirectional heuristic for maximizing the net present value of large‐scale projects subject to limited resources

We consider the scheduling of large‐scale projects to maximize the project net present value given temporal and resource constraints. The net present value objective emphasizes the financial aspects of project management. Temporal constraints between the start times of activities make it possible to handle practical problem assumptions. Scarce resources are an expression of rising cost. Since optimization techniques are not expedient to solve such problems and most heuristic methods known from literature cannot deal with general temporal constraints, we propose a new bidirectional priority‐rule based method. Scheduling activities with positive cash flows as early and activities with negative cash flows as late as possible results in a method which is completed by unscheduling techniques to cope with scarce resources. In a computational experiment, we compare the well‐known serial generation scheme where all activities are scheduled as early as possible with the proposed bidirectional approach. On the basis of a comprehensive data set known from literature containing instances with up to 1002 activities, the efficiency of the new approach is demonstrated. © 2003 Wiley Periodicals, Inc. Naval Research Logistics, 2003     

快速解命中的一种有效方法

基于对命中函数的考察,提出了一种新的命中解解算方法,实际应用及仿真计算结果表明,该方法能明显提高解算精度和缩短迭代时间。     

恒加试验的最优线性不变估计

给出了恒加试验的最优线性无偏估计与最优线性不变估计。与有关文献给出的二步估计相比,这两个估计有较大改进,并且在计算上简单可行。     

目标定位和跟踪的神经网络最小二乘估计

首先概述了单目标定位和跟踪的现有研究结论和新问题,接着叙述了神经网络最小二乘估计方法,最后给出了它对第１类和第２类火控系统应用的数学模型     

GF(q)上多元多项式与钟控序列

本文讨论了有限域GF(q)(q=p~α,p≥2为素数,α≥1为正整数)上多元多项式与钟控序列的周期和线性复杂度的关系。当前馈函数g(x_1,x_2,…,x_n)∈GF(q)[x_1,x_2,…x_n]为一次多项式时,我们给出了钟控序列到达最大周期与线性复杂度的充要条件。     

线性分式规划原始单纯形算法有限性问题

本文用一个数值例子说明用[1] 和[2] 中的原始单纯形算法求解退化的线性分式规划(LFP) 可能会出现基循环,从而得不到最优解。于是就此情形引入了Bland规则,并建立了有限性算法。