分组坐标轮换法与机械优化设计

minimax 问题是工程优化设计中普遍存在的问题。本文首次采用分组坐标轮换法求解该问题,通过分析获得了该算法收敛的充分条件(如果收敛,还可计算最大轮换次数)。一些数值计算验证了文中的结论,本文还把该算法用于四连杆实现函数机构的优化设计.     

How old is too old? An economic approach to replacing military aircraft

The US Air Force is asking when to replace its aging aircraft. We develop a framework to identify economically optimal replacement strategies that recognizes cost trade-offs and incorporates age effects. We also preview a stochastic methodology. We use the framework to identify an optimal strategy, defined by the replacement age, for a generic fleet and conduct a sensitivity analysis. Quantitative illustrations show that the range of strategies that provides close-to-optimal outcomes widens as the operating and support (O&S) cost growth rate decreases and the ratio of the acquisition price to the initial O&S cost increases. A wider range implies more decision-making leeway.     

表格驱动程序设计方法

本文提出了一种具有广泛应用前景的程序设计方法——表格驱动(Formdriven)程序设计方法(以下简称表格驱动法)。文中首先描述了表格驱动法的基本思想、解的结构;分析了这种方法的特点与适用性;剖析了基于表格驱动法的办公自动化软件ALL-IN-1的设计思想,并指出表格驱动法特别适合于设计软件开发工具。     

基于加权动态规划和航迹关联的小目标检测技术

低信噪比下运动小目标的检测一直是成像目标检测中的一个热点问题。提出了一种新的小目标检测算法,采用方向加权的动态规划算法和二值航迹关联,克服了低信噪比下目标机动和传感器的不稳定对小目标检测的影响,实现了对深空背景下运动方式任意、速度最大达1像素/帧的弱小运动目标的有效检测。     

基于组织的社会技术系统程序设计模型及语言

互联网上越来越多软件系统呈现出社会与技术交融、环境多样和开放、系统动态和演化等特点,如何支持这类复杂软件系统的构造是软件工程面临的一项重要挑战。本文将社会技术系统视为多Agent组织,把社会组织学的概念和思想引入到面向Agent程序设计范畴,提出了以Agent、组织、角色和职位为核心的基于组织程序模型ORAP来支持社会技术系统的开发;通过对多Agent组织不同层次动态性的识别和分析,设计了一组程序设计机制来实现社会技术系统的动态演化,包括角色绑定和组合机制、基于角色的交互机制以及组织的自我管理机制;论文提出了基于上述模型和机制的社会技术系统程序设计语言OragentL,给出了OragentL程序在组织、角色和组合三个层次的语法形式定义,介绍了OragentL的编译器及其运行支撑环境OragentBurg,并通过案例分析和演示展示了本文研究工作的有效性。     

When Programming Trumps Policy and Plans: The Case of the US Department of the Navy11

Traditionally, policy and planning have been institutionally weak in the Naval Staff (Office of the Chief of Naval Operations – OPNAV). In their place, the N8 (Programming) has dominated resource decision-making, and, by default, decisions relating to policy and planning. Recent uncertainty over defense authorization and appropriations has resulted in calls for a greater role to be played by the N3/5, Policy and Plans Directorate. The article argues that reform of the Department of the Navy’s planning process is urgently needed. OPNAV’s weak planning and overly dominant programming practices are compared with those of the Departments of the Army and Air Force and are shown to be out of conformance with them. The article concludes with specific and detailed recommendations for reform of both the current planning and programming processes.     

分布式约束优化问题及其求解算法

分布式约束优化问题(DCOP)能够对多智能体系统(MAS)中的各种分布式推理任务进行建模,广泛应用于分布式规划、调度、资源分配等问题中。首先从DCOP的概念出发,引入一个典型的DCOP实例,在此基础上对DCOP问题求解的两类主流算法进行了详细介绍和比较分析。针对DCOP对现实问题建模中出现的部分集中式、硬约束、开放式、隐私和anytime等5个方面的问题进行了阐述,并介绍了相应的扩展算法。在动态实时问题,自稳定性与误差容错以及在物理分布式环境下仿真等问题仍需进一步研究。     

求解整数规划问题的一个搜索方法

本文整数规划问题给出一种搜索方法,它类似于求解连续变量优化问题的迭代方法,从一个好的初始可行解出发,寻找一个搜索方向,沿着这个方向求出改进的可行解,然后又开始下一次迭代。此方法简单易行,可以求出问题的最优解或近似最优解,对于整数线性规划问题和整数非线性规划问题的求解都适用,并且容易推广到求解大规校整数线性规划问题。文中附有计算例子,说明方法是有效的。     

高级程序设计语言综述

为读者选择语言提供了总体思路,给出了高级语言的发展的趋势。     

Programming challenges and impediments to reform: identifying pragmatic solutions

Using two recently published essays by the current writer that assesses the dismal record of performance of Planning, Programming, and Budgeting System in enabling communist legacy defence institutions in Central and Eastern Europe to develop viable defence plans, this essay argues the need for deep reforms in the region’s defence institutions. To guide this reform effort, pragmatic solutions are suggested to improve the ability of these organisations to produce viable defence plans. Recommended reforms are: (1) conduct conceptual and cultural “audits,” (2) make operational and financial data central to decision-making, (3) change current organisational sociology, (4) examine planning methods and practices, and (5) stress the need to adopt policy frameworks to drive the operation of defence institutions.