An exact network flow formulation for cell‐based evacuation in urban areas

In the literature two common macroscopic evacuation planning approaches exist: The dynamic network flow approach and the Cell–Transmission–Based approach. Both approaches have advantages and disadvantages. Many efficient solution approaches for the dynamic network flow approach exist so that realistic problem instances can be considered. However, the consideration of (more) realistic aspects (eg, density dependent travel times) results in non‐linear model formulations. The Cell‐Transmission‐Based approach on the other hand considers realistic traffic phenomena like shock waves and traffic congestion, but this approach leads to long computational times for realistic problem instances. In this article, we combine the advantages of both approaches: We consider a Cell‐Transmission‐Based Evacuation Planning Model (CTEPM) and present a network flow formulation that is equivalent to the cell‐based model. Thus, the computational costs of the CTEPM are enormously reduced due to the reformulation and the detailed representation of the traffic flow dynamics is maintained. We investigate the impacts of various evacuation scenario parameters on the evacuation performance and on the computational times in a computational study including 90 realistic instances.     

领悟社会支持与高中生学习投入的关系:自尊和未来教育规划的链式中介作用

目的:探究领悟社会支持对学习投入的影响,以及自尊和未来教育规划的链式中介效应。方法:采用领悟社会支持量表、自尊量表、未来教育规划问卷和学习投入量表,对446名高中生进行施测。结果:①各变量两两之间均存在显著地正相关(r=0.32~0.50,p<0.001);②领悟社会支持能正向预测高中生学习投入;③在领悟社会支持对高中生学习投入的预测中存在三条路径:一是自尊的中介作用;二是未来教育规划的中介作用;三是自尊和未来教育规划的链式中介作用,效果值分别为:14.81%,27.78%和16.67%。结论:领悟社会支持不仅直接影响高中生学习投入,还通过自尊和未来教育规划影响高中生学习投入。     

基于MDFT框架的海上能源通道航线动态决策建模与仿真

针对常规海上航线决策没有考虑决策选择过程的问题,引入多选择决策场理论（ MDFT）,构建了海上能源通道航线动态决策模型。基于信息完备、信息不确定和特殊情况等三种情景想定,进行了海上能源通道航线规划决策及其实验仿真。结果表明,决策过程受时间压力和信息质量的制约,时间压力会制约决策质量,同时还可能引起“偏好逆转”现象,信息不确定也会导致确定性决策。该模型不仅综合考虑了海上能源通道航线选择时需关注的各类要素,且模型参数可动态调整,进而能合理模拟真实的海上状况和决策过程。     

振动遗传算法在无人机三维航路规划的算法研究

针对代价函数权重需要根据环境变化而变化的问题,结合飞行约束条件提出归一化的代价函数,当环境发生变化时,不用再修改代价函数,增强了算法的鲁棒性。为了弥补传统定步长寻径算法耗时长的缺陷,设计了一种基于B样条曲线与遗传算法的高时效寻径算法。利用遗传算法在地图中所寻合适的控制点,再结合B样条曲线生成航路。为了增强遗传算法的全局搜索能力,遗传算法中加入振动法则,使得种群在进化中后期依旧保持一定的多样性。仿真结果表明该算法与精英蚁群算法相比,规划时间大幅缩短;与振动遗传算法相比,航路代价明显降低。     

Resourcing for defense: Solving the roles and missions puzzle

One of the most important issues facing the post‐Cold War U.S. defense establishment concerns the future allocation of combat tasks and responsibilities among different branches of the armed forces. The challenge is to reduce unnecessary redundancy across roles and missions when resources are highly constrained, without compromising military effectiveness. Defining the policy problem as one of resource allocation rather than operational effectiveness, we develop a methodology for allocating roles and missions. Our methodology focuses at the highest level of force aggregation and uses a mathematical programming model to produce cross‐service cross‐mission trade‐offs that will yield the best total force combat and non‐combat potential within resource consumption constraints.     

MONEY IN SOUTHERN AFRICAN POLITICS

“Lack of money is the root of all evil”

George Bernard Shaw

In a environment where party financing, private funding in particular, is generally a laissez-faire business, as in most Southern African countries, there is the real risk that interest groups and wealthy individuals will buy influence in political parties and in so doing erode public confidence in the political system. Because they are neither open to public scrutiny nor subject to any legislative restraints, huge private donations can, and often do, come with strings attached. This lack of transparency and openness, as well as mechanisms to prevent it, provides ample ground for influence peddling, where those who pay the piper call the tune. Left unchecked, this party funding scenario will continue to be a serious indictment of the region's democratisation projects, crippling democracy and stunting economic development. This article consid     

基于多传感器信息系统的仿人机器人步态规划算法改进研究

摘要：基于视觉传感器、角度传感器和力／力矩传感器组成的多传感器信息系统,对仿人机器人理想的步态规划算法进行了改进．以双目视觉立体标定原理处理视觉传感信息,从而判断目标距离及目标路径的可达性;通过角度传感器获取实时旋转角度,经参数调整减小执行误差;根据力／;5矩传感信息对支撑脚踝关节的侧摆角度进行增量式补偿．经实验验证,改进后的算法对机器人的步态稳定性控制具有良好的效果．     

基于车辆稳态动力学特性的汽车动态轨迹规划

针对高动态环境下驾驶辅助系统（ADAS）的轨迹规划问题,提出基于车辆稳态动力学特性的动态轨迹规划算法。该算法首先在交通车轨迹预估基础上建立搜索空间,接着利用车辆稳态动力学模型作为轨迹发生器,在搜索空间中对轨迹发生器产生的轨迹进行评价,最后根据驾驶意图和优化条件选择最优轨迹。文章采用基于最优加速度预瞄理论的轨迹跟踪算法,建立了七自由度车辆动力学模型,并在Simulink环境下搭建仿真平台,进行超车仿真实验。实验结果表明,本文提出的算法可行、有效。     

资源不确定军事任务计划预测调度模型与算法

针对军事任务计划执行环境中普遍存在的资源不确定性,提出了基于资源缓冲区的军事任务计划预测调度算法。算法首先基于平台有效资源功能向量进行任务—平台分配,分配过程中通过调节平台有效资源功能向量,获得不同的平台资源缓冲区分配方案;然后基于NSGA-Ⅱ算法框架对军事任务计划进行多目标求解,进而获得问题的Pareto最优解集。文章通过仿真算例对算法的可行性和有效性进行验证,实验结果表明,该算法能够有效求解资源不确定军事任务计划问题。     

通信受限下分布式多传感器协同跟踪算法

在分布式多传感器目标跟踪系统中,由于局部融合中心(LFC)的物理限制(如:有限的频率信道、处理器容量有限等),只能接收有限个传感器的传送数据。此外,信息传输的方式也将影响传感网的使用寿命,因此,研究了通信受限下的分布式多传感器目标协同跟踪问题。首先对监视区内分布的传感器进行聚类分簇形成若干个子网,接着从通信能耗的角度出发,对传感器采集信息的传递路径进行最优路径规划;进而对子网局部状态进行估计,在子网信息融合中,分别采用最大距离和、最大化信息增量两种准则进行最佳传感器选择,最后通过各子网全局航迹融合实现分布式多传感器协同跟踪。仿真验证了算法的有效性。