基于凸优化理论的三维微分对策制导规律求解

针对三维微分对策制导律(DGL)求解问题,引入凸优化理论,将DGL求解归结到Hamilton系统的求解,设计了DGL求解算法,通过对代价函数梯度特征的凸分析,推导出对策系统鞍点存在的充要条件和求解方法,解决了以往通过对微分对策模型简化求解导致的模型不能客观反映作战过程的问题.     

面向多最优解组合优化问题的决策求解算法

针对具有固定物品总和、多最优解特征的组合优化问题,以固定总和实数子集问题和购买鸡翅问题为例,给出了这类多最优解组合优化问题的形式化表示。在分析枚举等经典算法基础上,提出了基于整数状态表示和实数状态表示的0-1决策递归搜索多最优解动态规划算法。针对该算法在最优解数量较大时,时间复杂度趋向O(mⁿ)的问题,提出了基于相同决策路径合并和基于0-x决策的两种改进算法。实验中两种改进算法的计算时间基本符合与O(nb+nm)的正比关系,表明对于这类多最优解组合优化问题具有良好的求解性能。     

伺服系统速度环控制器参数的自整定及优化

针对交流伺服系统速度环控制器参数自整定及优化的需求,为使工程技术人员避免繁琐的调参过程,提出了一种基于转动惯量辨识的速度环控制器参数自整定及优化方法。首先,采用频率法分析伺服系统速度环控制器参数的设计规则;随后,采用遗忘因子递推最小二乘法,对控制器参数设计所需的系统转动惯量进行辨识;最后,在此基础上利用设计的变步长迭代算法,完成控制器参数寻优过程。仿真结果表明,采用遗忘因子递推最小二乘法,能够有效辨识出电机轴端的转动惯量。提出算法进行整定及优化后,速度环控制器参数能使系统具有良好的动态响应和鲁棒性。     

Influence of shaped charge structure parameters on the formation of linear explosively formed projectiles

The research of LEFP (linear explosive forming projectile) is of great value to the development of new warhead due to its excellent performance. To further improve the damage ability of the shaped charge warhead, a special shell overhanging structure was designed to increase the charge based on the traditional spherical charge, in which case the crushing energy of LEFP could be guaranteed. LS-DYNA was used to simulate different charge structures obtained by changing the number of detonation points, the length of shell platform, the radius of curvature and the thickness of liner. The RSM (response surface model) between the molding parameters of LEFP and the structural parameters of charge was established. Based on RSM model, the structure of shaped charge was optimized by using multi-objective genetic algorithm. Meanwhile, the formation process of jet was analyzed by pulsed X-ray photography. The results show that the velocity, length-diameter ratio and specific kinetic energy of the LEFP were closely related to the structural parameters of the shaped charge. After the optimization of charge structure, the forming effect and penetration ability of LEPP had been significantly improved. The experimental data of jet velocity and length were consistent with the numerical results, which verifies the reliability of the numerical results.     

盲信道均衡的动量恒模算法

介绍了盲信道均衡恒模算法(CMA)的基本原理,对恒模算法的缺点进行了分析,通过修正恒模算法的迭代公式提出一种动量恒模算法(MCMA)。利用QPSK信号,采用计算机仿真的方法对动量恒模算法与恒模算法进行盲均衡性能比较,模拟结果显示,MCMA算法比CMA算法在收敛速率、降低稳态均方误差和符号间干扰具有明显的优势,有效地改进了信道均衡性能。     

含软判决引导的修正恒模盲均衡算法

在CMA算法基础上,提出了一种混合的修正恒模盲均衡算法MCMA SDD。新算法对CMA算法进行修正,同时引入软判决思想。理论分析和计算机模拟表明,该算法克服了CMA收敛速度慢、稳态误差大的缺点,同时可以纠正相位偏转。与CMA SDD算法相比,具有更快的收敛速度和更低的稳态均方误差。     

基于GA的模糊PID控制在质量矩导弹系统上的应用

针对三轴稳定的质量矩导弹姿态控制系统,提出了一种基于遗传算法进行参数寻优的模糊PID控制方法。在建立姿控系统数学模型的基础上,首先设计了姿态控制的优先级函数,然后根据实际的姿态控制系统,设计了遗传算法的性能指标函数和具体的优化策略,最后给出了模糊PID控制器的设计思路。仿真结果说明了该方法的有效性。     

驱护舰编队防空火力分配决策模型研究

驱护舰编队防空火力分配的确定和优化是最大限度地发挥驱护舰编队的整体作战能力的关键。为了高效地发挥驱护舰编队火力优势,用运筹学理论建立了驱护舰编队防空火力分配决策问题的优化模型,分析了各种解决此模型的方法的优缺点,并尝试运用粒子群优化理论(PSO)求解该模型。实例计算分析表明,所提出的方法较好地解决了驱护舰防空火力分配决策问题,为舰艇编队指挥员进行防空火力分配决策提供了一种有效的参考处理方法。     

Optimizing the quality control station configuration

We study unreliable serial production lines with known failure probabilities for each operation. Such a production line consists of a series of stations, existing machines, and optional quality control stations (QCSs). Our aim is to decide on the allocation of the QCSs within the assembly line, so as to maximize the expected profit of the system. In such a problem, the designer has to determine the QCS configuration and the production rate simultaneously. The profit maximization problem is approximated assuming exponentially distributed processing times, Poisson arrival process of jobs into the system, and the existing of holding costs. The novel feature of our model is the incorporation of holding costs that significantly complicated the problem. Our approximation approach uses a branch and bound strategy that employs our fast dynamic programming algorithm for minimizing the expected operational costs for a given production rate as a subroutine. Extensive numerical experiments are conducted to demonstrate the efficiency of the branch and bound procedure for solving large scale instances of the problem and for obtaining some qualitative insights.

相似文献   

基于树分解/合并策略的QoS多播路由方法

针对多约束QoS多播路由的NP-Complete特性,提出一种可控的多播树分解与合并策略,使多播树的生成在兼顾低费用的同时具有多样性,有效克服多播路由优化的局部极值问题。基于该策略设计蚁群算法,分解蚂蚁种群为与多播目标点相对应的蚂蚁子群,引入基于“死点”惩罚和多播树奖惩的信息素更新机制,提高了算法的收敛速度。仿真实验表明,该方法能有效地解决QoS多播路由问题,且随着网络规模的增大保持了良好的性能。