Multi-objective design optimization of composite submerged cylindrical pressure hull for minimum buoyancy and maximum buckling load capacity

This paper presents the design optimization of composite submersible cylindrical pressure hull subjected to 3 MPa hydrostatic pressure. The design optimization study is conducted for cross-ply layups [0s/90t/0u], [0s/90t/0u]s, [0s/90t]s and [90s/0t]s considering three uni-directional composites, i.e. Carbon/Epoxy, Glass/Epoxy, and Boron/Epoxy. The optimization study is performed by coupling a Multi-Objective Genetic Algorithm (MOGA) and Analytical Analysis. Minimizing the buoyancy factor and maximizing the buckling load factor are considered as the objectives of the optimization study. The objectives of the optimization are achieved under constraints on the Tsai-Wu, Tsai-Hill and Maximum Stress composite failure criteria and on buckling load factor. To verify the optimization approach, optimization of one particular layup configuration is also conducted in ANSYS with the same objectives and constraints.     

Multi-objectives nonlinear structure optimization for actuator in trajectory correction fuze subject to high impact loadings

This paper presents an actuator used for the trajectory correction fuze, which is subject to high impact loadings during launch. A simulation method is carried out to obtain the peak-peak stress value of each component, from which the ball bearings are possible failures according to the results. Subsequently, three schemes against impact loadings, full-element deep groove ball bearing and integrated raceway, needle roller thrust bearing assembly, and gaskets are utilized for redesigning the actuator to effectively reduce the bearings' stress. However, multi-objectives optimization still needs to be conducted for the gaskets to decrease the stress value further to the yield stress. Four gasket's structure parameters and three bearings' peak-peak stress are served as the four optimization variables and three objectives, respectively. Optimized Latin hypercube design is used for generating sample points, and Kriging model selected according to estimation result can establish the relationship between the variables and objec-tives, representing the simulation which is time-consuming. Accordingly, two optimization algorithms work out the Pareto solutions, from which the best solutions are selected, and verified by the simulation to determine the gaskets optimized structure parameters. It can be concluded that the simulation and optimization method based on these components is effective and efficient.     

时序驱动的详细布局方法

针对超大规模集成电路布局过程中时序优化问题,提出一种时序驱动的详细布局方法。对设计进行时序分析并获取时序违反路径集合,对路径上两个连续固定单元间的线网进行平滑处理,以减小路径曲折度以及减少线长。再针对每一个可移动单元与其相邻的线网建立二次规划时序模型,求解局部最优布局位置。对于给定的测试电路,实验结果表明,最差的时序违反与总的时序违反均有明显改善,采用ICCAD2015竞赛的测试模板和评价方法,总的时序性能有45~350 min的提升。     

果蝇算法优化的相关向量机在电子装备中的应用

针对电子装备渐变故障预测问题,提出一种基于果蝇算法优化相关向量机的故障预测方法。该方法将原始时间序列数据进行相空间重构处理,并基于折交叉验证和果蝇算法优化相关向量机模型的核函数参数,从而建立故障预测模型,并以某型雷达发射机速调管监测数据对模型性能进行了验证。实验结果表明,相比已有方法,该方法在全局优化、收敛速度、预测精度以及预测可靠性方面都具有一定优势。     

片上多核处理器的双向环设计与物理实现

针对少量强核构成的片上多核处理器,设计了一种双向环。该结构包含5个不同类型的环链路层,分别用于传输命令、大量数据以及小量数据;采用源路由方式,并设计了专门的拥塞控制网络,防止报文的相互覆盖;路由器采用无缓冲无阻塞结构,单节拍通过环的每个跳步,以降低环的传输延迟并实现可预知的确定延迟传输。针对环的链路距离长位宽大的挑战,通过实验选择了合理的中继器插入方法,并采用相邻导线交替插入反相器以信号线反向交错排布等串扰优化等方法,对环进行了物理设计和和长链路进行了延时优化。最终实现结果表明,本文所设计的环达到了1GHz的工作主频,并具备高达256GByte/s的链路带宽,完全满足高性能数字信号处理的需求。     

防空导弹实时优化射表计算的延拓法

文章给出防空导弹轨道实时优化射表的计算方法。首先介绍最优轨道对参数的连续依赖性和优化算法的算法模型的收敛性理论。特别对于每条最优轨道存在一个吸引区域 D* ,使得只要初始迭代轨道属于这个区域 ,由算法模型构造的轨道序列将收敛到所要的最优轨道。根据这个理论 ,设计了计算射表的延拓法。理论证明和计算实践证实由这种算法能有效地计算为实时优化提供初始迭代值的射表     

桁架式水塔结构的优化设计

本文将桁架式水塔结构等效成单自由度弹簧质量系统,考虑其在风载（包括定常与随机波动两部分）作用下,以桁架中的最大应力及频率耦合等作为约束条件的优化设计问题,导出了设计变量的最优化设计数学模型。     

基于松弛序列凸优化的轮式机器人协同轨迹规划

针对威胁环境下的多智能体协同轨迹规划问题,以轮式机器人为对象,研究了基于序列凸优化方法的协同轨迹规划方法。首先通过对轮式机器人模型的分析,给出单独轮式机器人实际物理约束,同时以状态量、控制量加权为性能指标,考虑运动学方程、避障避碰约束、个体物理性能约束、终端约束,建立多轮式机器人协同轨迹规划问题;其次,对运动学方程、避障避碰约束进行凸化,证明了满足凸化后的避障避碰约束的解也满足原始避障避碰约束,并进行了几何直观解释;然后,将凸优化子问题进行离散化与松弛化,采用序列凸优化架构进行问题求解;最后,通过数值仿真,对比了松弛序列凸优化方法与现有非线性优化求解器的求解效率。结果表明,松弛序列凸优化方法在尽可能保证最优性的同时大大缩短了计算时间,具有一定工程意义。     

天基光学短弧观测约束域的粒子群优化定轨方法

针对空间目标定轨问题,提出一种利用两段天基光学短弧观测数据的粒子群优化定轨新算法。在介绍天基光学短弧观测测量帧集、测量约束域及目标函数构造的基础上,为解决已有的基于网格搜索思想寻优的算法存在的多解、局部最优解及运算量过大等问题,提出了一种利用粒子群优化算法在约束域内对目标函数值寻优达到定轨目的的新算法。对算法的性能进行了仿真验证。多次仿真结果表明:该算法大大降低了计算量,且有效地解决了目标函数多解和局部最优解问题,对目标定轨的精度与定轨算法的克拉美罗下限接近。     

键联接的优化设计及可靠性计算

用优化设计的方法建立了键联接的优化数学模型,并对键联接进行了可靠性分析。