深空探测行星借力飞行轨道自动设计与仿真

深空探测任务中直接转移轨道一般需要较大的能量,借力飞行技术可以降低从地球的发射能量,经过多次天体借力到达目标星体,并可以一次探测多个天体.通过多天体的P-Rp图推测可能存在的多天体交会借力飞行方案,求解兰伯特问题,根据圆锥曲线拼接法搜索能量等高线图中匹配的C3,最后找出发射窗口.设计并研制了轨道方案与窗口自动搜索软件,通过EVEEJ的借力飞行方案,及其在STK软件中的仿真,验证设计方法的可行性及软件的正确、可靠性.给出了借力飞行技术的完整设计方法,为进行太阳系深空探测活动的行星借力飞行轨道设计提供参考.     

多次行星借力的小推力飞行控制策略研究

针对小推力深空探测器采用多次行星借力的飞行控制策略优化问题,文章提出了一种将推进段作为整体与滑行段进行打靶拼接的方法,最大限度地减少了优化设计时的待优化参数个数.首先,文章给出了小推力飞行轨道的间接优化设计模型和基于B平面理论的行星借力模型.随后,建立了给定开关机时序条件下的小推力借力飞行控制策略优化模型;最后,采用遗传算法和序列二次规划算法循环对该模型进行优化求解,并以地球–金星–地球–木星小推力飞行轨道控制策略优化设计为例进行仿真分析,仿真结果验证了优化模型的正确性和有效性,表明本文的研究方法可对小推力多次行星借力飞行的控制策略进行优化设计.     

用水与换热网络同步优化综合

通常采用分步策略进行用水与换热网络综合集成,难以实现新鲜水与公用工程的同时权衡。因此本文提出首先序贯求解用水网络综合子问题的非线性规划和换热网络综合子问题的混合整数非线性规划,以获取可行初值;然后,进行流程完备结构拓扑,并构造连接方程实现相关变量的传递;最后,联立用水与换热网络优化综合问题,基于GAMS平台采用DICOPT求解器,通过MATLAB接口控制求解程序流程,采用随机初值策略强化获得全局最优解的机会,以获取高置信度的全局最优解。最后通过对单杂质和多杂质系统实例计算,结果表明本文提出的方法能实现用水和用能网络的权衡,不但获得了比文献结果更好的优化网络方案,而且找出了包括文献结果的若干近优方案,为过程设计提供了多种备用选择。     

不确定工业过程运行指标异步更新强化学习决策算法

运行指标决策问题是实现工业过程运行安全和生产指标优化的关键.考虑到多运行指标决策问题求解的复杂性和工业过程生产条件动态波动引发生产指标状态的不确定性,提出了一种策略异步更新强化学习算法自学习决策运行指标,并给出算法收敛性的理论证明.该算法在随机自适应动态规划框架下,利用样本均值代替计算生产指标状态转移概率矩阵,因此无需要求生产指标状态转移概率矩阵已知.并且通过引入时钟和定义其阈值,采用集中式策略评估、多策略异步更新方式用以简化求解多运行指标决策问题,提高强化学习的学习效率.利用可测量数据,自学习得到的运行指标能够保证生产指标优化,并且限制在规定范围之内.最后,采用中国西部某大型选矿厂的实际数据进行仿真验证,表明该方法的有效性.     

基于高斯伪谱法的分区联立动态优化策略

最优控制轨迹通常由不同类型的弧段组成,采用数值方法求解这类问题时,其优化轨迹的不连续性可能导致求解失败,为此基于高斯伪谱法,提出一种分区联立动态优化求解策略.该策略通过对优化时域合理分区,避免了由于控制轨迹不连续而导致的病态优化问题.另外,通过联立求解分区后的优化问题,能够满足各分区间的连接条件,而且可以使用较少的节点获得高质量的优化解.最后以抗生素发酵过程为例验证了所提出算法的有效性.     

多工况大规模工业过程模型参数估计

在对工业过程进行建模时,模型规模庞大,现场运行工况点较多。构造全联立参数估计模型,模型规模随着工况数的增加成倍扩大,在初值较差的情况下,求解收敛性差。针对此问题,本文提出1种目标序贯式参数估计方法,按照特定规则分批调整目标,逐步添加约束,使得目标分批逼近设定值,最后在新的初值基础上求解联立参数估计优化命题。本文以PTA氧化反应工段模型为例,对其反应动力学常数进行参数估计,结果表明,目标序贯式参数估计收敛性强。     

计算机博弈中序贯不完美信息博弈求解研究进展

计算机博弈是人工智能的果蝇和通用测试基准.近年来,序贯不完美信息博弈求解一直是计算机博弈研究领域的前沿课题.围绕计算机博弈中不完美信息博弈求解问题展开综述分析.首先,梳理计算机博弈领域标志性突破的里程碑事件,简要介绍4类新评估基准,归纳3种研究范式,提出序贯不完美信息博弈求解研究框架;然后,着重对序贯不完美信息博弈的博弈模型和解概念进行调研,从博弈构建、子博弈和元博弈、解概念以及评估3方面进行简要介绍;接着,围绕离线策略求解,系统梳理算法博弈论、优化理论和博弈学习3大类方法,围绕在线策略求解,系统梳理对手近似式学习、对手判别式适变和对手生成式搜索3大类方法;最后,从环境、智能体(对手)和策略求解3个角度分析面临的挑战,从博弈动力学和策略空间理论、多模态对抗博弈和序贯建模、通用策略学习和离线预训练、对手建模(剥削)和反剥削、临机组队和零样本协调5方面展望未来研究前沿课题.对于当前不完美信息博弈求解问题进行全面概述,期望能够为人工智能和博弈论领域相关研究带来启发.     

基于定制内点法的多无人机协同轨迹规划

为提高多无人机(Unmanned aerial vehicles, UAV)协同轨迹规划(Cooperative trajectory planning, CTP)效率,在解耦序列凸优化(Sequential convex programming, SCP)方法基础上,提出一种高效求解凸优化子问题的定制内点法.首先引入松弛变量,构建子问题的等价描述形式,并推导该形式下的子问题最优性条件.然后在预测-校正原对偶内点法的框架下,构建一套高效求解最优性条件方程组的计算流程以降低子问题计算复杂度,并利用约束矩阵特征提出一种快速计算原对偶搜索方向的方法以提高规划效率.仿真结果表明,在解耦序列凸优化框架下,定制内点法可将协同轨迹规划耗时降低一个数量级,达到秒级.     

小推力轨道转移快速优化设计

在研究电推进系统中,为满足小推力转移轨道高精度在线生成的要求,伪光谱方法在电推进小推力轨道转移优化设计中的应用。首先对小推力航天器轨道转移最优控制问题模型进行无量纲化处理,以提高优化算法求解精度。然后采用基于勒让德-高斯-兰伯特配置点的勒让德伪光谱方法,将最优控制问题离散成约束参数优化问题,再利用适于求解大尺度非线性规划问题的TOMLAB/SNOPT优化软件包进行求解。通过数值仿真计算,求解生成了满足各类约束条件的小推力转移轨道,并利用余向量映射定理及极小值原理验证了所得轨道转移控制量的最优性。结果表明,勒让德伪光谱优化算法具有对初始猜测值不敏感、收敛速度快、精度高等优点。     

基于形状法和伪谱法的小推力借力优化研究

小推力借力飞行轨道优化是一个多变量多约束的非线性优化问题,根据形状法和伪谱法,提出一种混合优化策略,分为全局优化和局部优化两个阶段进行。在全局优化阶段采用LT-PGA模型,即通过求解形状法小推力Lambert问题,搜索满足约束条件的小推力发射窗口,得到发射、借力和到达时间点。在局部优化阶段采用伪谱法得到推力控制率,用连接点设置解决借力行星处状态量的不连续问题。数值仿真结果表明,改进方法不用事先指定推力开关机序列,优化效率高,为初始设计阶段小推力借力飞行的轨道优化问题提供有益参考。     

A combined shape control procedure of cable mesh reflector antennas with optimality criterion and integrated structural electromagnetic concept

A combined shape control procedure with optimality criterion and integrated structural electromagnetic concept for cable mesh reflector antennas is presented in this study. Using the optimality criterion, the shape control algorithm drives the distorted surface towards the ideal shape. The optimality criterion is implemented by pseudo inverse of sensitivity matrix of surface nodal displacements with respect to cable member dimensions to accelerate the iterative convergence. The following integrated structural electromagnetic design is performed to make good electromagnetic performance by a sequential quadratic programming optimization model. A distorted offset cable mesh reflector antenna is employed to show its effectiveness.     

Formulation and multigrid solution of Cauchy-Riemann optimal control problems

The formulation of optimal control problems governed by Cauchy-Riemann equations is presented. A distributed control mechanism through divergence and curl sources is considered with the boundary conditions of mixed type. A Lagrange multiplier framework is introduced to characterize the solution to Cauchy-Riemann optimal control problems as the solution of an optimality system of four first-order partial differential equations and two optimality conditions. To solve the optimality system, staggered grids and multigrid methods are investigated. It results that staggered grids provide a natural collocation of the optimization variables and second-order accurate solutions are obtained. The proposed multigrid scheme is based on a coarsening by a factor of three that results in a nested hierarchy of staggered grids. On these grids a distributed-Gauss-Seidel and gradient-based smoothing scheme is employed. Results of numerical experiments validate the proposed optimal control formulation and demonstrate the effectiveness of the staggered-grids multigrid solution procedure.     

Generalized Hopfield network based structural optimization using sequential unconstrained minimization technique with additional penalty strategy

This paper presents and examines a neuron-like framework of the generalized Hopfield network (GHN) that is capable to solve nonlinear engineering optimization problems with mixed discrete, integer and real continuous variables. The sequential unconstrained minimization technique (SUMT) was applied to construct the GHN for dealing with the design constraints. An additional penalty function for dealing with the discrete and integer variables was then imposed on the formulation of SUMT to construct an energy function of GHN for formulating the neuron-like dynamical system. The numerical solution process for such a dynamic system is simply solving a set of simultaneous first-order ordinary differential equations (ODE) that is the main feature of this optimization method. The experimental examples showed the presenting strategy is reliable. The suitable values or the adaptation technique for some parameters in computation was discussed in the paper. The presenting strategy indeed provides an alternative way of handling the engineering optimization dynamically and expands the usage of ODE. An asymmetrical three-bar truss design, a reinforced concrete beam design and a 10-bar structural design are contributed to illustrate the presenting neuron-like network method.     

The finite mass mesh method

The finite mass method is a purely Lagrangian scheme for the spatial discretisation of the macroscopic phenomenological laws that govern the flow of compressible fluids. In this article we investigate how to take into account long range gravitational forces in the framework of the finite mass method. This is achieved by incorporating an extra discrete potential energy of the gravitational field into the Lagrangian that underlies the finite mass method. The discretisation of the potential is done in an Eulerian fashion and employs an adaptive tensor product mesh fixed in space, hence the name finite mass mesh method for the new scheme. The transfer of information between the mass packets of the finite mass method and the discrete potential equation relies on numerical quadrature, for which different strategies will be proposed. The performance of the extended finite mass method for the simulation of two-dimensional gas pillars under self-gravity will be reported. Communicated by: G. Wittum     

A superelement formulation for the efficient layout design of complex multi-component system

In view of the prohibited computing time and the complexity of design procedure, a superelement formulation (SEF) is proposed to deal with the simultaneous optimization of component placement and the framework topology. In the iterative design process, each component is modelled as a movable superelement so that the sensitivity analysis with respect to the location design variables can be largely simplified by the SEF. Moreover, based on the Kuhn–Tucker optimality condition, two decomposition strategies are developed as variant approaches for the simultaneous design of multi-component system. By means of numerical examples, these approaches are compared to show their capability and efficiency for the system compliance minimization.     

Point-to-point control near heteroclinic orbits: Plant and controller optimality conditions

In this paper we consider the simultaneous optimization of the controller and plant in a one degree-of-freedom system. In particular we are interested in optimal trajectories between fixed points connected by heteroclinic orbits. We find that designing the plant dynamics to have a heteroclinic connection between target states enables a low energy transfer between the states. We use a nested optimization strategy to find the optimal plant dynamics and control effort for the transition. Additionally, we uncover plant optimality conditions which reduce the complexity of the optimization.     

Mesh refinement for shape optimization

The numerical solution of shape optimization problems is considered. The algorithm of successive optimization based on finite element techniques and design sensitivity analysis is applied. Mesh refinement is used to improve the quality of finite element analysis and the computed numerical solution. The norm of the variation of the Lagrange augmented functional with respect to boundary variation (residuals in necessary optimality conditions) is taken as an a posteriori error estimator for optimality conditions and the Zienkiewicz—Zhu error estimator is used to improve the quality of structural analysis. The examples presented show meaningful effects obtained by means of mesh refinement with a new error estimator.     

A topological derivative method for topology optimization

We propose a fictitious domain method for topology optimization in which a level set of the topological derivative field for the cost function identifies the boundary of the optimal design. We describe a fixed-point iteration scheme that implements this optimality criterion subject to a volumetric resource constraint. A smooth and consistent projection of the region bounded by the level set onto the fictitious analysis domain simplifies the response analysis and enhances the convergence of the optimization algorithm. Moreover, the projection supports the reintroduction of solid material in void regions, a critical requirement for robust topology optimization. We present several numerical examples that demonstrate compliance minimization of fixed-volume, linearly elastic structures.     

An adaptive neural network strategy for improving the computational performance of evolutionary structural optimization

This work is focused on improving the computational efficiency of evolutionary algorithms implemented in large-scale structural optimization problems. Locating optimal structural designs using evolutionary algorithms is a task associated with high computational cost, since a complete finite element (FE) analysis needs to be carried out for each parent and offspring design vector of the populations considered. Each of these FE solutions facilitates decision making regarding the feasibility or infeasibility of the corresponding structural design by evaluating the displacement and stress constraints specified for the structural problem at hand. This paper presents a neural network (NN) strategy to reliably predict, in the framework of an evolution strategies (ES) procedure for structural optimization, the feasibility or infeasibility of structural designs avoiding computationally expensive FE analyses. The proposed NN implementation is adaptive in the sense that the utilized NN configuration is appropriately updated as the ES process evolves by performing NN retrainings using information gradually accumulated during the ES execution. The prediction capabilities and the computational advantages offered by this adaptive NN scheme coupled with domain decomposition solution techniques are investigated in the context of design optimization of skeletal structures on both sequential and parallel computing environments.