A non-intrusive parallel-in-time adjoint solver with the XBraid library

In this paper, an adjoint solver for the multigrid-in-time software library XBraid is presented. XBraid provides a non-intrusive approach for simulating unsteady dynamics on multiple processors while parallelizing not only in space but also in the time domain (XBraid: Parallel multigrid in time, http://llnl.gov/casc/xbraid). It applies an iterative multigrid reduction in time algorithm to existing spatially parallel classical time propagators and computes the unsteady solution parallel in time. Techniques from Automatic Differentiation are used to develop a consistent discrete adjoint solver which provides sensitivity information of output quantities with respect to design parameter changes. The adjoint code runs backwards through the primal XBraid actions and accumulates gradient information parallel in time. It is highly non-intrusive as existing adjoint time propagators can easily be integrated through the adjoint interface. The adjoint code is validated on advection-dominated flow with periodic upstream boundary condition. It provides similar strong scaling results as the primal XBraid solver and offers great potential for speeding up the overall computational costs for sensitivity analysis using multiple processors.

     

A contact force model considering constant external forces for impact analysis in multibody dynamics

The adjoint method is an elegant approach for the computation of the gradient of a cost function to identify a set of parameters. An additional set of differential equations has to be solved to compute the adjoint variables, which are further used for the gradient computation. However, the accuracy of the numerical solution of the adjoint differential equation has a great impact on the gradient. Hence, an alternative approach is the discrete adjoint method, where the adjoint differential equations are replaced by algebraic equations. Therefore, a finite difference scheme is constructed for the adjoint system directly from the numerical time integration method. The method provides the exact gradient of the discretized cost function subjected to the discretized equations of motion.     

Adjoint-based fluid flow control and optimisation with lattice Boltzmann methods

A lattice Boltzmann (LB) framework to solve fluid flow control and optimisation problems numerically is presented. Problems are formulated on a mesoscopic basis. In a side condition, the dynamics of a Newtonian fluid is described by a family of simplified Boltzmann-like equations, namely BGK–Boltzmann equations, which are linked to an incompressible Navier–Stokes equation. It is proposed to solve the non-linear optimisation problem by a line search algorithm. The needed derivatives are obtained by deriving the adjoint equations, referred to as adjoint BGK–Boltzmann equations. The primal equations are discretised by standard lattice Boltzmann methods (LBM) while for the adjoint equations a novel discretisation strategy is introduced. The approach follows the main ideas behind LBM and is therefore referred to as adjoint lattice Boltzmann methods (ALBM). The corresponding algorithm retains most of the basic features of LB algorithms. In particular, it enables a highly-efficient parallel implementation and thus solving large-scale fluid flow control and optimisation problems. The overall solution strategy, the derivation of a prototype adjoint BGK–Boltzmann equation, the novel ALBM and its parallel realisation as well as its validation are discussed in detail in this article. Numerical and performance results are presented for a series of steady-state distributed control problems with up to approximately 1.6 million unknown control parameters obtained on a high performance computer with up to 256 processing units.     

A Dual Approach for Solving the Combined Distribution and Assignment Problem with Link Capacity Constraints

In this paper, we consider the combined distribution and assignment (CDA) problem with link capacity constraints modeled as a hierarchical logit choice problem based on random utility theory. The destination and route choices are calculated based on the multi-nominal logit probability function, which forms the basis for constructing the side constrained CDA (SC-CDA) problem as an equivalent mathematical programming (MP) formulation. A dual MP formulation of the SC-CDA problem is developed as a solution algorithm, which consists of an iterative balancing scheme and a column generation scheme, for solving the SC-CDA problem. Due to the entropy-type objective function, the dual formulation has a simple nonlinear constrained optimization structure, where the feasible set only consists of nonnegative orthants. The iterative balancing scheme explicitly makes use of the optimality conditions of the dual formulation to analytically adjust the dual variables and update the primal variables, while a column generation scheme is used to iteratively generate routes to the working route set as needed to satisfy the side constraints. Two numerical experiments are conducted to demonstrate the features of the SC-CDA model and the computational performance of the solution algorithm. The results reveal that imposing link capacity constraints can have a significant impact on the network equilibrium flow allocations, and the dual approach is a practical solution algorithm for solving the complex SC-CDA problem.     

Trajectory optimization for maneuvering satellites

This paper proposes an algorithm to compute optimal trajectories for a maneuvering satellite by using a nonlinear programming representation of an optimal control problem. In this problem, a satellite must be located at a given final position with given velocity from initial position and velocity without passing a prohibited region such as the atmosphere while achieving minimum fuel consumption. Optimal control theory is applied to obtain a set of ordinary differential equations subject to two-point boundary conditions (TPB) on the adjoint system. Then an exact penalty function method is employed to obtain the optimal trajectories by solving the TPB problem as initial conditions for the adjoint system and an unknown final time are regarded as decision variables. This formulation, where the optimal control technique and the nonlinear programming method are incorporated, permits more systematic and flexible algorithm implementation.     

Design sensitivity analysis and optimization of dynamic response

The paper presents methods of design sensitivity analysis and optimization of dynamic response of mechanical and structural systems. A key feature of the paper is the development of procedures to handle point-wise state variable constraints. Difficulties with a previous treatment where such constraints were transformed to equivalent integral constraints are noted and explained from theoretical as well as engineering standpoints. An alternate treatment of such constraints is proposed, developed and evaluated. In this treatment each point-wise state variable constraint is replaced by several constraints that are imposed at all the local max-points for the original constraint function. The differential equations of motion are formulated in the first-order form so as to handle more general problems. The direct differentiation and adjoint variable methods of design sensitivity analysis to deal with the point-wise constraints are presented. With the adjoint variable methods, there are two ways of calculating design sensitivity coefficients. The first approach uses an impulse load and the second approach uses a step load for the corresponding adjoint equation. Since the adjoint variable methods are better for a large class of problems, an efficient computational algorithm with these methods is presented in detail. Optimum results for several problems are obtained and compared with those available in the literature. The new formulation works extremely well as precise optimum designs are obtained.     

Nonstandard duality and nonlocal necessary optimality conditions in nonconvex optimal control problems

To partially implement the idea of considering nonlinear optimal control problems immediately on the set of Pontryagin extremals (or on quasiextremals if the optimal solution does not exist), we introduce auxiliary functions of canonical variables, which we call bipositional, and the corresponding modified Lagrangian for the problem. The Lagrangian is subject to minimization on the trajectories of the canonical system from the Maximum Principle. This general approach is further specialized for nonconvex problems that are linear in state, leading to a nonstandard dual optimal control problem on the trajectories of the adjoint system. Applying the feedback minimum principle to both original and dual problems, we have obtained a pair of necessary optimality conditions that significantly strengthen the Maximum Principle and admit a constructive realization in the form of an iterative problem solving procedure. The general approach, optimality features, and the iterative solution procedure are illustrated by a series of examples.     

Design sensitivity analysis for transient response of non-viscously damped dynamic systems

A design sensitivity analysis for the transient response of the non-viscously damped dynamic systems is presented. The non-viscously (viscoelastically) damped system is widely used in structural vibration control. The damping forces in the system depend on the past history of motion via convolution integrals. The non-viscos damping is modeled by the generalized Maxwell model. The transient response is calculated with the implicit Newmark time integration scheme. The design sensitivity analysis method of the history dependent system is developed using the adjoint variable method. The discretize-then-differentiate approach is adopted for deriving discrete adjoint equations. The accuracy and the consistency of the proposed method are demonstrated through a single dof system. The proposed method is also applied to a multi-dof system. The validity and accuracy of the sensitivities from the proposed method are confirmed by finite difference results.     

Construction of the minimal time control for processes with bounded control amplitudes and rates

Minimal time control of linear time-varying processes with bounded control amplitudes and rates is discussed. Based on the maximum principle, necessary conditions are given which include a specific modified adjoint solution. Two additional conditions that are associated with this modified adjoint solution are derived. These conditions are then utilized to develop a systematic method of constructing the extremal control as an explicit time function. The construction scheme consists of a set of decision rules in the form of inequalities. These decision rules are implemented on a digital computer. The computer program basically checks out the decision rules and determines the appropriate set of prestored parameters, which describes each extremal control. To obtain the minimal time control from the extremal controls, a point-by-point search from the grid on a hypercylindrical surface is required. As an illustration, the computer program is applied to the Ranger Block III attitude control system.     

A dual Newton strategy for scenario decomposition in robust multistage MPC

This paper considers the solution of tree‐structured quadratic programs as they may arise in multistage model predictive control. In this context, sampling the uncertainty on prescribed decision points gives rise to different scenarios that are linked to each other via the so‐called nonanticipativity constraints. Previous work suggests to dualize these constraints and apply Newton's method on the dual problem to achieve a parallelizable scheme. However, it has been observed that the globalization strategy in such an approach can be expensive. To alleviate this problem, we propose to dualize both the nonanticipativity constraints and the dynamics to obtain a computationally cheap globalization. The dual Newton system is then reformulated into small highly structured linear systems that can be solved in parallel to a large extent. The algorithm is complemented by an open‐source software implementation that targets embedded optimal control applications.     

Periodicity in optimal hierarchical checkpointing schemes for adjoint computations

We reexamine the work of Aupy et al. on optimal algorithms for hierarchical adjoint computations, where two levels of memories are available. The previous optimal algorithm had a quadratic execution time. Here, with structural arguments, namely periodicity, on the optimal solution, we provide an optimal algorithm in constant time and space, with appropriate pre-processing. We also provide an asymptotically optimal algorithm for the online problem, when the adjoint chain size is not known before-hand. Again, these algorithms rely on the proof that the optimal solution for hierarchical adjoint computations is weakly periodic. We conjecture a closed-form formula for the period. Finally, we assess the convergence speed of the approximation ratio for the online problem through simulations.     

基于观测器的受扰非线性系统近似最优跟踪控制

研究一类受扰非线性系统的最优输出跟踪控制问题.给出了有限时域最优输出跟踪控制律的近似设计算法.首先将求解受扰非线性系统最优跟踪控制问题转换为求解状态向量与伴随向量耦合的非线性两点边值问题,然后利用逐次逼近方法构造序列将其转化为求解两个解耦的线性微分方程序列问题.通过迭代求解伴随向量的序列,可得到由解析的线性前馈-反馈控制部分和伴随向量的极限形式的非线性补偿部分组成的最优输出跟踪控制律.利用参考输入降维观测器和扰动降维观测器,解决了前馈控制的物理可实现问题.最后仿真结果表明了该方法的有效性.     

Approximate causal output tracking for linear perturbed systems via sliding mode control

ABSTRACT

A new approach to the solution of what is termed causal output tracking problem for linear time-invariant systems in the presence of both matched and unmatched unmodelled disturbances is presented. The proposed solution is addressed through the design of a dynamic steady state estimator based on the desired system structure and an observer, considering the reference as the system output. The unmatched disturbance is estimated and used in the steady state estimator to achieve invariance, while the matched disturbance robustness is provided via sliding mode control using the super-twisting algorithm. A useful application of the presented techniques is output tracking for non-minimum phase systems; thus, the performed simulation results confirm the effectiveness of the proposed control scheme for this kind of systems.     

自适应事件触发通信机制下机理解析与数据驱动融合的ICPS双重安全控制

针对存在拒绝服务(DoS)攻击与执行器故障的工业信息物理融合系统(ICPS),将机理解析与数据驱动方法相结合,在新型自适应事件触发通信机制下,研究双重安全控制问题.首先,设计自适应事件触发机制,能够触发参数随系统行为动态自适应变化,节约更多网络通信资源;其次,基于系统最大允许时延建立攻击检测机制,可以有效区分大、小能量DoS攻击;再次,基于极限学习机算法(ELM)建立时序预测模型,用于大能量DoS攻击时重构修正控制量,以主动容侵攻击的影响,并给出与小能量攻击时机理解析的弹性被动容侵来提升系统对攻击的防御能力;然后,借助T-S模糊理论、时滞系统理论、新型Bessel-Legendre不等式等,推证得到系统鲁棒观测器及双重安全控制器的解析求解方法,使双重安全控制与通讯性能得到折衷协同提升;最后,通过实例仿真验证所提出方法的有效性.     

非线性离散系统的伴随辨识法

对非线性离散系统辨识的研究,通常是将其转化成一个非线性优化问题.为此需要计算目标函数对参数向量的梯度,以往的方法需要求解一个矩阵差分方程,计算量颇大.依据系统输出量测值来确定含在系统中的未知参数向量,首先引进伴随状态向量,代替矩阵差分方程求解的是计算一个向量差分方程,从而大大简化计算,然后将这种梯度计算方法结合进拟牛顿信赖域法中.最后给出了应用此方法的一个实际例子,数值仿真的结果说明方法是有效的.     

非结构网格的并行多重网格解算器

多重网格方法作为非结构网格的高效解算器,其串行与并行实现在时空上都具有优良特性.以控制方程离散过程为切入点,说明非结构网格在并行数值模拟的流程,指出多重网格方法主要用于求解时间推进格式产生的大规模代数系统方程,简述了算法实现的基本结构,分析了其高效性原理;其次,综述性地概括了几何多重网格与代数多种网格研究动态,并对其并行化的热点问题进行重点论述.同时,针对非结构网格的实际应用,总结了多重网格解算器采用的光滑算子;随后列举了非结构网格应用的部分开源项目软件,并简要说明了其应用功能;最后,指出并行多重网格解算器在非结构网格应用中的若干关键问题和未来的研究方向.     

Arbitrary High-Order Explicit Hybridizable Discontinuous Galerkin Methods for the Acoustic Wave Equation

We propose a new formulation of explicit time integration for the hybridizable discontinuous Galerkin (HDG) method in the context of the acoustic wave equation based on the arbitrary derivative approach. The method is of arbitrary high order in space and time without restrictions such as the Butcher barrier for Runge–Kutta methods. To maintain the superconvergence property characteristic for HDG spatial discretizations, a special reconstruction step is developed, which is complemented by an adjoint consistency analysis. For a given time step size, this new method is twice as fast compared to a low-storage Runge–Kutta scheme of order four with five stages at polynomial degrees between two and four. Several numerical examples are performed to demonstrate the convergence properties, reveal dispersion and dissipation errors, and show solution behavior in the presence of material discontinuities. Also, we present the combination of local time stepping with h-adaptivity on three-dimensional meshes with curved elements.     

Minimax filtering in linear stochastic uncertain discrete-continuous systems

Filtering of the states of a system, whose dynamics is defined by an Ito stochastic differential equation, by discrete and discrete-continuous observations is studied under the assumption that the intensities of continuous noises and covariance matrices of discrete noises are known only within to membership of certain uncertainty sets. A minimax approach is used to solve the problem. The filter is optimized with an integral quality criterion. Minimax filtering equations are derived from the solution of the dual optimization problem. A numerical solution algorithm for the problem is designed. Results of numerical experiments are presented.     

On the numerical implementation of continuous adjoint sensitivity for transient heat conduction problems using an isogeometric approach

A crucial problem of continuous adjoint shape sensitivity analysis is the numerical implementation of its lengthy formulations. In this paper, the numerical implementation of continuous adjoint shape sensitivity analysis is presented for transient heat conduction problems using isogeometric analysis, which can serve as a tutorial guide for beginners. Using the adjoint boundary and loading conditions derived from the design objective and the primary state variable fields, the numerical analysis procedure of the adjoint problem, which is solved backward in time, is demonstrated. Following that, the numerical integration algorithm of the shape sensitivity using a boundary approach is provided. Adjoint shape sensitivity is studied with detailed explanations for two transient heat conduction problems to illustrate the numerical implementation aspects of the continuous adjoint method. These two problems can be used as benchmark problems for future studies.