Abstract:The optimization of multibody systems requires accurate and efficient methods for sensitivity analysis. The adjoint method is probably the most efficient way to analyze sensitivities, especially for optimization problems with numerous optimization variables. This paper discusses sensitivity analysis for dynamic systems in gradient-based optimization problems. A discrete adjoint gradient approach is presented to compute sensitivities of equality and inequality constraints in dynamic simulations. The constraints are combined with the dynamic system equations, and the sensitivities are computed straightforwardly by solving discrete adjoint algebraic equations. The computation of these discrete adjoint gradients can be easily adapted to deal with different time integrators. This paper demonstrates discrete adjoint gradients for two different time-integration schemes and highlights efficiency and easy applicability. The proposed approach is particularly suitable for problems involving large-scale models or high-dimensional optimization spaces, where the computational effort of computing gradients by finite differences can be enormous. Three examples are investigated to validate the proposed discrete adjoint gradient approach. The sensitivity analysis of an academic example discusses the role of discrete adjoint variables. The energy optimal control problem of a nonlinear spring pendulum is analyzed to discuss the efficiency of the proposed approach. In addition, a flexible multibody system is investigated in a combined optimal control and design optimization problem. The combined optimization provides the best possible mechanical structure regarding an optimal control problem within one optimization.

Whiplash Gradient Descent Dynamics

Inertial Methods with Viscous and Hessian driven Damping for Non-Convex Optimization

An SDE Perspective on Stochastic Inertial Gradient Dynamics with Time-Dependent Viscosity and Geometric Damping

Optimal Convergence Rates for Damped Inertial Gradient Dynamics with Flat Geometries

Proximal Gradient Dynamics: Monotonicity, Exponential Convergence, and Applications

Accelerated Gradient Dynamics on Riemannian Manifolds: Faster Rate and Trajectory Convergence

Accelerated Gradient Methods Combining Tikhonov Regularization with Geometric Damping Driven by the Hessian

Convergence Analysis of a Class of Nonsmooth Gradient Systems.

On dissipative symplectic integration with applications to gradient-based optimization

Dissipative Gradient Descent Ascent Method: A Control Theory Inspired Algorithm for Min-max Optimization

Conformal Symplectic and Relativistic Optimization

On the stability of gradient descent with second order dynamics for time-varying cost functions

A Speed Restart Scheme for a Dynamical System with Hessian-Driven Damping and Three Constant Coefficients

Fast convex optimization via closed-loop time scaling of gradient dynamics

Accelerated gradient methods for nonconvex optimization: Escape trajectories from strict saddle points and convergence to local minima

Stochastic Inertial Dynamics Via Time Scaling and Averaging

Gradient Methods with Dynamic Inexact Oracles

Fast Convergence of Inertial Multiobjective Gradient-like Systems with Asymptotic Vanishing Damping

Convergence of Constant Step Stochastic Gradient Descent for Non-Smooth Non-Convex Functions

A discrete adjoint gradient approach for equality and inequality constraints in dynamics

Optimization on manifolds: A symplectic approach