Provides a full Lean formalization of dual-regularized Riccati recursions and their inertia/descent results for interior-point optimal control.
Abstract
We derive closed-form extensions of the sequential and parallel Riccati recursions for solving dual-regularized linear-quadratic regulator (LQR) problems, with $O(N)$ sequential time and $O(\log(N))$ parallel time, respectively. We show that these subproblems arise when using regularized primal-dual interior-point methods to solve smooth, constrained, non-convex, discrete-time optimal control problems via multiple-shooting, even in the presence of stagewise equality or inequality constraints, and without imposing any rank requirements on constraint Jacobians. We prove that, when certain inertia conditions on the Newton-KKT matrix are met, each nonzero primal step is a descent direction of an augmented barrier-Lagrangian merit function. We characterize these inertia conditions in terms of the positive-definiteness of the dual-regularized Riccati pivots (a weaker condition than the standard LQR positive-definiteness requirements), thereby yielding inexpensive certificates of the required inertia. We provide MIT-licensed implementations of our methods in C++ and in JAX, as well as a full formalization of our results in Lean. We benchmark our algorithm against leading optimal control and nonlinear programming solvers on complex trajectory optimization problems, establishing competitive performance on moderate problems and substantial gains as the horizon length, problem dimension, and constraint count increase.
Problem
Regularized interior-point methods for constrained discrete-time optimal control need provably correct Riccati recursions and inertia certificates, without rank assumptions on constraint Jacobians.
Approach
The authors derive closed-form sequential (O(N)) and parallel (O(log N)) Riccati recursions for dual-regularized LQR subproblems arising in primal-dual interior-point methods via multiple-shooting, characterize the Newton-KKT inertia conditions in terms of positive-definite Riccati pivots, and give a full formalization of the results in Lean alongside C++/JAX implementations.
Results
When inertia conditions hold, each nonzero primal step descends an augmented barrier-Lagrangian merit function; benchmarks against acados, IPOPT, Fatrop and others show competitive performance on moderate problems and large gains as horizon, dimension, and constraint count grow.
Figure 1: Representative optimized trajectories from the analytical benchmark set. Left: bounded-actuation cartpole swing-up. Right: quadpendulum navigation with obstacle-avoidance constraints. Faded configurations show intermediate states along each optimized trajectory.