Learning to Coordinate: Distributed Meta-Trajectory Optimization Via Differentiable ADMM-DDP

2025-09-01

AI Summary

To address the strong coupling among hyperparameters and high tuning cost in ADMM-DDP for multi-agent trajectory optimization, this paper proposes L2C, a distributed meta-learning framework. Methodologically, L2C models ADMM-DDP hyperparameters—including truncation strategies and penalty coefficients—as lightweight differentiable neural networks, and leverages ADMM’s convex structure to enable distributed meta-gradient computation. Integrating differentiable programming, Riccati recursion, feedback gain reuse, and matrix-based LQR solving, it establishes an end-to-end differentiable distributed optimization pipeline. Evaluated in IsaacSIM, L2C enables rapid drone formation reconfiguration and coordinated 6-DoF payload manipulation, demonstrating strong generalization across varying team sizes and task specifications. Empirically, L2C accelerates meta-training gradient computation by up to 88% compared to baseline approaches.

Technology Category

Search and Optimization: Learning to SearchMultiagent Systems: Adversarial AgentsIntelligent Robots: Learning & Optimization for ROB

Application Category

Economics, Online Markets and Human Computation: Cost models of using LLMs in production systemsGraph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphsSemantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Distributed trajectory optimization via ADMM-DDP is a powerful approach for coordinating multi-agent systems, but it requires extensive tuning of tightly coupled hyperparameters that jointly govern local task performance and global coordination. In this paper, we propose Learning to Coordinate (L2C), a general framework that meta-learns these hyperparameters, modeled by lightweight agent-wise neural networks, to adapt across diverse tasks and agent configurations. L2C differentiates end-to-end through the ADMM-DDP pipeline in a distributed manner. It also enables efficient meta-gradient computation by reusing DDP components such as Riccati recursions and feedback gains. These gradients correspond to the optimal solutions of distributed matrix-valued LQR problems, coordinated across agents via an auxiliary ADMM framework that becomes convex under mild assumptions. Training is further accelerated by truncating iterations and meta-learning ADMM penalty parameters optimized for rapid residual reduction, with provable Lipschitz-bounded gradient errors. On a challenging cooperative aerial transport task, L2C generates dynamically feasible trajectories in high-fidelity simulation using IsaacSIM, reconfigures quadrotor formations for safe 6-DoF load manipulation in tight spaces, and adapts robustly to varying team sizes and task conditions, while achieving up to $88%$ faster gradient computation than state-of-the-art methods.

Problem

• Meta-learns hyperparameters for distributed multi-agent trajectory optimization
• Adapts coordination parameters across diverse tasks and agent configurations
• Enables efficient end-to-end differentiation through ADMM-DDP pipeline

Innovation

• Meta-learns hyperparameters via lightweight neural networks
• Differentiates end-to-end through ADMM-DDP pipeline
• Accelerates training with truncated iterations and meta-learned penalties