Dual-Network PINNs for Optimal Control: A Reproducible Benchmark on the Mass-Spring-Damper System
arXiv:2606.15271v1 Announce Type: cross Abstract: This work presents a transparent and reproducible benchmark study of a direct dual-network Physics-Informed Neural Network (PINN) formulation for the optimal control of a mass-spring-damper system. The classical linear-quadratic optimal control problem is solved by two independent classical methods -- Pontryagin's Minimum Principle with single shooting, and direct transcription through trapezoidal collocation -- and recast as a constrained optimization problem solved by two feedforward neural networks: a state network whose boundary conditions
The paper demonstrates the growing maturity and application of Physics-Informed Neural Networks (PINNs) in solving complex control problems, driven by advances in AI and computational methods.
This work provides a reproducible benchmark for using AI, specifically PINNs, in optimal control, which is critical for autonomous systems and complex engineering challenges.
The ability to reliably apply dual-network PINNs for optimal control offers a new methodological tool for engineers and researchers, potentially accelerating development in domains requiring precise control.
- · AI/ML researchers and engineers
- · Robotics and autonomous systems developers
- · Aerospace and automotive industries
- · Control systems engineers
- · Traditional optimal control methods (potential efficiency loss if not integrated
- · Engineers without AI/ML expertise
Improved efficiency and performance in designing optimal controllers for dynamic systems using AI.
Accelerated development of more robust and adaptive autonomous vehicles and robotic systems.
Potential for AI-driven self-optimizing physical systems across various industries, leading to enhanced performance and resource efficiency at scale.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG