Linear Quadratic

Linear-quadratic (LQ) methods address optimal control problems and games where dynamics are linear and objective functions are quadratic, aiming to find optimal strategies or Nash equilibria. Current research focuses on developing efficient algorithms for solving these problems, particularly in multi-agent settings with incomplete information or stochasticity, employing techniques like policy optimization, mirror descent, and augmented Lagrangian methods. These advancements are significant for applications in areas such as multi-robot coordination, shared control systems, and reinforcement learning, offering improved scalability and convergence guarantees for complex decision-making scenarios.