<ul><li>The reinforcement learning (RL) and model predictive control (MPC) communities have developed vast ecosystems of theoretical approaches and computational tools for solving optimal control problems.</li><li>MPCritic is a machine learning-friendly architecture that seamlessly interfaces with MPC tools.</li><li>MPCritic utilizes the loss landscape defined by a parameterized MPC problem, focusing on 'soft' optimization over batched training steps.</li><li>The versatility of MPCritic is demonstrated on classic control benchmarks in terms of MPC architectures and RL algorithms it can accommodate.</li></ul>

MPCritic: A plug-and-play MPC architecture for reinforcement learning

Discover more