Task-optimal data-driven surrogate models for eNMPC via differentiable simulation and optimization

arXiv:2403.14425v1 Announce Type: new
Abstract: We present a method for end-to-end learning of Koopman surrogate models for optimal performance in control. In contrast to previous contributions that employ standard reinforcement learning (RL) algorithms, we use a training algorithm that exploits the potential differentiability of environments based on mechanistic simulation models. We evaluate the performance of our method by comparing it to that of other controller type and training algorithm combinations on a literature known eNMPC case study. Our method exhibits …

abstract algorithm algorithms arxiv contrast control cs.lg data data-driven differentiable environments exploits math.oc optimization performance reinforcement reinforcement learning simulation standard training type via