Leveraging human strategy in Bayesian optimization for industrial commissioning

Commissioning industrial processes is often tedious and difficult because certain settings, like controller parameters, need repeated adjustment to reach a specific goal. Passing knowledge on system tuning from experienced operators to junior colleagues is also challenging, as this knowledge is often tacit and hard to formalize.We approach the tuning problem with Bayesian optimization (BO). We discuss different ways of combining the best of both worlds, i.e., the insights of the expert with the goal-driven exploration of BO. With the goal of incorporating recorded expert tuning sessions into BO, we delve into two main directions.First, we explore the effect of context changes on a Human First-Computer Last approach, where the expert conducts an initial rough tuning, and BO performs a fine-tuning afterward. Second, we outline initial steps toward a subgoal-based BO method, based on the observation that humans often divide complex tasks into multiple subgoals achieved in sequence.We evaluate these methods on two simulation setups: a cascade controller of a pusher-slider robotic system and a projectile trajectory planning problem.