Performance-oriented model learning for control via multi-objective Bayesian optimization

•The problem of performance-oriented model learning for model-based control is formulated as a black-box, multi-objective optimization problem.•The performance-oriented model learning problem is solved using multi-objective Bayesian optimization, which is particularly suitable for handling black-box, expensive and noisy performance measures in closed-loop control.•Performance-oriented model adaptation can yield control-relevant models that result in a significant improvement in closed-loop performance compared to model identification using closed-loop data.