Optimizing Through Change: Bounds and Recommendations for Time-Varying Bayesian Optimization Algorithms