Information-Theoretic Multi-Objective Bayesian Optimization with Continuous Approximations
This addresses the challenge of efficiently optimizing multiple objectives in resource-constrained real-world applications like rocket launching, though it appears incremental as an extension of Bayesian optimization to continuous approximations.
The paper tackles the problem of multi-objective black-box optimization with continuous function approximations that trade off accuracy and cost, proposing iMOCA to maximize information gain per unit cost for approximating the Pareto front. Experiments show iMOCA significantly improves over existing single-fidelity methods on synthetic and real-world benchmarks.
Many real-world applications involve black-box optimization of multiple objectives using continuous function approximations that trade-off accuracy and resource cost of evaluation. For example, in rocket launching research, we need to find designs that trade-off return-time and angular distance using continuous-fidelity simulators (e.g., varying tolerance parameter to trade-off simulation time and accuracy) for design evaluations. The goal is to approximate the optimal Pareto set by minimizing the cost for evaluations. In this paper, we propose a novel approach referred to as information-Theoretic Multi-Objective Bayesian Optimization with Continuous Approximations (iMOCA)} to solve this problem. The key idea is to select the sequence of input and function approximations for multiple objectives which maximize the information gain per unit cost for the optimal Pareto front. Our experiments on diverse synthetic and real-world benchmarks show that iMOCA significantly improves over existing single-fidelity methods.