Multi Objective Bayesian Optimisation

Multi-objective Bayesian optimization (MOBO) is a powerful technique for efficiently finding optimal solutions when dealing with multiple, often conflicting, objectives in computationally expensive problems. Current research focuses on improving the scalability and efficiency of MOBO algorithms, particularly through the development of parallel and batch acquisition functions, and exploring different surrogate model architectures, including both mono-surrogate (using a single model for aggregated objectives) and multi-surrogate (using separate models for each objective) approaches. These advancements are impacting diverse fields, from telecommunications network optimization and drug policy design to engineering applications like heat exchanger and wind farm layout optimization, by enabling the efficient exploration of complex design spaces with multiple performance criteria.