Multi Agent Bayesian Optimization

Multi-agent Bayesian optimization (MABO) tackles the challenge of optimizing a shared objective function using multiple agents, each with limited information and potentially coupled constraints. Current research focuses on developing efficient algorithms, such as those based on primal-dual methods or the alternating direction method of multipliers (ADMM), to coordinate agents and handle both black-box and affine constraints while minimizing computational cost. These methods aim to improve sample efficiency and scalability compared to single-agent approaches, enabling applications in diverse fields like resource allocation and robotics where distributed optimization is crucial. The development of computationally tractable approximations, like Gaussian Max-value Entropy Search, is key to addressing the computational complexity inherent in multi-agent scenarios.