SOLID Lab focuses on the design and implementation of large-scale local and global optimization algorithms to solve problems that arise in diverse decision-making paradigms such as machine learning, stochastic optimization,and optimal control. Our algorithms combine mathematical techniques and emerging high-performance computing hardware to achieve computational scalability. We make these developments accessible to academic and industrial users by implementing algorithms on easy-to-use and extensible software libraries.

The problems that we are addressing are of unprecedented complexity and defy the state-of-the-art. For example, in our recent work, we developed a novel global optimization algorithm capable of solving k-center clustering problems (an unsupervised learning task) with up to 1 billion samples, while state-of-the-art approaches in the literature can only address several thousand samples.

We are currently using our tools to address engineering and scientific questions that arise in diverse application domains, including optimal decision trees, optimal clustering, deep-learning-based control, optimal power system planning, AI for bioprocess operation, and optimal design of zero energy buildings.