Learning to Value Constraints. Optimisation subject to constraints is key to improving efficiency in transport, energy and many other areas. This project will develop better optimisation algorithms by

Description

Learning to Value Constraints. Optimisation subject to constraints is key to improving efficiency in transport, energy and many other areas. This project will develop better optimisation algorithms by leveraging the power of machine learning to boost the handling of constraints. By developing more advanced constraint handling, the optimisation methods created in this project will enable larger and more complex optimisation models to be solved. A particular focus is optimisation in applications involving networks. The development of such machine-learning enhanced optimisation approaches is expected to lead to benefits in industries where optimisation plays an important role, including transport, logistics, and energy grid planning.. Scheme: Discovery Projects. Field: 4602 - Artificial Intelligence. Lead: Prof Andreas Ernst