|'DS for O'. Many decision support problems require the usage of advanced optimisation techniques in order to make good, or the best, decisions. Examples, arising in OR, CS, and AI, include direct practical resource allocation problems such as scheduling, timetabling, networks, facility location, transport and routing, and many others, However, the complexity of such techniques has increased to the point where their configuration and control can in itself become a significant challenge; these 'optimisation of the optimiser' tasks are already being addressed by machine learning and statistical techniques and so are good ground for applications of DS. This includes data driven model description, algorithm configuration and construction as well as active learning of best decision making strategies.
'O for DS'. Many problems within DS are themselves naturally framed as optimisation problems. Examples include feature selection, data classification, clustering, data mining, big data problems, expectation maximisation, error minimisation, etc. The group aims to support the exploration of a wide range of state-of-the-art optimisation techniques to these problems.
Overall, applications and theory of data science and optimisation require inherently different skills than those in developing search methods for specific domains. This group aims to bring together the relevant groups of people.