Tuesday 13 February 2024, 9:00am to Thursday 15 February 2024, 5:00pm
VenueLancaster University (Room TBC), Lancaster, UK
RegistrationFree to attend - registration required
This event is only for STOR-i students and STOR-i National Associates.
The notion that "nothing happens in the universe in which some rule of maximum or minimum does not appear" was known at least since Euler's time, and it gave rise to extraordinary theoretical results, which were further expanded when Kantorovič outlined its relevance even in the more earthly plywood industry. This, combined with the development of the first colossal programmable computer some 300 km from here, sparked interest in algorithms that could max / minimize constrained objectives and "allow businesses and traders to become more efficient and competitive".
Unfortunately, the problems faced by businesses and traders can rarely be solved to proven optimality by mathematical means in limited time. This has given rise to a line of research that does away with mathematics to achieve heuristic algorithms. The two communities, math for exact and no-math for heuristic - especially metaheuristic - algorithms, coexisted for decades with little overlap. However, the new millennium saw the rise of the field of matheuristics, i.e., heuristic algorithms that make use of the mathematical model of the problem of interest.
This masterclass will introduce the field, presenting both matheuristic extensions of well-known metaheuristics and innovative methods designed to obtain heuristic solutions based on mathematical programming results. Hands-on experience with simple matheuristic methods on simple instances will be proposed if time permits.