Séminaire : Higher, faster, stronger: some real-world applications (and solutions) of multi-objective optimisation, J. Fieldsend

The next seminar will be given on the October 24th, 2024, 2pm in B014 room, by Jonathan Fieldsend, Professor of Computational Intelligence at the University of Exeter, UK.

Title :
Some real-world applications (and solutions) of multi-objective optimisation

Abstract:
This talk will cover the application of multi-objective optimisation techniques to some of the practical problems he has been involved with during his academic career. Three industrial examples have been selected which span a wide range of problem characteristics, domains, and optimisation algorithm type used. The first applied problem, from air traffic control, is tuning a classification component of a short-term conflict alert system. The second, a sensor mesh routing problem in the heritage sector to maximise both network lifetime and time to first battery failure. The final problem is the stacked tray design in a wastewater separator — where the computational fluid dynamics evaluation takes multiple hours per design. The talk will step through the problem properties, the insights (and issues!) raised tackling real-world problems, and give some flavour of the range and applicability of algorithms employed.

The talk should be of interest to those interested in black-box and grey-box multi-objective optimisation, and the practicalities of working with, and addressing, industrial optimisation problems.

Journée du LISIC

La journée commencera par une présentation des salles hébergeant du matériel de recherche Les membres du laboratoire (permanents ou non-permanents) sont sollicités pour proposer un

Journée du LISIC 2024

Le laboratoire LISIC a le plaisir d’organiser la 4ème édition de la « Journée du LISIC » qui aura lieu le jeudi 11 juillet 2024

Campagne ATER 2024

Le laboratoire offre 5 postes d’ATER (contrat durée 1 an) pour la rentrée en septembre 2024 dont l’enseignement est effectué au sein des unités d’enseignement

Aller au contenu principal