Output details
15 - General Engineering
University of Sheffield
Diversity Management in Evolutionary Many-Objective Optimization
Partners in industry value a rich diversity of solutions from which to select a suitable solution when applying our multiobjective evolutionary optimisation algorithms to their problems. Well-distributed solutions can yield important insights by revealing a wide range of trade-off options in applications such as engine calibration (Ford, contact: Dr R Lygoe, blygoe@ford.com) and system architecture design (Rolls-Royce, contact: Dr I Griffin, i.griffin@rolls-royce.com). Since industrial problems can often involve many objectives, it is especially difficult to evolve a satisfactory range and distribution of trade-offs. This paper significantly improves the quality of solutions arising in such problems.