K.U.Leuven - Departement toegepaste economische wetenschappen
DTEW Research Report 9938 pages:1-28
As of this writing, many success stories exist yet of powerful genetic algorithms (GAs) in the field of constraint optimisation. In this paper, a hybrid, intelligent genetic algorithm will be developed for solving a cutting layout problem in the Belgian fashion industry. In an initial section, an existing LP formulation of the cutting problem is briefly summarised and is used in further paragraphs as the core design of our GA. Through an initial attempt of rendering the algorithm as universal as possible, it was conceived a threefold genetic enhancement had to be carried out that reduces the size of the active solution space. The GA is therefore rebuilt using intelligent genetic operators, carrying out a local optimisation and applying a heuristic feasibility operator. Powerful computational results are achieved for a variety of problem cases that outperform any existing LP model yet developed.