Survey
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project
Evolutionary Algorithms and Scheduling Colin Reeves School of Mathematical and Information Sciences Coventry University Evolutionary Algorithms • • • • Population-based heuristic search Genotype-phenotype mapping Fitness evaluation ‘Genetic’ operators: recombination, mutation • Selection mechanism Scheduling • Single/multiple machines • Flow shop/job shop/open shop Traditional Jobshop Variations • • • • • • Simple flows/Precedence constraints Serial/parallel machines Intermediate storage Ready times Independent/dependent processing times Due dates/earliness penalties EA Techniques • • • • • Encoding methods: direct/indirect Sequence/schedule (permutations) Operator design Performance measures Despatching rules Fitness landscapes • • • • Different operators = different landscapes Measuring ‘ruggedness’ ‘Big valley’ phenomenon ‘Path tracing’ EAs EA Strengths/Weaknesses • Population: good for uncertainty/multiobjective aspects • Parallel implementation • Hybridisation • May be inefficient (fitness evaluations) • Need for tuning • Not an excuse for lack of thought Future • Uncertainty modelling: probability vs. fuzziness • Multiple objectives • Rescheduling: robustness measures (Artificial Immune Systems?)