This paper focuses on deterministic single machine scheduling with nonzero ready times. The objective is to sequence the jobs such that both maximum tardiness (T_(max)) and number of tardy jobs (n_T) are minimized. However, no preference is established between n_T and T_(max) and we aim to find non-dominated solutions. The proposed approach has two main elements: 1) Individual Population Evolvers (IPEs) and 2) Population Mixer (PM). IPEs focus on optimizing the designated performance measure while PM mixes solutions to improve both measures. The experimentation performed shows that mixing both populations at the MIXER improve the set of non-dominated solutions when the total number of solutions kept constant. The MIXER approach results were compared with the Optimal Pareto Frontier in 10-job category and finally the proposed approach was compared with another multi-objective Genetic Algorithm Approach. The results show that the MIXER approach performed very well in both comparisons.
展开▼
