Study On Fast Intelligent Optimization And Multi-Objective Scheduling Algorithms

The optimization of complex systems often encounters some difficulties, such as large-scale, hard to model, time-consuming to evaluate, NP-hard, multi-modal, uncertain and multi-objective, etc. It is always a hot research topic in academic and engineering fields to propose advanced theory and effective algorithms. With complex control and manufacturing systems as backgrounds, in this thesis a class of fast intelligent algorithm is proposed to overcome the time-consuming evaluation and several effective hybrid intelligent algorithms are proposed to solve multi-objective and multi-modal scheduling problems.The main contents of this thesis are as follows.Firstly, simulation optimization, intelligent optimization and multi-objective scheduling algorithms are overviewed.Secondly, a class of fast genetic algorithm (GA) based on interpolated evaluation is proposed to overcome the time-consuming evaluation in optimization process and to improve the efficiency. Simulation results based on parameter estimation and model reduction of nonlinear systems demonstrate the effectiveness and efficiency of the proposed algorithm.Thirdly, by combining the quantum-based GA and permutation-based GA,and employing the non-dominated sorting technique to handle multiple objectives, an effective hybrid quantum-inspired GA is proposed for the multi-objective permutation flow-shop scheduling problems.Fourthly, by combining the parallel evolutionary search based on particle swarm optimizer (PSO) and several local searches based on neighborhood structures, and employing random weights to handle multiple objectives, an effective hybrid PSO is proposed for the multi-objective permutation flow-shop scheduling problems.Simulation results and comparisons based on a set of benchmarks showthat, the non-dominated solutions obtained by the proposed hybrid algorithms can well approximate the optimal Pareto front and are of good diversity. Moreover, the proposed algorithms are of high efficiency and good robustness on problems. It should be stressed that we are the first to propose the two hybrid algorithms based on quantum computing and particle swarm optimizer in international academic field for multi-objective scheduling problems.Finally, summarization and future work are provided.