Font Size: a A A

Multi-objective Lot-streaming Flow Shop Scheduling Based On Evolutionary Optimization

Posted on:2017-04-29Degree:DoctorType:Dissertation
Country:ChinaCandidate:Y Y HanFull Text:PDF
GTID:1108330509954782Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
The lot-streaming flow shop(LSFS) scheduling problem can be described that each job can be split into several sublots, and each sublot can be processed on the downstream machine when it is completed at the current one. In this way, it will reduce the production cycle and idle time, accelerate the manufacturing process, increase resource utilization rate, and enhance customer satisfaction. Thus, it has been widely employed in manufacturing systems, assembly lines, information service facilities, as well as chemical, textile, plastics, and semiconductor industries.In real-world applications, the LSFS scheduling problem is usually subject to the following cases:(1) optimizing a single objective or simultaneously optimizing multi ones;(2) existing unconstrained decision variables or ones with one or multi constraints;(3) considering determinate paprameters or indeterminate one(s). Previous research has demonstrated that it is difficult to obtain an optimal solution of the scheduling problem with constraints within polynomial time. Thus, it is highly challenging to solve the indeterminated multi-objective LSFS scheduling problem with constraints.Although various intelligent optimization methods have already been done to tackle the certain LSFS scheduling problem, they are not able to solve the uncertain multi-objective one with multiple constraints. In view of this, the corresponding mathematical models and the optimization algorithms of the LSFS scheduling problem are proposed based on the characteristics of multi-objective, constraints and uncertainties. The detailed contents are described as follows.First, a mathematical model of the multi-objective LSFS problem is constructed, and an improved NSGA-II(Improved Nondominated Sorting Genetic Algorithm II, INSGA-II) are proposed. In the proposed algorithm,(1) a variant of NEH heuristic is designed to generate several initial solutions, which remarkably enhances the quality of the initial population;(2) an estimation of distribution algorithm(EDA) and a mutation operator based on insertion and swap operators replace traditional crossover and mutation operators, which improves the search capability and convergence rate;(3) a simple and efficient restarting strategy is performed on the population when the diversity of the population is smaller than a given threshold so as to greatly enhance the diversity of the population.Following that, a blocking constraint is added to the above mathematical model to form a multi-objective blocking lot-streaming flow shop scheduling problem(BLSFS). Its mathematical model and a multi-objective artifical bee conlony algorithm are proposed. In the proposed algorithm,(1) two variants of NEH and MME(combination MinMax with NEH) heuristics are presented to generate several solutions with high quality;(2) two novel crossover operators are proposed by taking full of valuable information of non-dominated solutions so as to enhance the capability of the proposed algorithm in exploration;(3) an efficient Pareto local search operator is designed to improve the capability of the proposed algorithm in exploitation.Next, a multi-objective blocking lot-streaming flow shop scheduling problem with interval processing time is considered, in which(1) a multi-objective BLSFS scheduling problem with interval processing time is transformed into a traditional multi-objective optimization problem using the midpoint and radius of processing time;(2) an imporved differential evolution operator is adopted to strengthen the capability in exploration, and a local search with ideal point-based assistant method is applied to improve the capability of the algorithm in exploitation.Then, the multi-objective BLSFS scheduling problem with machine breakdowns is considered, in which(1) a crossover operator is proposed by taking full of valuable information of non-dominated scheduling strategy;(2) a rescheduling strategy is designed according to the characteristic of machine breakdowns.Last but not the least, a lot-streaming batch battery production of a solar energy company as the application object is focused on, and the practical LSFS scheduling problem is analyzed. In addtiton, the previous proposed optimization theory is applied into the practical production, and the simulation expriments demonstrate the feasibility and efficiency of the proposed algoirhtms.The above research results can not only enrich and deepen the existing scheduling theories but also promote the application of these theories in such fields as shop scheduling of batch battery, thereby directly serve the national economic and social development.
Keywords/Search Tags:Lot-streaming flow shop, Blocking, Multi-Objective Optimization, NSGA-II, Artifical bee conlony algorithm, Interval, Machine breakdowns
PDF Full Text Request
Related items