Research On Flexible Flow Shop Scheduling With Batch Process Machines Based On DES-GEP

With the development of the social economy,the market demand tends to product diversification and personalized customization.The modern manufacturing model has been evolving from the mass production to the flexible multi-varieties and small-batch production mode.The Flexible flow production can flexibly adapt to changes of market and implement mixed-line productions of multiple products.It is the mainstream way to deal with multivarieties and small-batch production for discrete manufacturing companies.As a typical application scenario of the production,The Flexible flow production with batch process machines is widely used in many fields such as semiconductor packaging.The scheduling process under this scenario not only complete equipment Assignment and job sequencing,but also consider how jobs will be grouped,so the Productive Operation Management is more complex.At the same time,dynamic events such as the dynamic arrival of jobs and fluctuations in working hours often occur in the actual Productive process,making scheduling decisions only based on real-time information.Therefore,The problem that enterprises need to solve urgently is how to propose an effective scheduling method based on real-time production data for the flexible flow scheduling problem with batch process machines to respond to the dynamic factors in the production process in a timely manner while improving production efficiency.The thesis introduces a scheduling rule method that is widely used in the field of dynamic scheduling.In view of the limitations of this method to solve our problem,the thesis is carried out from three aspects: model construction and algorithm design based on compound dispatching rules,the evaluation of the compound dispatching rules and dynamic generation of the compound dispatching rules.The specific work is as follows:(1)Model construction and algorithm design of flexible flow scheduling problem with batch process machines based on compound dispatching rules.On the basis of the flexible flow scheduling problem,parallel batch machines with incompatible job families and the mixed-line production of multiple products are considered.a multi-objective mathematical model is constructed based on the characteristics of this problem.A method for constructing compound dispatching rules based on sub-problems is proposed in view of the shortcomings of dispatching rules that are difficult to be solved directly because of the single dimension of the scheduling rules.What's more,considering the performance evaluation and the dynamic generation of compound dispatching rules,a scheduling framework based on DES-GEP is proposed.(2)Performance evaluation of compound dispatching rules based on DES.Firstly,Aiming at the problem that the evaluation of dispatching rules needs to be based on real-time working condition data and therefore the evaluation model is difficult to construct,the discrete-event simulation method is used for parametric modeling to construct a reusable scheduling simulation model;Secondly,The scheduling decision-making process of HFSP-BPM with time window is analyzed;Finally,the effectiveness of the time window is verified through calculation examples of different scales,and the conclusion that there is strong coupling between dispatching rules is drawn.In addition,the significance analysis of the parameters of the calculation example verifies the effectiveness of the setting method of the calculation example.(3)Automatic generation of compound scheduling rules based on improved GEP.In view of the shortcomings that the same scheduling rule cannot be applied to any scenario,an automatic generation algorithm of compound scheduling rules based on improved GEP is proposed.The classic dispatching rules are used as the elements in the set of endpoint,and genetic operations based on effective lengths are proposed.The non-dominant sorting method of NSGA-? is Introduced for multi-objective optimization problems.Aiming at the shortcomings of the algorithm,including fall into local optimum easily and prone to repeated individuals,improvement strategies including the repeated individual elimination algorithm,variable neighborhood search algorithm and adaptive genetic operator are proposed.Finally,the effectiveness of the improved GEP is verified through the analysis of numerical examples.