Research On The Methods For Distributed Blocking Flowshop Scheduling

In the context of global economic integration,with the optimization and integration of general-purpose factories and the cooperation of homogeneous production lines,distributed manufacturing has gradually become a common manufacturing mode.Distributed manufacturing enables the group to rationally configure,optimize and share the resources of the subordinate enterprises or factories,and coordinate multiple factories to establish cooperative relationship.In the environment of low-cost and low-risk,the strategies of labor division and collaboration are adopted to realize the efficient manufacturing.As an important part of production activities,job shop scheduling plays an important role in distributed manufacturing system.Different from traditional single-shop scheduling problems,scheduling in a distributed environment not only needs to consider the correlation between task characteristics and production constraints,but also needs to consider the coupling between global scheduling and local scheduling of each distributed factory,and the solution process is more complex.Hence,it is of great significance to build scheduling optimization model and design effective optimization method according to the characteristics of distributed manufacturing.With view of promoting the scheduling theory research of distributed manufacturing.The distributed job shop scheduling problem sometimes optimizes a single production target,sometimes it needs to take the optimization of modern enterprises and national macroeconomic development goals(such as energy conservation and emission reduction and other environmental protection policies)into account.Sometimes the decision-making space is limited by a single constraint,sometimes multiple constraints need to be met at the same time.Sometimes the scheduling environment is static,sometimes it is disturbed by uncertain factors.Based on the research background of distributed blocking flowshop manufacturing,this research has designed the corresponding distributed production scheduling architecture,and makes an in-depth research on static scheduling,green scheduling and dynamic scheduling of distributed blocking flowshop.The mathematical models of static scheduling,green scheduling and dynamic scheduling were constructed respectively,and three kinds of metaheuristic optimization algorithms were proposed based on different problem characteristics.Finally,a prototype system for distributed blocking flowshop scheduling was built,and the system implementation and application verification of the theoretical research were carried out in a technologyintensive enterprise.The main research work is introduced as follows:(1)The characteristics of general blocking flowshop scheduling problem are analyzed,the concept of distributed manufacturing is introduced,and the business process of distributed production scheduling is analyzed.On this basis,the key decision points of the distributed blocking flowshop scheduling business are further studied,the production scheduling architecture for distributed flow manufacturing is designed.A production scheduling management and control mode based on cloud-edge combination is proposed.(2)The distributed blocking flowshop scheduling problem in the static environment is studied.Based on the analysis of the characteristics of distributed flow manufacturing mode,a mixed integer programming model of the scheduling problem is constructed with production efficiency as the optimization goal,and a discrete fruit fly optimization algorithm(DFOA)is proposed.First of all,according to the characteristics of blocking flowshop constraints,three kinds of population initialization methods are proposed;in the smell search phase,four kinds of directed neighborhood search operators are designed to expand the search space of DFOA;then,a variable neighborhood descent search framework including two types of plug-in local search is proposed;in the visual search phase,the population update strategy is improved to ensure the fast convergence of DFOA.Finally,the effectiveness and superiority of DFOA are verified on the test benchmark.(3)The green multi-objective scheduling problem of distributed blocking flowshop which considers both production efficiency and energy saving is studied.The conflict relationship between manufacturing cycle and energy consumption is analyzed,and a multi-objective distribution estimation algorithm based on Pareto theory(MOEDA)is proposed.Firstly,a discrete probability model based on Bayesian network is established,and the individual population is updated by the probability model.Then,two kinds of speed adjustment operators are designed,and a kind of job insertion strategy is improved,which realizes the adaptive adjustment of the process speed during the procedure of job sequencing;Finally,the coverage and approximation performance of MOEDA to the front of Pareto are verified by comparative experiments.(4)Considering the event-driven distributed manufacturing mode and taking the machine breakdown as the background,the dynamic scheduling problem of distributed blocking flowshop is studied.The mechanism of breakdown simulation and repair is designed,and a dynamic scheduling strategy for distributed blocking flowshop is proposed.Taking makespan and system stability as optimization objectives,a distributed blocking flowshop interference management model is constructed,and a discrete Memetic algorithm based on differential evolution(DMA)is designed for rescheduling jobs.Firstly,a population initialization strategy based on job weight location is designed;then,the idea of differential evolution is introduced to search the rescheduled target space;then,a random reference local search strategy based on job block is proposed to prevent the algorithm from falling into local optimum.Finally,the effectiveness and superiority of the rescheduling algorithm are verified by comparative experiments.(5)Combined with the distributed blocking flowshop production architecture,a cloud-edge-based scheduling prototype system is developed,which realizes the cloud operation of the group scheduling business.Then,the multi-variety,small-scale and customized products are taken as the application object in a technology-intensive enterprise,the dynamic scheduling strategy and DMA rescheduling algorithm are verified by examples.This paper has 77 figures,24 tables and 186 references.