Research On Scenario-based Job Shop Scheduling Method In Uncertain Environment

As an important component of manufacturing,workshop carries a large number of production tasks and is also a key point for a large amount of real-time information.In the context of intelligent manufacturing,the market demand for products is constantly changing to individualization,and the production process of products has also become more diversified.This makes the actual production scheduling decision more complicated,leading to many uncertain factors in the production process of the manufacturing workshop.For example,machine failure,workpiece rework,material shortage,emergency order insertion,processing time change,etc.These uncertainties have seriously affected the stable operation of the production system.Therefore,under the current background,the manufacturing industry has put forward higher requirements for the calculation efficiency,actual operability,and real-time response ability to production disturbance factors for the workshop scheduling problem.With the continuous development of industrial Internet of Things and artificial intelligence technology,the level of intelligence in manufacturing workshops is constantly improving.A large amount of valuable data is stored in the manufacturing information system,which makes it possible to realize intelligent real-time decision-making in the production and processing process.Therefore,in the context of intelligent manufacturing,how to effectively use the relevant historical data of the manufacturing system and mine scheduling knowledge from it to guide the actual workshop production activities is a research hotspot in the industry and academia.This paper takes the discrete manufacturing dynamic job shop scheduling problem under the background of intelligent manufacturing as the research topic,and introduces the multi-scenario theory to the job shop scheduling research under uncertain environment.Aiming at the actual needs of discrete manufacturing workshops under uncertain environments,based on multi-scenario technology and online scheduling technology,with parametric modeling and data-driven as the orientation,the research idea of "technical theory + method research + simulation verification + system development" is adopted,The real-time decision-making method of job shop scheduling under uncertain environment is researched,and a prototype system is designed and developed.The main research contents of this paper are as follows:(1)Aiming at the problem of multi-scenario construction in the job shop under uncertain environments,the multi-scenario analysis method is first summarized;then the multi-scenario construction technology and scene reduction technology are introduced in detail,and the specific process of multi-scenario construction is given.;Finally,take the multi-scene generation of the job shop under the medium-scale production scale as an example to illustrate the scene construction,which provides the scene foundation and theoretical support for the subsequent chapters.(2)Aiming at the performance evaluation of job shop dynamic scheduling in an uncertain environment,this paper first designs an object-oriented dynamic scheduling simulation model;then based on equipment failure,emergency order insertion,random arrival of workpieces,and changes in the urgency of delivery time,etc.Uncertain factors construct job shop dynamic production scenarios;comprehensively consider 5types of scheduling indicators,perform online scheduling simulation analysis on 20 scheduling rules,and compare the scheduling performance of different strategies under different indicators;finally build job shop dynamics based on scheduling simulation results Scene knowledge base.In actual production applications,it guides workshop managers to make dynamic scheduling decisions,thereby improving the overall production efficiency of the enterprise.(3)Aiming at the high computational cost and insufficient robustness of traditional dynamic scheduling methods in uncertain environments,this paper introduces the integrated learning algorithm Random Forest(RF)into job shop dynamic scheduling research.First,a two-stage real-time scheduling framework is proposed;then,a data-driven real-time scheduling method(PCA-XGBoost-IRF)based on Improved Random Forest(IRF)is designed.The purpose is to take the status information of the production workshop as the input and the optimal scheduling strategy as the output to construct a mapping network from the system state to the optimal scheduling strategy in the production scheduling process.It is used to select the optimal scheduling rule for each machine in the job shop scheduling problem in an uncertain environment;finally,the validity and feasibility of the proposed method are verified based on case analysis.(4)In response to the actual needs of production scheduling in discrete manufacturing workshops,based on the research results of the above chapters,combined with Socket communication technology,a job shop dynamic multi-scenario scheduling system was developed based on the B/S architecture design.The purpose is to provide a low-cost,efficient and convenient simulation experiment environment for scientific researchers,and at the same time provide a decision-making basis for the actual production scheduling of enterprise workshop managers.This research also provides a reference for the research of data-driven intelligent scheduling system.