Study On The Rescheduling Method Of Flexible Job Shop Under The Arrival Of The New Job

The random arrival of a new job is often happened in the actual process of interference,and will affect the efficiency and stability of production scheduling.How to quickly solve the consequences of interference,reasonable selection of rescheduling mode is the most important thing.However,most of the scheduling decisions of current workshops in the face of disturbances in the production process only rely on the experience of workers,and the decisions have great uncertainty and randomness.Even if there are related disturbance studies are mostly focused on flow shop and job shop.Therefore,based on flexible job shop production under the disturbance of random arrival of new job,this paper studies the method of how to quickly select the rescheduling mode which is most meets the requirements.The content is as follows:（1）Firstly,the Dynamic Flexible Job Shop Scheduling Problem（DFJSP）is deeply studied,and the rescheduling strategy of the Flexible Job Shop under the arrival disturbance of new jobs is mainly studied.Two rescheduling strategies are designed to reduce the impact caused by interference,and the optimal solution of the initial scheduling model is found by combining with the improved genetic algorithm.Taking minimizing C_maxas the optimization objective of initial scheduling and a rescheduling method is selected to optimize the comprehensive performance and stable performance.Several experiments are conducted to analyze the influence of the new jobs on the results when it arrived at different stages,and the weighted method is used to determine the selected rescheduling strategy.A large number of high-dimensional labeled data are obtained in the experiments.（2）In order to quickly and efficiently select the most suitable rescheduling mode from the two rescheduling strategies when the random arrival of the new job is disturbed,PNN learns from large amounts labeled data and obtains the FJSP rescheduling mode.For improving the classification accuracy and efficiency of PNN,PCA is used to reduce the dimension of high-dimensional features.For improving the smoothing factor selection method based on manual experience in PNN network,a set of smoothing factor vectors is set,and the Sparrow Search Algorithm improved by adaptive t-distribution strategy is used to optimize the data.SSA has better global exploration and local development capabilities,and the introduction of an adaptive t-distribution strategy can make the sparrow quickly converge near the optimal value.（3）Finally,through the analysis of the actual production process of a wild silk product company,the method proposed in this paper is applied to the wild silk manufacturing workshop where new job frequently arrive,and a visual workshop scheduling platform is established.The system can meet the needs of daily selection of optimal rescheduling mode,effectively improve the resource utilization rate and information management level,reduce manual error and facilitate the use of operators.