Research On Shipyard Intelligent Production Scheduling Problem

There are still many problems in the production scheduling of shipyards in China.Production management is mainly on-site scheduling.The production plan is inconsistent with the actual situation,and there are problems of inefficiency and waste of resources.Considering the global development of intelligent manufacturing and the Chinese government’s promotion of"Made in China 2025",improving the intelligent production scheduling level of shipyards is the focus of shipbuilding transformation and upgrading.Studying the intelligent scheduling problem is beneficial to improve the digitalization level of the shipyard production,optimize the production scheduling plan and reduce waste of production resources.Using big data analysis technology to improve the rationality of the scheduling decision is important to realize the intelligent manufacturing of shipyards in the future.Considering the contradiction between the development trend of intelligent manufacturing and the current production management of domestic shipyards,this paper studies the intelligent production scheduling problem in shipyards.The research process is divided into three parts:（1）Simulation optimization.According to the actual production process,the plant simulation software is used to establish the simulation model of shipyard cutting machining center and the plane segmentation production center,and define the key processing attributes of the workpiece,and consider the random failure of the equipment.The genetic algorithm is used to solve the scheduling problem,and the optimized production scheme is obtained.The example operation shows that the genetic algorithm can effectively improve the performance of production scheduling.（2）Data conversion.The simulation data is exported,and the original production data is converted into machine learning training data through a series of processes such as data integration cleaning,feature conversion,standardization processing,and data marking.（3）Data mining.Three kinds of machine learning models,such as training decision tree,K-nearest neighbor and neural network,are proposed.A comprehensive decision-making mechanism is proposed to give full play to the value of each model.The classification accuracy,recall rate and ₁F,three indicators are used as the judgment criteria.The results show that the comprehensive decision-making mechanism can effectively improve forecast performance.