The Quality Forecasting Of Mass Customization Based On Support VectorMachines

Mass customization is a new mode of production which meets moderntechnological development and needs of customers; insuchamodeof production, thequalitycontrol methods of product design and product manufacture are different fromtraditional methods of mass production. Given the complexity of mass customizationand its characteristics, the problem of quality is comprehensive; quality forecastingand simulation, online quality control are good methods to solve the problem ofqualityinmasscustomization.The purpose of this paper is to realize the accurate quality forecasting in themanufacturing process under the condition of small-sample and provide necessarysupport for the physical simulation and quality pre-control. Based on analyzing thecharacteristicsofmass customizationmanufacturing,researchonthekeytechnologiesand quality control system of mass customization is made, the existing methods forquality forecasting are introduced and the shortages of them are analysed. Thereliability of support vector machines learning algorithm is demonstrated throughin-depth research on the basic theoryand characteristics of it. SVM is combined withpracticalproblemsinMC,thequalityforecastingmodelbasedonSVMis established,and the superior performance of SVM is proved through comparing with GM (1, 1)model.SVM theory is applied in quality forecasting of mass customization to solve thedeficiency of artificial neural network and grey theory; it could be widely used inmasscustomization.