Research And Application Of Decision-Making And Optimization Method For Cocoon Vacuum Water Bath Craft

Cocoon cooking process design is the process of decision-making and reasoning by using process knowledge,which is a key link of silk reeling production and has the characteristics of multivariate,nonlinear and strong coupling,etc.Due to the traditional manual trial cooking test method,it is difficult to resolve the problems such as how to acquire and reuse process knowledge,how to improve efficiency and quality of process design,and how to predict and control quality of reelability.To this end,this paper addresses the need for intelligent design of the cocoon cooking process and researches the method of decision-making and optimization of cocoon vacuum water bath craft.1)Construct cocoon cooking data mart.Firstly,taking the fiber inspection agency’s vacuum water bath cocoon cooking craft design scheme as the research object,then analyze the craft design process,characteristics and find out the existing problems.Secondly,the extraction transformation loading（ETL）technology is used to clean and integrate the scattered process data,and then a multi-source heterogeneous cocoon cooking data mart is constructed to provide data support for process decision-making research.2)A case-based reasoning method for optimal decision-making of the cocoon cooking process scheme is proposed.Firstly,the cocoon cooking case is described using an object-oriented triple representation.Secondly,the cocoon quality characteristics are used as the retrieval basis,the candidate cases with high similarity to the target cocoon quality are determined through a hierarchical retrieval strategy.Then,a multi-index comprehensive evaluation of the quality of the reelability in the alternative cases is carried out through the superior and inferior solution distance algorithm（TOPSIS）,and finally the vacuum water bath process plan of the optimal case is obtained by weighing the similarity of the cocoon quality and the quality of the solution.In this way,the rapid optimization and intelligent decision-making of scheme are realized.3)A process scheme optimization method is proposed based on the LSSA-LSSVM reelability quality prediction model.The quality characteristics of silkworm cocoons,vacuum water bath craft characteristics and reelability quality variables are extracted as input and output variables of the least squares support vector machine（LSSVM）.The latin hypercube sampling method（LHS）and the levy flight strategy are introduced to optimize the initialization method and position update method of the original sparrow search algorithm,and an improved sparrow search algorithm（LSSA）is obtained.The optimal combination of hyperparameters(^*,^2*)of LSSVM is obtained by LSSA and then the LSSA-LSSVM prediction model is constructed,and the model prediction results are used as the basis for the adjustment and optimization of the vacuum water bath craft scheme.4)An intelligent decision-making system for cocoon vacuum water bath craft is developed.First of all,the cocoon cooking data mart,the optimal decision-making method of the craft scheme,the prediction model of reelability quality are organically combined with the actual functional requirements of the fiber inspection agency.Then a visual interface for human-computer interaction is built.Finally,the intelligent design of the vacuum water bath craft scheme is realized.