THE PRODUCTION QUALITY MANAGEMENT BASED ON THE PRECEDENT APPROACH AND THE CLUSTERING OF USE CASES

Abstract

: In this work the problem of various discrepancies knowledge management is considered during realization of technological production processes of polypropylene tubes. Such discrepancies can make serious impact on manufacturing efficiency enterprise for the reasons of the compelled equipment stand still, sharp decline in final product quality, failures to meet time constraints of production shipment for the consumer, etc. In order to remove discrepancies it is required to define quickly the major factors exerting negative impact on production quality. And for each type of a product there can be the set of factors. The main way for work with the large volume of information concerning problems, their factors and ways of elimination is experience of experts. Such way owing to a human factor is not reliable and cannot be considered as the effective solution of the considered problem. In article as the decision the structure of the knowledge base on the basis of precedents (use cases) is offered. The precedent represents the information block including a basic situation and the decision corresponding to it. The offered structure is founded on hierarchy to Isikava's chart, one of popular instruments of quality control, and listed products. For filling of base precedents it is offered to use an algorithm of a clustering of data CLOPE. Results of work are the three-level structure of the knowledge base, model of a precedent, model of processes of addition of a new precedent and search of a precedent in the knowledge base, an algorithm of a clustering of precedents. It was revealed that the preliminary clustering allows reducing search time considerably. This approach can be used at a stage of technological preparation of production