Abstract:

With the multi-specification and small batch manufacturing, in order to achieve fast and accurate time-quota prediction, this study efficient use of enterprise data and proposes a new method based on similarity of processes. The products were encoded depending on styles, components and procedures, and standard time quota database was established to make the work time quick inquiry. This paper established the model of the evaluation indicators for processes similarity, analyzed with principsl component, obtained the weight of each index, and made fuzzy membership functions to calculate the similar coefficients of the benchmark process and sample process. The function relationship between the standard time quotas and the similar coefficients was determine to predict the unknown time by curve fitting with MatLab. The research results show that the thghest index weight is the process (0.0108) , and the lowest index weight is the specification (0.011) . In the case study the predicted time of "zipper" (201s) is close to the actual time by stopwatch (208s) which peoves the high accuracy and feasibility of the method.

Key words: standard time quota, membership function, garment process, principal component analysis