|
REGULATION AND ERROR FORECAST OF MANUFACTURE SYSTEM USING
NON-STATISTICAL THEORY
Xia
Xintao Wang Zhongyu Zhu Jianmin
(College of Mechanical and Electronical Engineering, Henan University of
Science and Technology, Luoyang 471003)
Li
Xudong
(Technology Department, Luoyang Bearing
Group
Ltd., Luoyang 471039)
|
|
Abstract: At present, a manufacture system is regulated using the
statistical method. The statistical method is based on probability and
statistics theory. Normal distribution is required, the standard
deviation and the coverage factor are beforehand certain for the output
of the manufacture system, and therefore it will be very difficult to
resolve the problems of the Regulation and error forecast of the
manufacture system with non-normal distributions, unknown or
non-comprehensive information by means of the classical statistical
theory. Therefore, a novel method is proposed. This method is the
non-statistical method based on fuzzy-set theory. By using this method,
many typical and untypable problems, such as normal distribution,
uniform distribution, and commingle distributions, can be researched,
the disperse interval of the systemic output can be straightway
estimated without any known characteristic parameter such as the
standard deviation and the coverage factor. Influence of the random
error’s change in size and direction on the systemic output can be
roundly and truly described because the algorithm of fuzzy interval-data
is introduced, and the average values and the maximum difference-values
of fuzzy interval-data are considered in processing the commence and
conclude data. Computer imitation and manufacture cases indicate the
very small error and the very high veracity of estimating for the
systemic output using the non-statistical method proposed.
Key words: Manufacture system Regulation Error Forecast
Non-statistical theory Fuzzy-set theory
CLC No: TH161.1
国家自然科学基金(50375011)、河南省杰出青年基金(02120001500)、河南省创新人才基金(2001-2005)和河南科技大学重大科技前期预研专项基金(2003ZD03)资助项目.
Received 20040311, received in revised form 20040825
|