English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Yield forecasting system in discrete manufacturing is a dynamic complex systems with a number of factors, face such a complex and large system is difficult to directly general system to the enterprise, therefore need to give a company's production forecast model. On the background of Wuzhong Instrument company application, this paper by RBF(radical basis function) neural networks and generalized regression neural network algorithm were compared in enterprise applications, to find the most suitable production forecast model in Wuzhong Instrument company, then Analysis and comparison shows the experimental results . The model has been applied to the valve, spool and valve seat yield prediction in the advanced planning and scheduling in which the practical arrangements played a major role for production planning of Wuzhong Instrument company. Introduction Wuzhong Instrument Company is the largest production base of automatic control valve’s research and development in china, meanwhile, has developed into the leader of the application of information technology in the same industry throughout the country[1]. With the increase in production, Wuzhong Instrument production plan has become a major restricted factor of expansion and development. In order to increase accuracy of production capacity forecasting, the domestic and foreign scholars have put forward a lot of production forecasting models[2]. but for discrete manufacturing, with small batch production and multiple species’ features ,the research of automatic adjustment valve prediction are comparatively far less[3].Therefore, we need to create a set of actual production forecasting model to apply in APS(advanced plan system) of Wuzhong Instrument company. Description In order to meet the established production planning and job scheduling needs, it is necessary for Wuzhong Instrument company which is in the condition of many varieties, small batch-to-order model to solve the problems of batch size and delivery date[4]. With the increasingly severe competition of the market, the Wuzhong Instrument Company begins to pay attention to customers’ needs. The production forecasts can predict the peak demand of parts. Then factory can produce some common parts in a relatively low period and reserve them for easing the pressure on the peak of production. Of course, more isn't always better, because warehouse backlog will result in inefficient flow of funds. How to reasonably and accurately predicate production is very important. Analysis of forecasting model Production forecast is that of enterprise according to the existing objective conditions and historical data for the enterprise future development trends and status of the production forecast. 4th International Conference on Mechatronics, Materials, Chemistry and Computer Engineering (ICMMCCE 2015) © 2015. The authors Published by Atlantis Press 1082 Enterprises also can surplus in production capacity or a large surplus, the company in accordance with "experience or data support" to determine and "moderate" to arrange some common parts and components to product, in order to avoid wasting manufacturing resources. In addition, post-production also play a relieve pressure[5]. By analyzing the data Wuzhong Instrument factory (2011-2014 year) as well as access to the relevant planning department staff, the indicators Wuzhong Instrument factory production lines were screened to obtain the initial index of production lines. Wuzhong Instrument obtain from the data analysis, the data of a total of 150,000, but only have the 2000 empty data, part of the field is null, the proportion is very small, so we can take the linear interpolation to fill. Data is based on daily production to record in the database , if the forecast according to daily production volume, the result would be very cumbersome and not intuitive analysis of its regularity, so we follow a month (storage date) such time interval data analysis. There are many product categories, so we conducted an analysis of its valve body, valve sleeve and valve core types. (1) Normalization process which can make parameterized in the same range, it can also speed up the training of the network. So before enter the neural network, which is data normalization process, will cover the range of data defined between [0, 1]. Research of yield forecasting model The yield prediction purpose is to use the program inventory to make no direct relationship between the program and decisive orders, making the production plan have some flexibility and adjustability, so as to solve the problem of inadequate capacity utilization, while reducing to a certain extent production cycle. Within a production cycle, when the order volume, resulting in production efforts becomes saturated, if the instruction will continue to release a production, the enterprise production capacity is saturated and lead to production cycles. So when production efforts saturated, it should consume inventory, the number of actual production plans should take a minimum. On the contrary, when the orders is little, the enterprise overcapacity, it should to replenish valve parts in orders to use when production capacity is saturated. (2) T represents minimum inventory than planned over the number of complete or incomplete at year; 1 P represents a reasonable production quantities of the product; 2 P represents the actual level of production of the product; 3 P represents the product of the month inventory; 4 P represents the smallest year inventory. (3) P represents the number of production forecast. RBF hidden units used as "base" constitutes the hidden layer space, the input vector is mapped directly to the hidden, after space center confirmation determine the mapping is linear [6] between the hidden layer spaces to the output space mapping. The purpose of RBF neural network output layer and the hidden layer achieved is different, the output layer is linear right adjustments, while the hidden layer of RBF parameters is adjusted, using a nonlinear optimization strategy . min max min i i x x x x x − = − 0 0 0 T P T T  >  = ≤  ( ) ( ) 2 T P P P P = − − − 3 4 1

Research on yield forecasting model based on RBF in discrete manufacturing industry application