Study And Development Of Iron&Steel Enterprise Product Quality Decision Support System

With the rapid development of the computer technology and database technology, as the core technology of the management information system,database technology have extensive application in a great deal of field. It can accelerate the speed of obtaining and processing information by this way for enterprises. However, it is not a simple thing for obtaining information. In a sense , data and forming information to the explanation of the data with data analysis.On one hand enterprise must set up one computer process system to substitute traditional craft homework way processing the daily business to come to want, set up a real business process system. On the other hand, enterprise's database has born the weight of the important data of enterprises, the analysis to these data is,not merely the query , update and operating of the data, but need to collect useful information from the data. For example, put in order the data according to different kinds and gather analysing , predict etc.Let the computer put the data in order, extract information, find information for us, this is just the function of computer DSS (Decision Support System).With the development of the technology of the database,it is possible that not only let transaction informationing , but also offers powerful technical support for DSS;At the same time because of the development of technology of data warehouse (Data Warehouse , DW ), DSS and data are excavated (Data Mining, DM) combine together, make DSS meet the needs of enterprises further.A characteristic of the modern iron and steel enterprise produces is that the craft is very complicated,and the process is various, a lot of links that steel production should go through smelting , casting in succession , hot rolling , cold-rolled ,etc. every step has influence on the quality of the final products. So it has extremely important meanings to study and set up a modern product quality information DSS , in order to help the business personnel of quality engineering department of iron and steel enterprise to carry on steel products quality analysis, hold the trends of product quality accurately, realize that manage and makes policy effectively to the whole course of whole product quality, lowering costs to the iron and steel enterprise, raise the quality, economic benefits and social benefit of improving enterprises.Because products quality inspection has many piece wide, it is impossible to set up a DSS just rely on database technology .It must set up data warehouse , let data warehouse as basic datum environment of DSS. At the same time,because simple quality analysis can not meet with real work's demands,it must rely on mathematics theory to set up mathematics model to deal with our complicated decision analyse. The goal of this thesis is to design one DSS based on product quality data warehouse of iron and steel enterprise, construction method to explore steel product quality information data warehouse, and design some mathematics analysis model, prediction and controls the product quality information on the basis of data warehouse .It has a lot of implementation methods to set up data warehouse at present,but after comparing and analysing the theory and method that we use warehouse moduling of the data, we adopt Analysis Manager of MS SQL Server 2000 Analysis Services to finish the extraction of the data, wash, translate and load, realize the memory and management of the data in the data warehouse, based on warehouse technology of the data, by online analyse process and Data Mining tools for the means, have designed a set of data analysis and DSS model that information and knowledge support,supports users of enterprises to respond the quality control and business operation of every procedure are optimized etc..Use the controlling components of Decision Cube, carry out the representing and analysing of the data. Linear regression model is introduced to product quality to predict in alloy steel of intensity and control to analyse according to steel characteristic at the mathematics model. It will not influence the comprehensive degree of the sample according to the suitable sample space in probability theory, and it has the same effect between different sample space.With this characteristic,we design a product quality course power index with a lot of different varieties, different time space.And analyse the course power index of quality through it, guides production.Another mathematics model in this thesis is using probability to one optimization method confirm primitive optimum size of blankThis algorithm rubs the fluctuation of every random variable to a total variable, and every variable must accord with a certain known probability distribution , figure out the homogeneous and equation of the total variable.Then study the probability of each point while fetching different value of total variable to seek to solve optimum. Advantage of this algorithm is it can avoid strenuous measurement work, using atotal variable fluctuation to represent different variable to change at random in all.It can only study the randomness of the total variable,needn't consider the fluctuation of other variables.But because of the complexity of the relation between them of randomness increase , so it reduces the optimization with increase of its variance, and it need long time to calculate. While adopting this kind of method, the law choosing the distribution must accord with the random variable of probability distribution function, otherwise it can cause greater error. Every random variable must obeys the normal distribution lawjust as in this thesis, the total variable (or belongs to a variable ) even obeys the normal distribution law, try to get the variance of the total variable under normal distribution in view of the above, adopt mathematics analytical method to calculate total variable fetch what surplus is cut and decreased minimumly at the value, thus got and solved optimumly.Random probability suppose product of total length is a constant, thus draw the optimum scheme.But the reflection actual conditions that the optimizing value can be very good, how great such approximate error is!, can only really get through examining actually, will produce the experiment.If adopt Monte Carlo law first before producing the experiment, according to every parameter random distribution characteristic, produce every parameter value at random, namely the computer samples, ask out to give and imitating the number of times definitely, how many the finished product can accomplish to press fixed shear after rolling completely, and how many is it require is it decrease to cut always and what is lumber recovery respectively to reach completely, the whole course that thus finished steeling rolling on the computer.lt is one of the real state that similar to reflection,Let the users have a preliminary understanding to the result of optimizing.If relatively is far from ideal scheme difference, pre-produce after revising the parameter properly, thus offer science,reliable theoretical foundation for producing actually.The author took the production management system of a large scale iron &steel enterprise LAIGANG as background,designed and realized a DSS of C/S and B/S construction based on data warehouse,made use of SQL Server 2000 as database and data warehouse server,and proposed a logical structure of software system,In this work the data flow of the system was analyzed,the flow direction of data and its process wre described,an a reliable support to the process of data analysis was provided.