Font Size: a A A

Study Of Self- Learning Random Neural Network And Simulation In Scheduling Management

Posted on:2005-12-05Degree:MasterType:Thesis
Country:ChinaCandidate:B B ChengFull Text:PDF
GTID:2156360152968148Subject:Technical Economics and Management
Abstract/Summary:PDF Full Text Request
Project management has been viewed as an important management method. And, the scheduling management (a broad sense including resourses distribution ) has been the most important factor in it. Traditionally people express schedule with Gantt Chart, etc. However, those methods can't cope with random process. Some other methods can cope with probability whereas can't manage feedback loops. Some great project management softwares may have able to do with GERT network, but some parameters are needed to put into the software. However those parameters are just people want to know. To calculate those parameters is a boring, complex thing that could held back the desiraion of the use of these softwares. Not to mention that those parameters would vary according to the change of the environment, which would have the softwares adjusted greatly time and time.In this article, by the use of computer, we combine the GERT and the Neural NetWork together, apply randomizer and Monte Carlo into our model which would skip the complex methmatics analysis. There are four aspects in the article. First, expatiate the theory of the GERT, Neural NetWork and Monte Carlo in choosing the proper randomizer to get data resources which conform to the reality. Second, it set forth a simulation model in IT industy. Third, apply Monte Carlo to simulation with the data resources and the GERT that have been choosed to forcast the result. So a self-getting data, self-learning and self-analysising system has formed. In this way, the complicated processes of calculation in GERT and parameters are simplified. And, the last, in this article set forth the module idea: one is the module of getting and analysising data, the second is module of randomizer (or data resource), and the third is the module responsible for the interface of the module one, module two and the programme(i.e. module four). Accordingly, the module four is the programme such as the GERT itself, which reflects the data relationships. The module idea ensures the popular application in vary conditions.
Keywords/Search Tags:GERT, Neural NetWork, Monte Carlo, Simulation
PDF Full Text Request
Related items