Development& Implementation of an Algorithm to Generate Correlated Uniformly Distributed Tri-Dimensional Vectors

Please enable javascript in your browser.

Saberian, Fatemeh | 2011

564 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 42177 (01)
University: Sharif University of Technology
Department: Industrial Engineering
Advisor(s): Akhavan Niaki, Taghi
Abstract:
Nowadays, simulation science supports a wide range of researches. This science is in a constant improvement process. Simulation is a process to fulfill the real world problems by means of experiments which resemble to the real world situation so much. One of the preconditions in statistical simulation is to produce the random number and variables based on the presumed attributes and parameters. In this research, at first, a simple and fast responding algorithm to generate correlated uniformly distributed tri-dimensional vectors is proposed. In this algorithm by producing three groups of random numbers, we can produce tri-dimensional vectors of correlated uniformly distributed based on desired correlation coefficients. At the second part, to implement this algorithm, different software solutions have been candidate. Considering the complexities and functionalities in these algorithms, the neural network approach has been applied. The implemented architecture of proposed neural network has been discussed in details.
Keywords:
Statistical Simulation ; Random Vector Generation ; Correlated Uniformly Distributed Vectors With Zero & 1 Parameters

No TOC