Using Artificail Neural Networks for Solving Assignment Problems in WorldWide Webs Optimization

Please enable javascript in your browser.

Mirzavand Boroujeni, Nasim | 2018

363 Viewed

Type of Document: M.Sc. Thesis
Language: Farsi
Document No: 51153 (01)
University: Sharif University of Technology
Department: Industrial Engineering
Advisor(s): Eshghi, Kourosh
Abstract:
In this thesis the capacity and flow assignment problems in worldwide webs is considered in order to minimize the total cost of networks and delays in network links. The main result of this problem is increasing the speed of message transmission and providing better quality of service delivery for network users. For pursuing these goals, Recursive Artificial Neural Networks is proposed for solving the problem and because of their high power and accuracy in reaching optimal solutions, they are convergent to optimal solutions very well. The method is tested on four different topologies and the results show that Recursive Artificial Neural Networks have high efficiency to reach optimal solutions in compare to other existing methods in literature
Keywords:
Optimization ; Recurrent Neural Networks ; Artificial Neural Network ; Capacity Assessment ; Flow Assignment ; Worldwide Webs ; Recursive Artificial Neural Networks

No TOC