Optical Multistage Interconnection Networks Using Neural Network Approach

Abstract

Now a days Optical Multistage Interconnection Networks (OMINs) are being used as communication media for distributed computing systems. Neural network solution is used in case of OMINs in order to avoid crosstalk. In this thesis neural network routing methodologies are discussed that can be used to generate control bits for a broad range of OMINs. The routing methodology makes use of an Artificial Neural Network (ANN) that functions as a parallel computer for generating the routes. The parallel nature of the ANN computation may make this routing scheme faster than conventional routing scheme. A neural network computation algorithm is used to solve for the optimal traffic routing in a general N-node communication network. It has been observed that neural network gives better results in terms of speed and avoid crosstalk. Artificial neural networks have been studied for many years in the hope of achieving human like performance in the fields of speech and object recognition. These networks composed of many nonlinear computational elements operating in parallel and arranged in patterns reminiscent of biological neural nets. The routing methodology makes use of an ANN that functions as a parallel computer for generating the routes. Neural Networks, which are simplified models of the biological neuron system, is a massively parallel distributed processing system made up of highly interconnected neural computing elements that have the ability to learn and thereby acquire knowledge and make it available for use. Neural Networks are a class of systems that have many simple processors-neurons-which are highly interconnected. A neural network is a special form of parallel computer as the function of each neuron is simple, and the behavior is determined predominately by the set of interconnection. Neural Networks have been used to solve a wide range of problems. In this thesis a set of methods have been discussed for developing neural networks that can be used to systematically and repeatably engineer neural networks to solve specific problems.