Date of Award

12-1987

Degree Name

Master of Science

Department

Computer Science

Access Setting

Masters Thesis-Open Access

Abstract

There has been extensive research in the area of Stochastic Learning in both Psychology and Computer Science. This thesis examines the literature of Stochastic Learning automata with respect to operation control. It traces the development of an optimal reinforcement scheme, and examines the applications of Stochastic Learning automata in routing. The paper simulates M/M/3 and M/M/5 queueing systems similar to the system described by Glorioso and Osorio. The simulation implements Lr-p, Li-p, and Lr-i learning models and compares their performance to Teller Window and Jockeying. A blocking factor is developed to provide more information for the Lr-p learning model. The results show an improvement in the performance of the learning with the blocking factor. The results also demonstrate a possible application of the learning automata in Queueing Theory to create a server device sensitive to the environment.

Download

Share

COinS