Journal of Natural Disaster Science

OPTIMUM POST-EARTHQUAKE RESTORATION OF A TELEPHONE SYSTEM USING NEURAL NETWORKS

Shigeru NODA
Associate Professor, Department of Social Systems Engineering, Tottori University, Tottori

(Received 7 May 1993 and revised 26 October 1993)

Abstract

An effective procedure is presented for the optimization of post-earthquake restoration of telephone service. Mean field approximation annealing and alternative routing system with trunk reservation policy are combined to optimize the restoration process. A neural decision-making network is modeled and formulated to minimize the likelihood of functional loss until repairs are completed. Functional optimization is considered a better measure than structural optimization. Statistical examples are presented to demonstrate the proposed procedure. Results show that both repair sequencing of the damaged facility and the restoration route can be effectively determined and that the neural system is very useful tool for determining the restoration process.

Key words

neural network, telephone system, traffic analysis, eathquake disaster, restoration process, traveling salesman problem