Design of Fault-Tolerant Wireless Sensor Networks Using the GRASP Algorithm

Abstract

This thesis presents a fault-tolerant design approach for Wireless Sensor Networks (WSNs) using the Greedy Randomized Adaptive Search Procedure (GRASP) metaheuris tic. It addresses the challenge of maintaining network connectivity and reliability in the face of random and targeted node failures. The WSN is modeled as an undirected graph, and GRASP is applied to iteratively construct and enhance network topologies via greedy initialization and local improvements. The approach integrates backup paths to ensure re dundancy and robustness. Performance is evaluated through simulation under diverse failure scenarios, demonstrating GRASP’s effectiveness in preserving high delivery rates and net work resilience. Furthermore, the proposed method is compared with existing approaches in the literature, highlighting its superior adaptability and fault tolerance. The thesis presents a detailed methodology, implementation process, and comprehensive performance analysis.