AI-Based QoS Optimization in Post-Disaster Ad Hoc Networks: Top-K Path Enumeration with Multicriteria Decision Analysis

Abstract

This paper presents the Topkws algorithm, an efficient method for enumerating the k shortest paths in an Ad hoc network. The algorithm is specifically designed for disaster management scenarios, such as earthquakes, and utilizes top-k ranking from nodes. The approach is based on the Topkws recommendation algorithm, which incorporates Machine Learning (ML) techniques. These ML techniques are used for AI-driven recommendation, enhancing the algorithm's ability to handle complex network scenarios. It employs Multicriteria Decision Analysis (MCDA) with a weighted sum method to order paths based on Quality of Service (QoS) attributes like hop count, bandwidth, delay, and cost. By optimizing QoS characteristics, the algorithm effectively reduces traffic in the network, thereby improving overall performance. The research demonstrates the importance of efficiently obtaining the necessary data elements in an ad hoc network using top-k issues, thereby ensuring the optimal use of resources in post-disaster scenarios.