Comparative Analysis of Floyd-Warshall and Simulated Annealing for Traffic-Based Shortest Path Problems

This paper presents a comparative study of the Floyd-Warshall (FW) algorithm and the Simulated Annealing (SA) heuristic for solving shortest path problems in static and dynamic traffic-based graphs. While FW is efficient on smaller static graphs, it struggles with dynamic and large-scale instances. SA, leveraging adaptive mutation and temperature control, demonstrates superior scalability and performance in dynamic settings. Using both synthetic and real-world traffic data from OpenStreetMap, we evaluate runtime, memory usage, solution quality, convergence, and improvement rate. Results show FW's strength in small graphs and lower memory usage, whereas SA excels in large, dynamic graphs, achieving high improvement rates despite increased memory consumption. Findings suggest that SA is a viable heuristic alternative to traditional algorithms for dynamic traffic routing scenarios.