A comparative analysis of pathfinding algorithms in static environments: modified A*, PSO, and FLA

Abstract

Pathfinding algorithms are crucial in the domain of autonomous navigation, impacting the efficiency and safety of robotic and AI systems. This paper presents a comparative analysis of three prominent pathfinding algorithms: the A* algorithm, Particle Swarm Optimization (PSO), and the Fick’s Law Algorithm (FLA), within a grid-based simulation populated with static obstacles. We evaluate the performance of each algorithm based on path optimality, computational efficiency, and adaptability to a standardized environment. The A* algorithm, known for its heuristic-based search, demonstrates superior performance in finding the shortest path by utilizing a grid-specific heuristic. PSO, inspired by social behavior in nature, showcases flexibility in path trajectory, offering smoother navigation around obstacles. FLA, a newer approach, strikes a balance between the deterministic nature of A* and the stochastic behavior of PSO, showing potential in applications with limited computational resources. Our findings suggest that while A* remains the optimal choice for grid-constrained navigation requiring precise pathfinding, PSO and FLA may offer advantages in scenarios where flexibility and computational simplicity are prioritized. This study enhances the understanding of pathfinding methodologies, paving the way for future research to refine these algorithms for dynamic environments and integrate adaptive heuristic mechanisms for improved real-world applicability.