Scalable Optimization for Drone and Mobile Robot Routing Using Genetic Algorithms

This study addresses the Drone Vehicle Routing Problem (TRPD) in a scenario involving a single drone and mobile robots, aiming to identify the optimal solution. A machine learning-based genetic algorithm (GA) approach utilizing function approximation is proposed, preserving a two-stage structure to enhance computational efficiency. The algorithm's performance is evaluated through numerical experiments and compared to the simulated annealing (SA) method under identical assumptions. The results demonstrate that the machine learning-based GA outperforms SA in solving small- and medium-sized problems, providing superior solutions in less time. However, for large-scale problems, the method exhibits slower performance due to the computational overhead associated with machine learning operations. These findings underscore the potential of integrating machine learning into optimization techniques to effectively address TRPD scenarios.