Comparison of the Genetic Algorithm and the Particle Swarm Optimization for Minimizing Fuel Consumption for UAVs

This paper compares the efficiency of two metaheuristics, namely the Genetic Algorithm (GA) and the Particle Swarm Optimization (PSO), for the problem of fuel minimization for uncrewed aerial vehicles (UAVs). When flying search-and-rescue or surveillance missions, it is critical to minimize fuel consumption in order to increase the autonomy of the vehicles. The method proposed consists of using the metaheuristic to compute optimized power settings for the UAV along the trajectory. The approach involves path smoothing in 3D using circular arcs for fixed-wing UAVs. Equations of motion and Newton's equation are used to calculate fuel consumption based on power settings. The metaheuristics are used to compute optimized power settings while adhering to physical constraints. The method is parallelized on a multicore processor for faster computation and optimization in case of trajectory changes. Results show that both the GA and the PSO are efficient at minimizing fuel consumption for the UAV with the PSO being slightly superior. The proposed method can reduce fuel consumption by up to 15% and the parallel implementation is 13.8× faster than a sequential implementation on the CPU.