Computation Resource Allocation Based on Particle Swarm Optimization for LEO Satellite Networks

A distributed computing architecture for computation offloading and data processing in low earth orbit (LEO) satellite networks without transmitting data back to the ground station, so that the task execution time is minimized, is introduced in this paper. Current computational resource allocation schemes are not applicable to the LEO satellite networks and can be classified as an optimal solution problem with constraints. To obtain the optimal solution of this problem, the group intelligent optimization algorithm is utilized to approach the minimum task execution time and the influence of CPU frequency and transmission rate, and channel delay is considered synthetically. Besides, considering the task execution time is mainly determined by the processing power, the processing power of each satellite is evaluated and transformed into the weight to adjust the velocity of the particle in particle swarm optimization (PSO) algorithm. Numerical results demonstrate that the improved algorithm has faster convergent speed than the PSO algorithm and costs less time for task execution.