Genetic Algorithm Parameter Optimization for Indigenous Telecommunication Satellite Constellation Design

Satellite constellations represent a rapidly evolving industry globally. This paper highlights the significance of this domain, defines pertinent concepts, and frames the satellite constellation design as an optimization challenge. Subsequently, it explores an optimal solution for a ۱۰-satellite constellation in Low Earth Orbit (LEO). This optimization problem comprises ۳۰ variables, resulting in a high-dimensional optimization scenario. Furthermore, this study employs a Genetic Algorithm (GA) optimizer, which necessitates the tuning of four distinct parameters. A well-defined objective function is introduced, considering not only coverage but also computational efficiency and cost. The research focuses on determining suitable values for the initial population size, aiming to achieve optimal parameter settings for each simulation. The results demonstrate that population size is a critical parameter influencing the performance of the Genetic Algorithm. While increasing the population size leads to higher computational cost, it also results in a significant improvement in the optimization outcome. Across five different scenarios, the investigation shows that an optimal solution is achieved when the population size is approximately ۳.۵ times the number of problem variables. This yields a satellite constellation capable of providing ۱۷ hours of coverage out of a possible ۲۴ hours for the designated target, which in this case is Tehran. Additionally, a sensitivity analysis is conducted to examine the impact of improper parameter selection, revealing risks of premature convergence and reduced solution quality.

کلیدواژه ها:

Satellite constellations ، optimization ، Genetic Algorithm ، Telecommunication