Analysis of a hybrid Genetic Simulated Annealing strategy applied in multi-objective optimization of orbital maneuvers
Optimization of orbital maneuvers is one of the main issues in conceptual and preliminary design of spacecraft in different space missions. The main issue in optimization of high-thrust orbit transfers is that the common optimization algorithms such as Genetic Algorithm and Simulated Annealing are not effectual in finding optimal transfer when they are purely used in optimization. In such problems, modified algorithms are required to find the optimal transfer. Such modifications involve consecutive search and dynamic boundary delimitation. This paper presents a direct approach to optimize high-thrust orbit transfers using a hybrid algorithm based on Simulated Annealing and Genetic Algorithm. This multi-objective optimization method considers optimum fuel transfers while minimizing the error of orbital elements at the end of orbital maneuver. Trajectory optimization is conducted based on converting the orbit transfer problem into a parameter optimization one by assigning proper mathematical functions to the variation of thrust vector direction. Optimization problem is solved using intelligent boundary delimitation in a general optimization method. Taking advantage of nonlinear simulation, a technique is proposed to acquire good quantity for optimization variables, which results in enlarged convergence domain. Numerical example of a three dimensional optimal orbit transfer is analyzed and the accuracy of proposed algorithm is presented. Optimality and convergence of the proposed algorithm is discussed by comparing the results obtained by different approaches. Results confirm the efficiency of the proposed hybrid algorithm in comparison to Simulated Annealing and Genetic Algorithm.