VECTORIZED DYNAMIC SIZE GENETIC ALGORITHM FOR FAST FLYBY SEQUENCE GENERATION

Optimizing flyby sequences for an interplanetary mission poses a significant challenge due to the unknown number of optimal flyby sequences involved. Traditional global optimization algorithms typically require a fixed number of design variables, which limits their applicability to such problems. Even when the number of flybys is determined, the sheer volume of possible combinations makes it computationally expensive to optimize each sequence. This study presents a Vectorized Dynamic Size Genetic Algorithm (VDSGA), a method aimed at efficiently generating optimal flyby sequences. To enhance computational speed in the evaluation of objective functions, a Neural Network Lambert's Approximator (NNLA) is employed within the approach. VDSGA demonstrates a significant reduction in computational time for generating optimal flyby sequences. The accuracy and feasibility of the preliminary design solutions via this approach are also investigated.