Improved starting points for heuristic searches in time-optimal slew-maneuver problems

Two methods, Latin hypercube selection and Sobol sequences, are considered for gen-erating a starting point in the search space for a time-optimal, constrained slew maneuver, with the starting point then being used in a heuristic search algorithm. Both methods provide faster early progress, as compared to an initial point generated using random num-bers with a uniform probability distribution. Covariance Matrix Adaptation-Evolutionary Strategy is employed for the search, which is terminated after a fixed number of genera-tions. The resulting partially-converged solutions are kinematically feasible, making them useful either as initial guesses for subsequent, more rigorous searches or as approximate stand-alone solutions.