Trajectory design using collocation and genetic algorithms: Aircraft turning maneuver

This paper deals with the off-line numerical optimization of trajectories using a direct Hermite-Legendre-Gauss-Lobatto collocation technique. The optimal trajectory synthesis problem is converted to a constrained discrete parameter optimization problem. Since the resulting nonlinear programming problem (NLP) requires good initial guesses for the state and control input histories, a genetic algorithm (GA) with a shooting method is used to find a near-feasible, near-optimal initialization. The GA and HLGL framework is applied to a minimum-time turning maneuver for an aircraft.