Evolutionary algorithm for multi-objective optimization and its application in unmanned flight vehicle trajectory control

To make sure that unmanned flight vehicle safely landed on the ground, it is necessary to control its trajectory. By adopting proper control law and optimization, the vehicle can achieve a perfect landing, and resources can be most economically assigned. It is a multi-parameters and multi-objectives optimization (MPMO) problem. Two primary problems exist in traditional way: must simplify equation and easy to trap in constrained results. To solve these problems, an evolutionary algorithm using following strategies is adopted: 1. An interface for Simulink toolbox of Matlab, serving as core of the fitness function computing module; 2. Norm based Regret Function serving as fitness function; 3. Adaptive crossover and mutation probability; 4. Elitist strategy. Result proves that the "Improved Genetic Algorithm (IGA)" has better ability in dealing with multi-objective optimization. Finally, the trajectory optimization problem of an unmanned flight vehicle is solved, and the result is satisfying.