Hybrid Genetic Algorithm Collocation Method for Trajectory Optimization

摘要

RESEARCH of recent decades has uncovered many advances innthe development of numerical trajectory optimization. Thesenmethods are attempts to solve discretized versions of some form of annoptimal control problem that, in general, cannot be solved analyticallynusing Pontryagin’s minimum principle [1]. Numerousntechniques exist, each one parameterizing the problem and enforcingnthe equations of motion differently. One such technique is that ofndifferential inclusions, which constrains the discretized states atnadjacent nodes to lie on the attainable sets given the admissiblencontrol inputs [2]. Alternatively, direct shooting methods discretizenthe control history and integrate the equations of motion to obtain thenstate histories and evaluate the constraint violations [3]. A thirdntechnique, collocation, enforces the equations of motion throughnconstraints on the derivative of an interpolating function and the statenequations [4]. These various methods have been applied to a numbernof problems, including space vehicle reentry trajectories [5],nkinematic path planning for unmanned aerial and ground vehiclesn[6], spacecraft slewing maneuvers [7,8], and low-thrust orbitntransfers [9–12].