【摘要】 对基于微分进化(DE)算法的小推力探测器逃逸轨道优化问题进行了研究。首先在极坐标下建立了二维的轨道动力学模型,然后以燃料最省为性能指标,建立了最优化模型。通过对推力方向角初值、方向角导数初值和加速度幅值初值等的猜测,获得逃逸时间较短时协态变量初值序列。仿真发现,随着时间变化协态变量初值呈现一定规律性,于是使用了一组指数函数来预测当逃逸时间变得更长时的协态变量初值,再将此初值代入DE算法中进行寻优,从而能较容易、较快速地寻找到最优轨迹和最优控制。更多还原

【Abstract】 A differential evolution algorithm is used to find optimal solutions for an indirect method for low-thrust escape trajectories.Two-dimensional trajectory dynamic model is presented,and then optimal control theory is applied to produce a boundary value problem.The initial values of thrust angle,its time rate of change and the magnitude of thrust acceleration are guessed instead of initial co-state variable values.Plotting the optimal initial values versus the time of flight shows that they closely follow power law curves.Using the curve fits from the first ten days of data points generates estimates of the co-state variables for more days.Simulation result shows that the optimal solutions can be found quickly and easily by this method.更多还原