摘要
小卫星的任务规划是为确定飞行任务目标并根据飞行任务目标制定控制小卫星运行的指令序列。小卫星的任务规划可以归结为在资源一定的情况下的规划与调度问题。对于0-1整型线性和二次型规划(0-1ILQP)问题，本文提出Hopfield神经网的解法。即将0-1ILQP的目标函数和约束条件用罚函数的方法写成能量函数形式，然后利用Hopfield神经网求解出该能量函数的全局最小点，从而求解出原0-1ILQP的最优解。最后，作为一个算例，给出了用此方法成功的在优化卫星任务规划中的应用。
Scheduling small satellite's mission is to determinate its mission objectives and make decision to control its operation according to the mission objectives. So it does factually belong to planning and operation problem. To the 0-1 integral linear and quadratic programming (0-1 ILQP), a new approach is presented in the paper to solve the problem by Hopfield neural networks. First, the objective function and the constraints are integrated as an integer in the form of energy function with penalty function approach. Then, the global minimal point of the energy function, which also is the optimal solution of the 0-1 ILQP, is obtained by using Hopfield neural networks. Finally, an application to optimizing the mission schedule of satellite is presented as an example that is solved successfully with the approach.
出处
《系统仿真学报》
CAS
CSCD
2001年第5期579-581,共3页
Journal of System Simulation
