遗传模拟退火算法在初始对准中的应用仿真

捷联惯导系统粗对准结束后,可以用遗传算法来搜索三个误差角,且由于遗传算法的全局寻优能力,在速度上具有很大优势。但遗传算法的局部寻优能力不足,因此得到的结果在精度上也受到了限制。模拟退火算法容易陷入局部最优解,但是具有很强的微调能力。因此,将遗传算法和模拟退火算法结合起来,能很好地解决初始对准的速度和精度的问题。仿真结果证明遗传模拟退火算法可以很好地改善单一遗传算法的局部寻优能力,使得结果精度更高。

Simulation of Genetic Simulated Annealing Algorithm in Initial Alignment

Genetic algorithm could be used to search for error angles of the strapdown inertial guidance system after coarse alignment is completed, for there is great superiority in speed because of its global searching ability. But there are also limitations in local searching ability of the algorithm, so as for the precision of searching results. Simulated annealing tends to be trapped in local optima, but it also possess powerful fine adjustment ability. So the combination of genetic algorithm and simulated annealing could be sufficient to initial alignment for speed and precision. Simulations showed that genetic simulated annealing algorithm could greatly improve local searching ability of the genetic algorithm, and could get results which are better in precision.