初始种群范围及运行参数对遗传算法解算GPS病态方程的影响

Study of Effects of Initial Population Range and Running Parameters on GPS Ill-conditioned Equations Solved with Genetic Algorithm

通过实验研究了GPS快速定位病态方程中参数的搜索范围对基线及双差模糊度解算精度的影响,通过GPS基线解算实例研究了种群大小、交叉概率、变异概率及最大进化代数等参数设置对GPS基线及双差模糊度解算精度的影响。计算结果表明,如果测站近似坐标精度达到±0.5m以内,仅利用2个历元的单频载波相位观测数据,利用遗传算法可得到较准确的模糊度浮点解,有利于模糊度的快速固定。种群大小、交叉概率、变异概率及最大进化代数对遗传算法解算精度有一定的影响,通过实验将其分别设置为80,0.75,0.02,400,计算结果表明这些参数的设置是合适的。

The impact of parameter search range on the solution precision of GPS ill-conditioned equations was researched by experiment.Also the impacts of four basic running parameters of population size,crossover probability,mutation probability and max evolutional algebra,etc.on the solution precision of GPS ill-conditioned equations were studied by experiment.The calculation results showed that the precision of parameter solutions of GPS ill-conditioned equations could be greatly improved based on genetic algorithms if the precision of the approximate coordinates of stations was within 0.5m,which was highly advantageous to fix integer ambiguities rapidly.The calculation results also showed that the four basic running parameters had some impacts on the solution precision of GPS ill-conditioned equations,and the four parameters were set as 80,0.75,0.02,400,respectively by experiments.Experiments showed that the setup of the parameters were suitable.