基于实数编码遗传算法的GPS短基线整周模糊度搜索

摘要

将遗传算法用于GPS短基线整周模糊度的搜索计算,从实数编码人手,利用二进制编码染色体杂交前后两对染色体之和不变,且子代的最大值为父代取或的值,最小值为父代取与的值的特点,提出一种基于实数编码的遗传算法.该算法在杂交过程中利用该特点在这个最大最小值范围内随机产生一个子代,另一个子代利用它们之和不变的性质来确定.而初始种群的产生和变异同样可利用产生随机数的方法实现.算例表明,基于实数编码的遗传算法在搜索效率和可靠性方面都优于基于二进制编码的遗传算法.%Genetic algorithm is used in searching integer ambiguity of GPS short-baseline data. On the basis of real-coding and the properties that the sum of the chromosomes is invariant, and the biggest value of filial generation is the OR of parent, and the smallest is the AND of parent,an improved algorithm is proposed. The first filial generation is a random value between the biggest and the smallest values and the other is decided by properties that the sum of the two filial generations is invariant. The creation and variation of the initial population can also be achieved by random. Test results show that both the efficiency and reliability of this algorithm based on real-coded genetic algorithm are better than the algorithm based on binarycoding.