实数编码改进遗传算法的非线性最小二乘平差

Nonlinear least squares adjustment based on real coded improved genetic algorithm

遗传算法在处理测量领域中的非线性问题时,算法中的种群数目大小、个体中的参数分量的数量以及参数的取值区间都会对算法的效率产生影响。针对基本遗传算法在处理非线性问题时,容易陷入局部最优值、速度慢、收敛区间小等问题,本文采用了一种新的交叉策略,并对变异算子中的变异步长作动态的自适应改变。最后通过实例解算验证了这种改进的遗传算法比基本遗传算法更加稳定、精度更高、收敛速度更快、收敛区间更大。

Using genetic algorithms in dealing with nonlinear problems in measurement,the number of population size,the number of parameters of individual and the ranges of parameters will affect the efficiency of the algorithm.For the problems with simple genetic algorithm to deal with nonlinear cases,including easy to fall into local optimum,slow and small convergence zone and so on,this paper used a new crossover strategy,and made the steps of mutation operator adaptive change.Finally,an example proved that the improved genetic algorithm was more stable,higher accuracy,faster convergence,greater convergence interval than simple genetic algorithm.