基于遗传算法解决附有不等式约束的最小二乘平差问题的研究.pdf

1 1 1,2 朱建军 , 欧阳文森 , 文小岳 ( 11 , 410083; 21 , 410001) : 测量数据处理中经常有些先验信息可以利用, 这些先验信息可以总结成等式或不等式附等 式约束的 差理论目前已经十分成熟, 因而如果是等式约束, 则可用附等式约束的间接 差方法来 处理但如果是不等式约束, 则计算相对困难Holland 等人提出的遗传算法在最优化计算中取得 了非常好的效果, 本文尝试将遗传算法引入附不等式约束的 差计算中本文首先介绍了附不等式 约束的最小二乘 差模型, 分析了基于遗传算法解决该问题的理论依据, 进而通过用内罚函法将不 等式约束 差转化为无约束 差, 以方便运用遗传算法, 最终调用Matlab 遗传工具箱来求解 差结 果通过实例分析, 该算法同其它常用的算法进行比较, 证明该方法具有快速的收敛性, 求解结果 良好 : 内罚函数法; 不等式约束; 遗传工具箱 + : P207 12 : B Abstract: In survey data rocessing, sometimes, there has rior information that can be used. This rior information can be ex ressed by equality or inequality constraints on the arameters. The equality_constrained adjustment theory has been maturated and com rehended, so if the constraints are equality, the roblem can be solved by the equality_constrained indirect adjustment theory. However, if the constraints are linear inequality, the roblems become liner inequality constrained adjustment ( LICA) and the com utation will be very difficult. Genetic algorithms ( GA, Holland, 1975) has shown a great effect on the field of o timization, so this a er is trying to introduce the GA into LICA theory. It firstly resents the LICA roblem. Then it analyzes the combination rinci le of GA and LICA and transforms the inequality constraints to equality constraints by inner enalty function for GA o eration. At the last, the adjustment is carried out with the hel of Matlab GA toolbox. Exam le shows that this algorithm is feasible. Key words: inner_ enalty function; linear inequality_constrained adjustment; GAtoolbox 1 2005) , ,