摘要
抽油机优化设计是一个多目标、多条件约束的优化问题 ,同时也是一个NP计算问题。采用包括复合形法在内的优化方法均不能保证得到全局最优解。为此 ,提出了一种以上冲程最大扭矩因数最小为目标函数的用于常规式抽油机优化设计的遗传算法。该算法首先利用复合形法在可行域内构造遗传算法的初始种群 ,然后在此种群中进行遗传变异 ;同时对抽油机杆件尺寸采用浮点数编码方案 ,采用简单的排序选择算子 ,对个体进行杂交、变异操作。与复合形法优化计算结果比较 ,可以看出采用遗传算法的抽油机优化设计方法是稳定而高效的。
The optimization of beam pumping units is a sort ofmulti-objective problem with multi-condition constraint, and isalso a problem of NP. It is hard to obtain a global optimum byusing the general optimization method. Thereby a geneticalgorithm for the design of the conventional pumping unit is putforward, which takes the mimimum peak torque factor on theupstroke as the objective function. The algorithm generates theinitial seeds in the solution space firstly by adopting the complexmethod, and then execute operation of hereditary variationamong those colony. And the dimensions of the pumping unit'srod components are optimized by float coding method. Thecomputation result proves the genetic algorithm for the pumpingunit is stable and effective.
出处
《石油机械》
北大核心
2004年第7期28-30,34,共4页
China Petroleum Machinery
