LS-SVM-GA算法在油田产量预测中的应用研究

油田产量预测是油田开发动态分析最重要的内容之一,也是油田开发优化决策的基础。在介绍最小二乘支持向量机(LS-SVM)及遗传算法(GA)的原理基础上,建立LS-SVM-GA模型,并用该模型对某气田天然气产量进行预测。通过二个性能指标将其与LS-SVM和BP神经网络模型进行对比,结果表明,在样本有限保证一定精度的情况下,LS-SVM-GA模型的预测精度较高,范化能力较强,能够利用该模型对气田天然气产量进行预测。

Application of LS-SVM-GA Algorithm in Oil Production Forecasting

Yield prediction is one of the most important elements.Oil field development the dynamic analysis is also the basis for oil field development optimization decision.In introducing the Support Vector Machines(LS-SVM) and genetic algorithm(GA) based on the principle of the establishment of LS-SVM-GA model and the model of a gas field with gas production is predicted.Performance by two of its with the LS-SVM and BP neural network model,results showed that,in the sample limited guarantee the accuracy of the case,LS-SVM-GA model of higher accuracy,range of resistant able to use the model to forecast natural gas production field.