遗传算法在油田产量预报中的应用

油田开发方案设计中，产重预报正确与否与油田开发成败息息相关，它决定了油田生产规模，投资决策等。正确预测油田产量对改善开发生产效果和提高措施作业效益有重要作用，油田产量预报方法很多，如递减曲线法、神经网络法、Poisson分布法等。各种油气田产量预报模型的应用关键在于参数估计，常规参数估计方法为试凑法和最小二乘法，工作量大，可靠性差，所以引入了一种新型遗传算法求取参数最优估计值，并通过遗传算法在Г模型中的应用实例说明了该方法在油气产量预报中的可行性。

APPLICATION OF GENETIC ALGORITHM IN PROUCTION FORECAST OF OILFIELD

In the development plan making of oilfield, production forecast ties up to the success of reservoir development. It determines the developing scale and investment, and plays an important role in improving the developing efficiency and operation performance. There are many kinds of methods to forecast the production of oilfield, such as decline curve method, neural network method, and Poisson distribution method, etc. The key point to use all these methods is parameter estimation. The widely used methods are trial solution and the least square method. They need huge calculations and their results are less reliable. A new kind of intelligent genetic algorithm was employed to get the optimal parameter. Using it in the model has proved its reliability for production forecast of oilfield.