油田措施配置多目标随机规划

措施规划对于延长油田稳产年限 ,合理地安排稳产措施从而提高采油速度及最终提高采收率是十分必要的 .有些学者建立了油田稳产措施规划的确定性规划 ,即模型中所有的参数均为确定值 ,而实际油田生产中有许多因素是不确定的 .也有文章建立了措施规划的随机规划模型 ,但建立的规划模型仅涉及一个目标函数 ,即所谓的单目标规划 .而在实际的油田措施配置中 ,需要对多项指标提出优化要求 .本文针对油田开发实际中存在的不确定现象 ,利用规划论中处理随机现象的机理 ,建立了油田措施的多目标随机规划模型 ,并采用具有擅长全局搜索、高度鲁棒性特点的遗传算法进行求解 .应用结果表明 ,该模型使措施配置更能反映生产实际.

A Multiple Objective Optimization under Parameter Uncertainty for Measure Program of Oil Field

measure programming is very essential to extend the time limit of stable production, to plan reasonable measures to oil field to enhance the oil recovery rate and the final recovery . Some determinate models have been established on the programming of oilfield production-stabilizing measures, in which parameters are all pre-determined. In fact, many production-factors can not be determined in advance. In this paper, the multi-object model with random parameters has been constructed for the measure programming . The model can be efficaciously solved with the genetic algorithm, which has the characteristic of whole area searching and highly effective. The method has been applied to the practical oil field and shows that the random programming method is very efficient to make oilfield development plan and can be used as a powerful tool.