Stochastic Programming Modeling And Solving For Oilfield Development Programming

In the oilfield development, increasing the oil output and reducing the production cost are the Decision Maker (DM) mainly concerned. These objectives are interactional, and seek an optimal strategy is essential. In this paper, mathematical programming are applied to deal with oilfield development problem. Because of in the oilfield development reality some variables are uncertain and random, the stochastic programming models for oilfield development are established. In these models, the variables associate with some objectives and constrained are assumed to be random and normally distributed. When we solve these models, a chance constrained compromise programming model is proposed as a deterministic transformation. Then the problem can be solved by nonliear progamming method. The main contents of this paper are listed as follows:For the case of linear relationship between variables and measure, linear stochastic programming models of oilfield development are established. According to the difference of the decision variables, they can devide into general model and particular model. The general model contains total profits objective, oil increment objective, water increment objective and investment objective. The particular model contains total profits objective, oil production objective, average cost objective and investment objective. A chance constrained compromise programming model is proposed as a deterministic transformation in solving these models, then solved by nonlinear programming method. Two methods are applied to deal with multi-objective programming. One is linearization of the nonlinear components, then linear objective programming method is applied to solve the problem. And the other is solve the nonlinear programming. Then compare these two results.For the case of nonlinear relationgship between oil production and measure, nonlinear stochastic programming models of oil development are established, including two single-objective models and one multi-objective model. Genetic algorithm approach based on nonlinear optimization layer by layer approach is applied to solving the nonlinear multi-objective stochastic programming model after deterministic transformation. Several different rules are made in the solving process of genetic algorithm, and seven kinds of oil development strategies are obtained from the different rules.For the results of the nonlinear multi-objective programming model, evaluation and ordering of these strategies under the rule of similar figures method are given.