遗传算法与惩罚函数法在油田开发规划中的应用

APPLICATION OF GENETIC ALGORITHM AND PENALTY FUNCTION METHOD TO OIL FILED DEVELOPMENT PLAN

主要针对油田开发规划优化模型中的产量分配优化模型的求解方法,进行了讨论。遗传算法和惩罚函数法作为通用、高效的优化算法,已应用到工程计算的各个领域。然而传统算法通常只适用于目标函数和约束条件函数可微的情况,而且不能保证求到全局最优解。文中把“遗传算法+惩罚函数法”的通用方法,应用于油田开发规划中,它非常适合于求解复杂的非线性规划问题。此算法既克服了传统优化算法的不足,得到较为理想的全局最优解,同时也改善了遗传算法的局限性。

This paper discusses solutions of optimized production distribution models in the field development plan.Being general,effective and optimal,genetic algorithm and penalty function methods have been applied to all engineering calculations.However,conventional methods are suitable only under conditions of differentiable objective function and constraint function, and optimum solution may not be obtained.The integrated method of 'genetic algorithm and penalty function' is applied to the field development plan,solving the complicated non-linear problems.Compared with the conventional optimum algorithms,the methods can not only obtain the optimum solution,but also reduce limitations in genetic algorithm.