摘要
油田措施规划能够改善增产措施投资结构,有利于提高油田经济效益。建立了措施规划的目标函数以及约束条件,并提出一种改进型多目标优化的方法。考虑到NSGS-II重复个体互不支配,存在大量重复个体被复制到下一代的可能,提出了将重复个体按一定概率复制到下一代,既保证收敛性又能够提高种群的多样性。针对遗传算法全局搜索能力强而局部搜索能力弱的特点,提出将单纯型法辅助NSGA-II进行局部搜索。考虑到单纯型法目前是针对单目标情况设计的,重新定义搜索成功含义,以及单纯型法是在无约束条件下进行搜索,通过对解空间的变换和惩罚函数解决。算例表明,该改进型多目标遗传算法能够提供很好的Pareto解集,通过增油量和成本关系,油田决策者可以宏观把握在不同投资规模下,最大的措施增油量,从而可以做出更加合理的决策。
Oilfield stimulation programming can improve investment structure of stimulation measure and increase economic effects, accordingly, the object function and constraint condition are established and an improved multiple objective optimization algorithm is proposed. Because repeated NSGS-II individuals don’t dominate each other, they may be copied to the next generation, therefore, this paper proposes a method of copying duplicate individuals to the next generation according to a certain size to ensure the convergence and improve the population diversity. According to strong global search capability and weak local search capability, simplex method is proposed to assist NSGA-II algorithm for local search, meanwhile, by the consideration of the simplex method which designed for single target, the meaning of the search is successfully redefined, in addition, the simplex method is conducted without constraints, which can be fixed by solution space transformation and penalty function. A numerical example shows that this improved multiple objective genetic algorithm can provide good Pareto solution set, meanwhile, the relationship between oil stimulation and the cost can help oilfield decision makers macroscopic handle the maximum amount of oil stimulation at different investment scale and make a more rational decision.
出处
《油气藏评价与开发》
CSCD
2016年第5期54-57,62,共5页
Petroleum Reservoir Evaluation and Development
关键词
增产措施
多目标规划
遗传算法
stimulation measure
multiple objective programming
genetic algorithm
