This paper discusses the long-term planning problem of oilfield development. The planning aims to deploy the workload of stimulation treatments with uncertain indicators. And the uncertainties of initial stimulation effect, annual effect error and new recoverable reserves are considered. With their uncertainty sets and the adaptability of decisions, a multiobjective multistage robust integer optimization model is constructed. In this model, the total development cost is minimized and the total new recoverable reserves is maximized. And the model makes decisions on the workload of each stimulation treatment in each year under constraints of annual oil production and workload balance. In addition, an algorithm for solving multiobjective multistage robust integer optimization model is proposed, which can obtain the finitely adaptive robust efficient solution set. Finally, a numerical example of long-term oilfield development planning is presented. The planning model with given uncertainty sets is solved, to verify the validity of the proposed model and algorithm.

本文讨论了油田开发的长期规划问题。该规划旨在部署指标不确定的刺激处理工作量。并考虑了初始增产效应、年效应误差和新增可采储量的不确定性。利用它们的不确定性集和决策的适应性，构建了多目标多阶段鲁棒整数优化模型。在这个模型中，总开发成本最小化，总新可采储量最大化。模型在年产油量和工作量平衡的约束下，对每年各增产处理的工作量进行决策。此外，提出了一种求解多目标多级鲁棒整数优化模型的算法，从而得到有限自适应鲁棒有效解集。最后，给出了一个长期油田开发规划的数值例子。求解给定不确定集的规划模型，验证所提出的模型和算法的有效性。