一种用于海外项目投资优化组合的混合优化方法

Multi-objective portfolio optimization method for overseas oil & gas projects

 华蓓 陈亚强 穆龙新 常毓文 翟光华

 （中国石油勘探开发研究院，北京 100083）

 (Research Institute of Petroleum Exploration ＆ Development, PetroChina, Beijing 100083, China)

 在海外复杂合同模式及经营环境多变的情况下，针对如何实现产量、投资、效益、风险等多个目标优化配置的目标，文中确定了海外项目多目标投资组合优化的基本思路，表征了海外不同合同模式涉及的复杂商业规则和约束条件，建立了考虑时间维度及风险因子的多目标优化数学模型。通过将不同的优化算法进行结合，提出了一种求解多目标优化模型的混合优化方法。该方法先通过排队过滤生成满足目标和约束条件的投资组合解；然后以该解的特征参数作为约束条件进行线性优化，求出投资组合局部最优解；最后以该最优解作为初始投资组合通过遗传算法求解得到一系列投资组合可行解。利用该方法对海外重点地区项目开展了多目标投资组合优化，提出了不同情景下的优化方案和经营策略。

 Portfolio optimization is a multi-objective one which means we need to consider several factors such as production, investment, revenue and risk of overseas projects at the same time in the circumstances of complicated contract models and fickle business environments. Thus, a multi-objective optimization model has been developed for overseas projects. In this model, the intricate business rules, corporate strategies and constraints in the overseas contract models are mathematically characterized. And both the time dimension and the risk factor have been taken into account. Based on the analysis of different optimization methods, a hybrid optimization method was developed and put forward for this multi-objective model. The hybrid method starts with the "rank & cut" method and the best portfolio generated by it is regarded as the constraints for the linear planning. And the genetic algorithm is used to seek the alternative portfolios near the local optimum solution generated by linear planning. This method was well applied in the optimization and planning process of overseas projects and some optimization operation strategies were raised and suggested according to different scenarios.

10.6056/dkyqt201702022