Optimization of China’s strategic petroleum reserve policy: A Markovian decision approach

China is in the process of building Strategic Petroleum Reserve (SPR) as a utility for its oil supply security. In this paper we develop a stochastic dynamic programming model to optimize China’s stockpile policy with the objective of minimizing the discounted SPR policy costs over a finite time horizon. It is shown that a deterministic and Markovian policy is optimal to the model. A recursive procedure is designed to construct the value functions and derive the optimal stockpile acquisition and release rates over time. Post-optimality analysis is performed to investigate sensitivities of the optimal policy to primary parameter assumptions.

► A Markovian decision process model is proposed to optimize China’s oil stockpile policy.
► A deterministic and Markovian policy is proven optimal.
► A recursive procedure is designed to derive the optimal policy.
► Sensitivity analysis discloses parameters that contribute significantly to the variability in the optimal policy.
► The optimal policy is sensitive to the stockpile capacity construction cost.