This thesis develops a probabilistic dynamic programming model which may be used to obtain optimal fill-rate and withdrawal-rate strategies for Korea oil stockpiling.
Oil stockpile could be considered as an insurance against possible future oil supply disruptions. Any excess oil stockpile seems expensive during normal periods but its value will be greatly appreciated during the emergency periods.
In our model the planner can control the quantity of oil imports, fill-rate and withdrawal-rate. The impact of oil supply disruption upon Korea is estimated through a macroeconomic production function which has a constant elasticity of substitution (CES) between energy input and non-energy input. The parameters used in the CES function are determined by benchmarking at a normal year.
The time paths of the optimal stockpiling level and stockpiling strategies are analyzed by varying two important parameters-the probability a disruption occurs and the size of price jump induced by the disruption: The effect of changes in the disruption probability and the size of price of price jump induced by the disruption upon the stockpiling level in 1988 is analyzed. This result can be used for the decision maker to respond easily and quickly to the expected change of the future oil market situations.