Long-Term Petroleum Product Supply Analysis Through a Robust Modelling Approach

Linear programming approach to economic modelling of petroleum refining has important shortcomings that make it less useful and less robust for the purposes of impact assessments of related policies. These have to do with its natural inability to calibrate observed data and obtaining jumpy responses of the decision variables to smooth exogenous shocks due to the large number of substitutions between the refining processes. Relying on positive mathematical programming literature, in this paper we propose a method that solves these issues. The main idea is that a refining model has to have a non-linear objective function via inclusion of an implicit total cost function that captures the aggregated impact of all other relevant factors that are not explicitly modelled. We discuss in some detail the issues relevant for practical implementation of the proposed approach for interested practitioners