A dynamic game theoretic framework for process plant strategic upgrade and production planning

Abstract

A dynamic potential game theoretic production planning framework is presented in which production plants are treated as individual competing entities and competition occurs dynamically over a discrete finite time horizon. A modified Cournot oligopoly with sticky prices provides the basis for dynamic game theoretic competition in a multi-market nonlinear and nonconvex production planning model wherein market price adapts to a value that clears cumulative market supply. The framework is used to investigate a petrochemical refining scenario in which a single inefficient refiner faces elimination by its competitors; we demonstrate that there exist conditions under which the threatened refiner may upgrade itself in order to become competitive and escape the threat, or alternatively in which the threat of elimination will never be carried out and the refiner is effectively safe in the given market configuration. Globally optimal dynamic Nash equilibrium production trajectories are presented for each case.