Numerical solution of dynamic oligopoly games with capital investment

Abstract

This paper discusses how numerical techniques may be used to solve the simultaneous functional equations that arise in general dynamic stochastic games. Unlike the conventional linear-quadratic approach, our methods may be used to address general model specifications that may include non-quadratic objective functions, non-linear equations of motion, and constraints on decision variables. As an illustration, we apply our methods to a dynamic duopoly game in which competing firms play short-run quantity game subject to production cost that can be lowered through investment in capital stock in the long run.