A Finite Horizon Model for the Optimal Upgrade and Replacement of Mainframe Computer Systems

This paper proposes a finite dynamic programming model to explain the optimal decision-making criteria for the upgrade and replacement of mainframe computer systems. The optimal investment rule is the solution to a stochastic dynamic programming model that specifies the system administrator's objective of maximizing profits through three main choices: 'keep', 'upgrade', or 'replace'. It explains several important factors of the model such as the Markov transition probability, its matrix, and a discretization method of the finite DP model. The calibration and simulations prove that the proposed finite DP model can be solved with a proper stochastic Markov transition probability matrix and a discretization method of state variables and suggest that simulated data tracks actual data well enough to show that the key factors of the model can explain the firm's optimal mainframe computer upgrade and replacement decisions. Therefore, this research plays an important role in that it provides the basis for a stochastic optimal stopping model for the investment in mainframe computer systems within the telecommunications industry that describes such investment behavior by focusing on the unique features of computer systems, which are generally associated with technological development.