A decision model for software maintenance

In this paper we address the problem of increasing software maintenance costs in a custom software development environment, and develop a stochastic decision model for the maintenance of information systems. Based on this modeling framework, we derive an optimal decision rule for software systems maintenance, and present sensitivity analysis of the optimal policy. We illustrate an application of this model to a large telecommunications switching software system, and present sensitivity analysis of the optimal state for major upgrade derived from our model. Our modeling framework also allows for computing the expected time to perform major upgrade to software systems.