This work involves the novel application of a traditional technique for multivariate decision making to a complex problem in software selection. In particular, we formalize the problem statement "how does one select an appropriate commercial real-time operating system for a specific application" and we review two previous solutions to this decision making problem. Along the way, we will consider various aspects of the theoretical argument, broaden its application, and address the deficiencies of the two other solution strategies. Then, we present an improved solution to the problem using the well-developed Analytical Hierarchy Process (AHP), which is not traditionally used by software engineers. We then demonstrate the application of this method using the same criteria used in the previous methods. By explicitly representing preference, by providing tools that allow users to set and inspect their judgments, and by affording users with systematic evaluation procedure, the contribution is to help the decision maker to better identify an appropriate real-time operating systems solution without the need for intensive performance testing.
