Knowledge Before Solutions: Some Reflections on a Successful O.R. Case Study

This paper describes how a capacity planning problem arising in health care services design and optimization was successfully tackled with mathematical programming techniques. What made the project successful was not the design of a sophisticated algorithm providing optimal solutions, but rather the iterative development of an integer linear programming model of the problem, solved by a general-purpose MILP solver. This approach was made possible by the characteristics of the mathematical model itself and the user-friendly tools that were used. As a result, the problem expert could autonomously challenge and improve the model and the data in a countless number of iterations with little or no intervention of the O.R. expert. This allowed to reduce the development cost to zero and the development time to a few days.