Advanced Lessons on the Craft of Optimization Modeling Based on Modeling Sudoku in Excel

In a previous paper we (Weiss and Rasmussen [Weiss, H. J., R. A. Rasmussen. 2007. Lessons from modeling sudoku in excel. INFORMS Trans. Ed.7(2), http://ite.pubs.informs.org/Vol7No2/Weiss/]) demonstrated lessons that can be learned by formulating Sudoku in Excel using Solver's standard tools. In this paper we use the advanced tools and solver engines available with the Premium Solver Platform in order to demonstrate more sophisticated lessons regarding optimization modeling. Optimization modeling is a skill developed by building and testing alternative formulations for new problems. This paper gives advanced undergraduate and graduate students an opportunity to develop the craft of optimization modeling, by presenting the construction of two new alternatives for modeling Sudoku in Excel. We do not present these models because they are more efficient than the previous models; in fact one of them does not work well at all. The main reason is to highlight strengths and weaknesses when using different modeling approaches, and to display some of the additional modeling capabilities available using the Premium Solver Platform rather than the standard Solver.