The University of California San Francisco (UCSF) spends approximately $600 million annually on goods and
services. To lower the total cost of goods purchased while maintaining quality and service, UCSF Strategic
Sourcing establishes strategic agreements with select vendors. These agreements are marketed to the campus
community through vendor shows that allow vendors to showcase their value proposition directly to UCSF
researchers and buyers. This paper formulates the creation of the UCSF vendor-show schedule as a bipartite
matching problem with side constraints, which is then solved using a binary integer program. Of the 3,818
vendor-show combinations, the model identifies 153 optimal assignments and increases potential incentives by
62 percent compared to the manually generated 2007 schedule. The model also improves vendor diversity and
selects the most suitable vendors for each show. Last, it dramatically reduces the time required to generate the
schedule. Strategic Sourcing management is pleased with the results and implemented the model for the 2010
vendor-show season.