Re-engineering a diverse service system: Multi-criteria resource assignment

Abstract

Balancing supply and demand in service organizations is a complex problem that should be approached using mathematical modeling. This work developed a resource assignment formulation that deals with the mismatch between personnel profile and tasks required skills, the e ect of travel on resource’s capacity, and the unknown relationship between tasks and workload. The formulation proved to be computationally unsolvable with regular computing capabilities; hence, an alternative formulation was developed. The University of Puerto Rico Agricultural Extension Service was used as test bed. Results showed that the alternative formulation performed very well, average utility results were less than 1% lower than the global optimum across 400 replicas, and utility=size for the full-size problem was in the 79th percentile across the 400-replica’s optimal solution. This work introduces JAM, a contribution to clustering analysis, where cuts at di erent levels are used to extract clusters of di erent sizes and densities from Ward’s dendrograms.