Optimization models for placing nurse recruiters

Abstract

This thesis addresses the problem of placing active duty nurse recruiters at recruiting stations for the United States Army Recruiting Command (USAREC). The problem can be formulated as an integer programming problem which is generally known as the uncapacitated plant location problem. The objective is to maximize the yearly production of nurse commissions, a random component of the problem. To account for this random variability, Poisson regression was used to estimate the average number of commissions from a school based on distance to recruiter, nurse unemployment, local nurse salary, and number of nursing students in the graduating class. When implemented, the problem generates a large number of variables and constraints. The cpu time required to solve the problem optimally is not practical. Instead, a greedy heuristic was used. Based on several small random problems, the heuristic provides solutions within 5% of optimality on the average. To illustrate possible uses of solutions to the problem, several applications are also discussed