A Sequential Model for Recruitment: High Volume, Random Yields, and Rigid Demand

Topic:

A Sequential Model for Recruitment: High Volume, Random Yields, and Rigid Demand

Time&Date: 

 10:30 am -12:00 pm, 2019/12/6 (Friday)

Venue:

 Room 619, Teaching A

Speaker:

Prof. LI Qing (Hong Kong University of Science and Technology)

Abstract:

We model the process as a large scale dynamic program. Due to the curse of dimensionality, it is impossible to compute the optimal policy. We instead consider an upper bound, which is obtained when the information about all candidates is available at the beginning, and a lower bound, which is obtained when the recruiter sets the number of offers to make in each phase before assessing candidates. Both bounds are easily computable. We show that when the "volume" (i.e., the numbers of applicants and target) is large enough, the upper bound, the lower bound, and the optimal policy all converge to the same limit. Therefore, the lower bound is an effective heuristic. With a simple yet effective heuristic in hand, we can compute the number of offers to make in each phase and the number of phases there should be. We apply our modeling framework and heuristics to a recruitment process of graduate students in a business program. Our counterfactual analysis shows that the outcome of the recruitment process, measured by the total assessment score of the candidates enrolled minus the penalty cost of the total enrollment deviating from the target number, can be improved by up to 8%, if our model recommendations are adopted. In addition, the outcome is not sensitive to the number of phases.