Optimal Resource Planning Based on Data Matching Performance and Monetary Cost

This article addresses the problem of planning the organization of exam administration in a university in terms of exam group sizes. The problem is modelled as a multi-objective optimization problem, with two categories of objectives: data matching performance and monetary cost. Data matching is used in this context to match student registration records with exam records, with some measure of accuracy, precision, recall, and other related data matching metrics. When students are organized into smaller exam rooms, data matching performs better, but monetary cost is greater due to the higher number of exam supervision staff required. In this study, we model this problem to come up with a set of optimal exam room configurations. Keywords— Data Matching; Multi-objective Optimization; Machine Learning