Personnel Scheduling and Testing Strategy during Pandemics: The case of COVID-19

TL;DR

This paper develops optimization models to determine staff scheduling and testing strategies during pandemics, aiming to minimize infection risk while maintaining organizational efficiency, demonstrated through real and synthetic data.

Contribution

It introduces two MINLP models for optimal personnel scheduling and testing strategies during pandemics, solved with genetic algorithms and commercial solvers, to reduce infection risk.

Findings

Reducing infection risk by 25-60% with optimized strategies.

Models effectively balance staff occupancy and testing to prevent disease spread.

Proposed approach adapts to different contact network scenarios.

Abstract

Efficient personnel scheduling plays a significant role in matching workload demand in organizations. However, staff scheduling is sometimes affected by unexpected events, such as the COVID-19 pandemic, that disrupt regular operations. Since infectious diseases like COVID-19 transmit mainly through close contact with individuals, an efficient way to prevent the spread is by limiting the number of on-site employees in the workplace along with regular testing. Thus, determining an optimal scheduling and testing strategy that meets the organization's goals and prevents the spread of the virus is crucial during disease outbreaks. In this paper, we formulate these challenges in the framework of two Mixed Integer Non-linear Programming (MINLP) models. The first model aims to derive optimal staff occupancy and testing strategies to minimize the risk of infection among employees, while the second model aims at only optimal staff occupancy under a random testing strategy. To solve the problems expressed in the models, we propose a canonical genetic algorithm as well as two commercial solvers. Using both real and synthetic contact networks of employees, our results show that following the recommended occupancy and testing strategy reduces the risk of infection 25\%--60\% under different scenarios.