Epidemic Dynamics and Intervention Measures in Campus Settings Based on Multilayer Temporal Networks

Abstract

This study simulates the spread of epidemics on university campuses using a multilayer temporal network model combined with the SEIR (Susceptible-Exposed-Infectious-Recovered) transmission model. The proposed approach explicitly captures the time-varying contact patterns across four distinct layers (Rest, Dining, Activity, and Academic) to reflect realistic student mobility driven by class schedules and spatial constraints. It evaluates the impact of various intervention measures on epidemic spreading, including subnetwork closure and zoned management. Our analysis reveals that the Academic and Activity layers emerge as high-risk transmission hubs due to their dynamic, high-density contact structures. Intervention measures exhibit layer-dependent efficacy: zoned management is highly effective in high-contact subnetworks, its impact on low-contact subnetworks remains limited. Consequently, intervention measures must be dynamically adjusted based on the characteristics of each subnetwork and the epidemic situations, with higher participation rates enhancing the effectiveness of these measures. This work advances methodological innovation in temporal network epidemiology by bridging structural dynamics with SEIR processes, offering actionable insights for campus-level pandemic preparedness. The findings underscore the necessity of layer-aware policies to optimize resource allocation in complex, time-dependent contact systems.

Keywords: campus epidemics; intervention strategies; multilayer temporal network.

Figure 1

The schematic diagram of East China Normal University’s multilayer time-varying network shows active individuals (red nodes) migrating across four subnetworks (R/D/C/A layers) via red arrows from $t=1$ to $t=4$.

Figure 2

The infection situation in various subnetworks of the campus under different basic reproduction numbers. (a) Daily new infection proportion (%) over time for different basic reproduction numbers ($R_0=1.5,2,3,5$). The blue solid line, cyan dashed line, yellow dotted line, and red dotted–dashed line represent $R_0=1.5,2,3,5$, respectively. (b) Final outbreak size (%) for each layer (Academic, Community, Rest, and Dining) under varying $R_0$ values. Blue, cyan, yellow, and red represent the Academic, Community, Rest, and Dining layers, respectively. Error bars represent 95% confidence intervals. (c) Epidemic peak (%) for each layer at different $R_0$ values. (d) Epidemic duration (days) for each layer across different $R_0$ values.

Figure 3

The suppressive effects of four subnetwork closure measures on the epidemic situation under different basic reproduction numbers. (a) Final outbreak size control rate (%) for each intervention measure across varying $R_0$ values. Blue, cyan, yellow, and red represent $C\to A$ (Community to Academic), $C\to R$ (Community to Rest), $A\to C$ (Academic to Community), and $A\to R$ (Academic to Rest), respectively. (b) Epidemic peak control rate (%) for each intervention measure at different $R_0$ values. (c) Epidemic duration control rate (%) for each intervention measure across varying $R_0$ values. (d) Cumulative infection control rate (%) for each layer under the $A\to R$ intervention measure across different $R_0$ values. Warm orange, bright red, deep purple-red, and deep burgundy represent Academic, Community, Rest, and Dining, respectively.

Figure 4

The impact of participation rate on epidemic control measures for four subnetwork closure measures. (a) Final outbreak size control rate (%) for each intervention measure under varying participation rates. Blue, cyan, yellow, and red represent $C\to A$ (Community to Academic), $C\to R$ (Community to Rest), $A\to C$ (Academic to Community), and $A\to R$ (Academic to Rest), respectively. (b) Epidemic peak control rate (%) for each intervention measure under varying participation rates. (c) Epidemic duration control rate (%) for each intervention measure under varying participation rates. (d) Peak time control rate (%) for each intervention measure under varying participation rates.

Figure 5

The impact of zoned management on epidemic control indicators and the changing effects of participation rates in zoned management on subnetworks. (a) Final outbreak size control rate (%) for different layers under varying $R_0$ values. Blue, cyan, yellow, and red represent the Rest, Academic, Dining, and Community layers, respectively. (b) Epidemic peak control rate (%) for each layer at different $R_0$ values. (c) Epidemic duration control rate (%) for each layer across varying $R_0$ values. (d) Final outbreak size control rate (%) as a function of participation rate, showing the overall infection control rate and infection control rate for each subnetwork.