Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

An Integrated Epidemic Simulation Workflow for Submodular Intervention Strategies

Reet Barik, Marco Minutoli, Mahantesh Halappanavar, Ananth Kalyanaraman

TL;DR

A workflow that will allow researchers to simulate the spread of an infectious disease under different intervention schemes under different intervention schemes is built using the Covasim simulator for COVID-19 alongside a network-based PREEMPT tool for vaccination.

Abstract

Owing to the ongoing COVID-19 pandemic and other recent global epidemics, epidemic simulation frameworks are gaining rapid significance. In this work, we present a workflow that will allow researchers to simulate the spread of an infectious disease under different intervention schemes. Our workflow is built using the Covasim simulator for COVID-19 alongside a network-based PREEMPT tool for vaccination. The Covasim simulator is a stochastic agent-based simulator with the capacity to test the efficacy of different intervention schemes. PREEMPT is a graph-theoretic approach that models epidemic intervention on a network using submodular optimization. By integrating the PREEMPT tool with the Covasim simulator, users will be able to test network diffusion based interventions for vaccination. The paper presents a description of this integrated workflow alongside preliminary results of our empirical evaluation for COVID-19.

An Integrated Epidemic Simulation Workflow for Submodular Intervention Strategies

TL;DR

A workflow that will allow researchers to simulate the spread of an infectious disease under different intervention schemes under different intervention schemes is built using the Covasim simulator for COVID-19 alongside a network-based PREEMPT tool for vaccination.

Abstract

Owing to the ongoing COVID-19 pandemic and other recent global epidemics, epidemic simulation frameworks are gaining rapid significance. In this work, we present a workflow that will allow researchers to simulate the spread of an infectious disease under different intervention schemes. Our workflow is built using the Covasim simulator for COVID-19 alongside a network-based PREEMPT tool for vaccination. The Covasim simulator is a stochastic agent-based simulator with the capacity to test the efficacy of different intervention schemes. PREEMPT is a graph-theoretic approach that models epidemic intervention on a network using submodular optimization. By integrating the PREEMPT tool with the Covasim simulator, users will be able to test network diffusion based interventions for vaccination. The paper presents a description of this integrated workflow alongside preliminary results of our empirical evaluation for COVID-19.

Paper Structure

This paper contains 13 sections, 1 theorem, 2 equations, 4 figures.

Table of Contents

  1. Introduction
  2. Related Works on Epidemic Simulators
  3. Influence Maximization-based Intervention
  4. Experimental Framework
  5. Covasim
  6. PREEMPT
  7. Experimental Evaluation
  8. Experimental Setup
  9. Experimental Results
  10. Effect of seed selection strategies
  11. Effect of vaccinating in batches of uniform size
  12. Effect of vaccinating in batches of non-uniform sizes
  13. Conclusion

Key Result

theorem 1

Given a graph $G_i=(V,E)$, a set of initially infected individuals $B \subseteq V$, and an intervention set $S$, the function $\lambda_{G_i}$ of eq:lives-saved is a submodular function of $S$ if $G_i$ is a rooted tree.

Figures (4)

  • Figure 1: Our integrated workflow for submodular epidemic intervention: The simulation is allowed to run unhindered for a month followed by regular vaccination rounds of certain batch sizes every week. Nodes to vaccinate are specified by a seed selection strategy, which could internally implement various strategies to identify those seeds.
  • Figure 3: Effects of the seed selection strategy on disease spread: The x-axis represents the % of population vaccinated at a single round, on the 31st day of the simulation. The y-axis represents the cumulative #infections after 5+ months of simulation as a % of the population infected.
  • Figure 4: Effects of temporally spacing out vaccination rounds on disease spread using a uniform batch size: For every curve labeled as 'X_Y_Z', 'X' stands for the seed selection strategy; 'Y' stands for the (uniform) batch size used at each round; and 'Z' stands for the number of rounds. Every vertical dashed line represents a vaccination round. Also shown for comparative reference, is the 'No vaccine' curve that corresponds to zero vaccines given out at each round. We use 20,000 as the total number of vaccines.
  • Figure 5: Effects of varying the batch sizes as the disease spreads: The plots are labeled as 'X_Y' where 'X' stands for the seed selection strategy and 'Y' represents the two batching strategies---uniform or non-uniform. The uniform strategy applies the same number of vaccines per round (1,000 vaccines per round). The non-uniform strategy uses varying batch sizes. For these experiments, we used: 2,000 vaccines in each of the first 5 rounds; 1,000 vaccines in each of the next 5 rounds; and 500 vaccines per round in the final 10 rounds. Also shown is the 'No vaccine' curve for reference.

Theorems & Definitions (3)

  • definition 1: EpiControl
  • theorem 1
  • definition 2