Introduction
Contributions
Preliminary concepts and system model
SEIRD epidemic model
Key network science concepts
Network clustering
Homophily
System model
Approach
Social distancing optimization
Direct contact approach
Clustering approach
Contagion potential approach
Optimization formulations
Scalable solutions
-
Sampling approach. similar to MCMC Gibbs Sampling, all nodes at- tempted to re-locate exactly once within the epoch, assuming all other nodes are fixed
-
Grid-based approach. one run of the optimization approaches (discussed in “Social distancing optimization” section) within the grid
Sampling strategy
Grid-based strategy
Approach
Hybrid strategy
Observations
Results
Parameter | Notation | Value |
---|---|---|
Number of iterations | - | 25 |
Population size | ν | 50–4000 |
Simulation area length | X | 50–400 ft., |
Simulation area breadth | Y | 50–400 ft |
Simulation duration | T | 50, 100 time epochs |
SEIRD parameters (Korolev 2021) | σ, γ, α, β | 0.25, 0.1, 0.05, 0.55 |
Contact threshold | d | 6 ft |
Distance threshold | τ | 25 ft |
Init. sus. and inf. fraction | - | 0.7, 0.3 |
Convergence index | π | 0.3–0.5 |
Grid count | Z | 1–64 |
Default parameters
Nodes | 75 | 90 | 105 | 120 | 135 | 150 |
Area (sq. ft) | 100 × 50 | 100 × 60 | 100 × 70 | 100 × 80 | 100 × 90 | 100 × 100 |
Optimization versus sampling approaches
Approach | Coeff. 0 (c0) | Coeff. 1 (c1) | Coeff. 2 (c2) |
---|---|---|---|
Opt-1 (Direct contact) | 0.98 | − 4.24 | 3.39 |
Sam-1 (Direct contact) | − 2.62 | 11.70 | − 1.59 |
Opt-2 (Clustering) | 1.82 | − 7.54 | 4.63 |
Sam-2 (Clustering) | 0.04 | 0.08 | 0.07 |
Effect of Convergence Index
Performance of distributed strategy
Performance of grid-based solution
Scalability analysis
Nodes | 50 | 250 | 1000 | 4000 |
Area (sq. ft) | 50 × 50 | 100 × 100 | 200 × 200 | 400 × 400 |
Grids | 1 | 4 | 16 | 64 |