In a network of devices in close proximity such as Device to Device (D2D) communication network, we study the dissemination of public safety information at country scale. In order to provide a realistic model for the information dissemination, we extract a spatial distribution of the population of Ivory Coast from census data and determine migration pattern from the Call Detail Records (CDR) obtained during the Data for Development (D4D) challenge. We later apply epidemic model towards the information dissemination process based on the spatial properties of the user mobility extracted from the provided CDR. We then propose enhancements by adding latent states to the epidemic model in order to model more realistic user dynamics. Finally, we study dynamics of the evolution of the information spreading through the population.
There are three different simulations available in this package:
1. pyshp
2. progressbar
3. pickle
4. numpy
$ tar zxvf archive.tar.gz
$ cd archive
$ pip install -r requirements.txt
usage: simulation_latent_heterogeneous.py [-h]
--output OUTPUT
--duration DURATION
[--tau TAU] [--mu MU]
[--sim-id SIM_ID]
[--cell-id CELL_ID]
Process SIR simulation with latent states and heterogeneous return probability.
optional arguments:
-h, --help show this help message and exit
--output OUTPUT output directory
--duration DURATION simulation duration in days
--tau TAU simulation step (fraction of day)
--mu MU simulation mu for latent state (fraction of the
population)
--sim-id SIM_ID simulation step (fraction of day)
--cell-id CELL_ID initial cellID
$ python simulation_latent.py --output ./output/latent/ --duration 7 --tau 0.1 --cell-id 0 --sim-id 1 --mu 0.3