Peer-to-peer resource sharing

Code for paper "Untapped Capacity of Place-based Peer-to-Peer (P2P) Resource Sharing for Community Resilience" by Zhengyang Li, Katherine Idziorek, Anthony Chen, Cynthia Chen.

Requirements

• Python 3.x.
• Common libraries such as numpy, pandas, matplotlib, scipy, etc.
• Gurobi 10.0.2 (for optimization)
• Networkx 3.1 (for network analysis)

Installation guide

Clone this repo. How to do this? Check this link.

File organization

• data: The derived data used in this project.
• src: The source code of this project.
  • community.py: The code for generating community-based social networks and P2P resource-sharing networks.
  • resource_sharing_model.py: The P2P resource-shairng model.
  • evaluation_metrics.py: The evaluation metrics for resilience loss.
• demo: The demo code and expected output.
  • demo.ipynb: The demo code shows how to use the source code to conduct analysis.
  • q_vec: The resoruce inventory of households before P2P resource sharing.
  • v_vec: The resource demand of households after P2P resource sharing.
• results: The scenario analysis results.
• figs: The figures generated in this project.

Demo

See the Demo folder for the demo code and the expected output.

Framework

The framework of this the place-based P2P resource sharing is shown as below. For more details, please refer to the paper.

1. Place-based social network construction

This module requires the following data:

• Degree distribution.
• Distance decay function.
• Household-to-household distance.
• The proportion of strong and weak ties in the community.

Degree distribution.

• Laurelhurst.

- South Park

Distance decay function.

• Laurelhurst.

- South Park.

Household-to-household distance.

• data/laurelhurst_distance_matrix.npy
• data/southpark_distance_matrix.npy

The proportion of strong and weak ties in the community.

• Laurelhurst: strong ties 0.41296, weak ties 0.58704.
• South Park: strong ties 0.298, weak ties 0.702.

Using the proposed procedure, an example of the generated social network is shown as below.

• Laurelhurst.

- South Park.

2. Place-based resource-sharing network construction

This module requires the following data:

• Resource distribution. See data/laurelhurst_resource_distr_PMF.csv and data/southpark_resource_distr_PMF.csv.
• Sharing preferece. See data/laurelhurst_share_pref_PMF.csv and data/southpark_share_pref_PMF.csv.
• Sharing priority.

For strong ties, the sharing priority is 3; for weak ties, the sharing priority is 2; for strangers, the sharing priority is 1.

3. P2P Resource-sharing Model

This module redistribute the surplus resources to the needy households based on the sharing preference and sharing priority. Here blow is an example of the resource-sharing model.

• Resource sharing model. resource_sharing_model.py

• Before sharing.

• After sharing.

4. Resilience loss analysis

See evaluation_metrics.py. This module calculates the resilience loss of the community after the disaster. Here is an example of the evaluation metrics.

Citation

If you find this code useful in your research, please consider citing:

"Untapped Capacity of Place-based Peer-to-Peer (P2P) Resource Sharing for Community Resilience" by Zhengyang Li, Katherine Idziorek, Anthony Chen, Cynthia Chen.