https://github.com/jescresswell/route-optimization

Produces optimal routing for freight deliveries, accounting for multiple trucks working in tandem, and multiple dispatch depots.
https://github.com/jescresswell/route-optimization

Last synced: 9 days ago
JSON representation

Produces optimal routing for freight deliveries, accounting for multiple trucks working in tandem, and multiple dispatch depots.

• Host: GitHub
• URL: https://github.com/jescresswell/route-optimization
• Owner: jescresswell
• Created: 2019-02-22T16:15:22.000Z (almost 6 years ago)
• Default Branch: b1
• Last Pushed: 2019-02-22T19:39:10.000Z (almost 6 years ago)
• Last Synced: 2024-08-14T07:10:02.992Z (4 months ago)
• Language: Python
• Homepage:
• Size: 4.9 MB
• Stars: 8
• Watchers: 1
• Forks: 3
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• jimsghstars - jescresswell/route-optimization - Produces optimal routing for freight deliveries, accounting for multiple trucks working in tandem, and multiple dispatch depots. (Python)

README

        
# route-optimization

This repo solves the vehicle routing problem and was implemented for a large Canadian freight company. We take a manifest of deliveries to be made in a day and produce an set of individual manifests and routes for a fleet of trucks. The algorithms can account for vehicle capacities, and having multiple truck depots to dispatch from. Our optimized solution represented a 5% decrease of travel distance on average compared to previous practice at the company, amounting to significant fuel savings.

First we clean the manifest data, a list of deliveries to be made, in the geocodepandas.py file. This fills in missing latitude/longitude data for the delivery locations by geocoding addresses. Then we use GraphHopper to create a distance matrix, an array of distances between each pair of locations in the manifest. Finally, we use the distance matrix to solve the route optimization problem in tsp.py.

Conventions - Address file should start with depot first (starting/ending point for travelling salesman problem (TSP)).

Geocoded file should be in CSV format with (lat/long/address) with the first row being the header "Latitude,Longitude,Address".

Distance matrix file is in CSV format with row/column indices matching that of geocoded file (row=origin, column=destination).

Dependencies: GraphHopper folder assumed to be in same location as dmatrix.jar, with subfolder north-america_canada_ontario-gh containing open source map data from Open Street Maps.

Usage: java -jar dmatrix.jar manifest.csv

produces dmatrix.csv