Shortest Path Calculation and Edge-Weight Modeling for Gridded Supply Chain Networks

Project done over the course of my 12-week summer data science internship at Lawrence Livermore National Laboratory, under the mentorship of Dr. Ya Ru Fan.

This project explores two different computational problems in application to supply chain networks:

1.) Exploring efficient Dijkstra’s algorithm data structures for building shortest-path dictionaries between resource nodes

2.) Comparing different models for predicting out-of-sample edge weights, in which we develop a radius-based nearest neighbors decision tree model that is able to reach perfect accuracy metrics.

The poster that I presented at the end of my internship can be found here.