Designing an efficient and equitable humanitarian supply chain dynamically via reinforcement learning

TL;DR

This paper proposes a reinforcement learning approach, specifically PPO, to dynamically optimize humanitarian supply chains for efficiency and fairness, outperforming heuristic algorithms.

Contribution

It introduces a PPO-based model that prioritizes satisfaction rate, offering a novel dynamic optimization method for humanitarian logistics.

Findings

PPO model outperforms heuristic algorithms in efficiency.

The model emphasizes equitable satisfaction across stakeholders.

Dynamic approach adapts to changing supply chain conditions.

Abstract

This study designs an efficient and equitable humanitarian supply chain dynamically by using reinforcement learning, PPO, and compared with heuristic algorithms. This study demonstrates the model of PPO always treats average satisfaction rate as the priority.