A Neural-Evolutionary Algorithm for Autonomous Transit Network Design

TL;DR

This paper introduces a neural-evolutionary algorithm that combines graph neural networks and evolutionary strategies to optimize autonomous transit network design, significantly improving performance over baseline methods.

Contribution

It presents a novel hybrid approach integrating graph neural networks with evolutionary algorithms for transit network planning, outperforming existing methods on benchmark instances.

Findings

Outperforms learned policy by up to 20%

Outperforms plain evolutionary algorithm by up to 53%

Effective for realistic transit network benchmarks

Abstract

Planning a public transit network is a challenging optimization problem, but essential in order to realize the benefits of autonomous buses. We propose a novel algorithm for planning networks of routes for autonomous buses. We first train a graph neural net model as a policy for constructing route networks, and then use the policy as one of several mutation operators in a evolutionary algorithm. We evaluate this algorithm on a standard set of benchmarks for transit network design, and find that it outperforms the learned policy alone by up to 20% and a plain evolutionary algorithm approach by up to 53% on realistic benchmark instances.