A Neural-Evolutionary Algorithm for Autonomous Transit Network Design
This addresses the optimization challenge in transit network design for autonomous buses, representing an incremental advancement in hybrid algorithms for this domain.
The paper tackles the problem of designing public transit networks for autonomous buses by proposing a neural-evolutionary algorithm that combines a graph neural net policy with evolutionary operators. The result shows improvements of up to 20% over the learned policy alone and up to 53% over a plain evolutionary algorithm on benchmark instances.
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.