Personalized route prediction improves by learning hidden costs
A Decision-Focused Neural Optimization Framework for Personalized Route Reproduction from Vehicle Trajectories
Machine Learning
Summary
This paper tackles the problem of predicting which routes individual drivers choose without having to check many alternatives. The authors create a neural network that learns hidden travel costs based on personal traits, trip details, and traffic conditions, then finds the shortest path with those costs. They train the system so it better matches the routes people actually take, even though the shortest path calculation isn’t easy to train directly. Their approach does better than older methods and can help explain why different drivers prefer different routes.
What this means in practice
- •For navigation system developers: Generate personalized route suggestions by learning driver-specific travel cost preferences from historical trajectory data.
- •For urban traffic planners: Better understand heterogeneous driver route preferences to improve traffic models and network management decisions.
Authors
Gyeongjun Kim, Yeseul Kang, Keemin Sohn
Abstract
This study formulates individual route reproduction as a shortest-path problem over learned driver-specific latent link costs. The central idea is that, once such latent costs are inferred from contextual information, observed routes can be reproduced without enumerating alternative route sets. We propose a neural pipeline that includes a perception model that embeds context covariates, which comprises individual characteristics, trip-specific attributes, and network-level traffic states, into the personalized link costs. A constrained optimization (CO) layer, which determines the shortest path (SP) based on these estimated costs, follows the perception encoder. To enable end-to-end training, we employ decision-focused learning to align the predicted shortest paths with observed routes. The implicit maximum likelihood estimation (iMLE) provides an approximate gradient of the loss function that contains the non-differentiable CO layer. Furthermore, a regularization term anchors the latent cost distribution to the empirical scale of observed link travel times, mitigating the scale ambiguity inherent in shortest-path supervision. Empirical evaluations demonstrate that the proposed framework outperforms baseline route choice models in path reproduction. The learned latent costs, interpreted as proxies for perceived travel costs, provide plausible explanations for heterogeneous route choices.