MixTTE: Multi-Level Mixture-of-Experts for Scalable and Adaptive Travel Time Estimation

TL;DR

MixTTE is a scalable, adaptive framework that combines link-level and route-level modeling with novel attention and mixture-of-experts techniques to improve travel time estimation accuracy in large urban networks.

Contribution

It introduces a multi-level mixture-of-experts model with external attention and incremental learning for scalable, real-time traffic prediction.

Findings

Significantly reduces prediction errors compared to baselines.

Successfully deployed in DiDi, enhancing accuracy and stability.

Efficiently models city-scale traffic dynamics and long-tail scenarios.

Abstract

Accurate Travel Time Estimation (TTE) is critical for ride-hailing platforms, where errors directly impact user experience and operational efficiency. While existing production systems excel at holistic route-level dependency modeling, they struggle to capture city-scale traffic dynamics and long-tail scenarios, leading to unreliable predictions in large urban networks. In this paper, we propose \model, a scalable and adaptive framework that synergistically integrates link-level modeling with industrial route-level TTE systems. Specifically, we propose a spatio-temporal external attention module to capture global traffic dynamic dependencies across million-scale road networks efficiently. Moreover, we construct a stabilized graph mixture-of-experts network to handle heterogeneous traffic patterns while maintaining inference efficiency. Furthermore, an asynchronous incremental learning strategy is tailored to enable real-time and stable adaptation to dynamic traffic distribution shifts. Experiments on real-world datasets validate MixTTE significantly reduces prediction errors compared to seven baselines. MixTTE has been deployed in DiDi, substantially improving the accuracy and stability of the TTE service.