EXHOLD Experience-Aware Real-Time Hold Control for Large-Scale Ride-Hailing Matching at DiDi

Jul 16, 2026

Upload by Skyler

Large-scale ride-hailing matching depends on hold control to improve end-to-end passenger–driver experience by selectively delaying driver–order pairs, reducing cancellations and excessive waiting while mitigating wasted driver effort. Industrial strategies often rely on heuristic thresholding across multiple predictive models, which can be brittle under non-stationary traffic and difficult to optimize for multiple experience objectives. EXHOLD introduces a deployable two-stage framework that decouples experience-aware pair assessment from hold-time execution, learning discrete experience tiers and solving a constrained monotone hold schedule.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages12

Reading Time30 min

Introduction

Why hold control is challenging in production

Limitations of heuristic model-threshold strategies

Our approach: decouple experience-aware assessment from hold-time execution

Method overview

Stage I: experience-tier decision model

Stage II: constrained optimization for monotone hold-time schedule

Experimental results

Online randomized A/B experiments in DiDi Brazil

Ablations and behavioral analyses

EXHOLD: Experience-Aware Real-Time Hold Control for Large-Scale Ride-Hailing Matching at DiDi Xu Liu Didichuxing Co. Ltd Beijing, China [leoliuxu@didiglobal.com](leoliuxu@didiglobal.com) Kai Wan Didichuxing Co. Ltd Beijing, China [peterwan@didiglobal.com](peterwan@didiglobal.com) Zihao Lu Didichuxing Co. Ltd Beijing, China [luzihao@didiglobal.com](luzihao@didiglobal.com) arXiv :2607 .09090v 1 [ cs .LG] 10 Jul 2026 Abstract In large-scale ride-hailing matching systems, hold control is a highleverage mechanism for improving end-to-end passenger–driver experience: by selectively delaying certain driver–order pairs, the system can wait for better opportunities, reduce cancellations and excessive waiting, and mitigate wasted driver effort, ultimately increasing trip success. In practice, many industrial hold strategies rely on heuristic thresholding over multiple predictive models, which can be brittle under non-stationary traffic conditions and hard to optimize for multiple, experience-oriented objectives. We propose EXHOLD, a deployable two-stage framework that decouples experience-aware pair assessment from hold-time execution. In Stage I, we learn a decision model that assigns each driver–order pair to discrete and interpretable experience tiers, optimizing a unified objective that aggregates multiple satisfaction-related signals across the matching funnel. In Stage II, we solve for a monotone hold-time schedule via constrained optimization over empirical quantiles, explicitly enforcing service guardrails that bound unnecessary holding of promising matches while maximizing overall experience improvement. We evaluate EXHOLD through online randomized A/B experiments in DiDi’s production ride-hailing matching system in Brazil. The results demonstrate consistent gains in both marketplace efficiency and passenger–driver experience: EXHOLD increases trip completion and driver income, while significantly reducing passenger cancellations before and after acceptance and improving key funnel efficiency signals such as faster call-to-acceptance. We further conduct targeted ablations and behavioral drill-down analyses, showing that both stages of EXHOLD are essential to the observed gains, and that the policy makes calibrated decisions under spatiotemporal heterogeneity. EXHOLD has been ramped up and is currently serving the Brazil market in production. Keywords Ride-Hailing; Decision Making; Representation Learning; Contextual Bandits; Constrained Optimization 1 Introduction Large-scale ride-hailing platforms solve a continuous stream of matching decisions under strict latency, reliability, and servicequality constraints [28] . For each incoming order, the system evaluates many candidate driver–order pairs (DO pair) and must decide not only who to match, but also when to match [25, 33] . A critical Accepted at the 32nd ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2026) . This is an arXiv author version. lever in this process is hold control: selectively holding certain DO pair candidates for some system-decided time so the system can wait for better opportunities, reduce cancellations and excessive waiting, and mitigate wasted driver effort—ultimately improving end-to-end passenger–driver experience and trip success [1, 23] . Hold control is powerful but delicate: an overly aggressive policy may suppress good matches and harm service levels of the platform, while an overly conservative policy yields little impact [1, 28] . Why hold control is challenging in production. Hold decisions are inherently multi-objective, sequential, and tightly coupled with the matching pipeline [18] . First, experience degradation manifests through heterogeneous funnel outcomes: Passenger Cancellation Before driver Acceptance (PCBA), driver non-response, Passenger Cancellation After driver Acceptance (PCAA), as well as Driver Cancellation After Acceptance (DCAA), have distinct user harm and efficiency im

FAQs

What evidence supports EXHOLD’s effectiveness?

EXHOLD is evaluated with online randomized A/B tests in DiDi’s production system in Brazil, showing improved trip completion and driver income, reduced passenger cancellations before and after acceptance, and better funnel efficiency such as faster call-to-acceptance, backed by ablations and behavioral analyses.

Jul 16, 2026

Upload by Skyler