The article details the creation of the didi-ride-skill for OpenClaw, explaining how a single voice command triggers a full ride‑hailing workflow, the underlying MCP toolset, engineering trade‑offs such as file splitting, attention handling, cron isolation, key management, testing strategies, and future roadmap.
This article dissects the core architecture of a miniature ride‑hailing platform, covering domain‑driven design, layered microservice structure, Redis GEO for fast location queries, distributed locking, Netty‑based real‑time messaging, and hot‑cold data separation to handle massive traffic and ensure reliability.
Douyin has introduced a ride‑hailing feature that aggregates third‑party services like Gaode Taxi, offering a limited set of functions while highlighting the strategic shift toward an aggregation model and its potential impact on the competitive landscape of Chinese ride‑hailing platforms.
This article introduces the GeoHash algorithm, explains how latitude and longitude are recursively bisected into binary strings, compressed with Base32, and demonstrates its use for efficiently locating nearby drivers in ride‑hailing services while discussing precision trade‑offs and edge cases.
This article explains the GeoHash algorithm, demonstrates how binary subdivision of latitude and longitude yields compact base‑32 strings, and shows how these hashes can efficiently locate nearby ride‑hailing drivers while highlighting precision limitations and edge cases.
The article examines ride‑hailing location failures—no fix and drift—explains Android vs iOS positioning, satellite and network sources, and presents a monitoring framework plus Wi‑Fi prompts and sensor‑fusion Kalman filtering that together reduce drift, improve accuracy, and boost order fulfillment.
This article presents Didi's comprehensive approach to designing, implementing, and governing a global data metric system for international ride‑hailing, covering business scenarios, metric‑related challenges, organizational structures, process flows, model architecture, time‑zone handling, tooling, and multi‑level governance.
This article walks through the full lifecycle of building a ride‑hailing platform, covering the impact of emergencies, requirement analysis, high‑level and detailed architecture, long‑connection management, GeoHash location algorithms, and experience‑optimisation techniques to ensure reliable, real‑time dispatch.
The article details how Didi’s ride‑hailing team built a full‑link load‑testing simulation‑degree measurement system that quantifies test coverage across five dimensions—interface, scenario, category, link, and module—using normalized metrics, traffic prediction, and scoring formulas to identify gaps, improve stability, and guide future capacity‑planning enhancements.
By separating test‑traffic triggering from result verification, Didi’s ride‑hailing backend uses live‑traffic inspection and replayed offline tests with bucketed validation rules to achieve near‑zero‑cost, full‑coverage QA, catching hundreds of bugs annually and dramatically improving service reliability for drivers and passengers.
Didi’s ride‑hailing order testing has progressed from manual device‑simulated orders to a tool‑based framework and now a self‑service visual platform that lets engineers drag‑and‑drop, share, and auto‑populate scenarios, dramatically cutting effort while supporting hundreds of test cases for thousands of monthly users.
Didi’s ride‑hailing backend evolved from a monolithic service to a DDD‑split architecture and the DuKang framework, then standardized components, rendering gateways, function access, and logging to solve multi‑terminal inconsistency, complex rendering logic, cross‑process integration, and data visibility, creating a unified cross‑end rendering platform supporting over twenty pages.
Didi balances ride‑hailing supply and demand by forecasting regional needs with time‑series and deep‑learning models, then optimally repositioning drivers through integer programming and refining policies via imitation and offline reinforcement learning, ultimately enhancing passenger experience and platform efficiency.
Laser (Latent Surrogate Representation learning) estimates long‑term ride‑hailing market effects by inferring hidden surrogate variables from short‑term outcomes using an iVAE and inverse‑probability weighting, thereby reducing experiment cost and latency while achieving more accurate causal effect predictions than existing baselines.
The article surveys Didi’s driver‑passenger matching challenges and presents a suite of solutions—from greedy nearest‑driver and Kuhn‑Munkres bipartite matching to stable marriage, dynamic and one‑to‑many assignments, reinforcement‑learning, routing and queueing models—while validating assumptions statistically, integrating preference‑aware machine learning, and outlining multi‑objective and digital‑twin future research.
This article explains the origin and system design of a ride‑hailing platform, compares direct and aggregation models, defines coverage and performance requirements, and details a unified fleet onboarding process together with a thread‑pool based concurrent price‑estimation solution that uses caching, priority grouping, and circuit‑breaker protection to achieve scalable, reliable service.
The article proposes a comprehensive method to eliminate driver‑location flicker and angle errors in ride‑hailing apps by calculating point‑to‑path distances, mapping and inserting points for smooth motion, handling yaw scenarios, refreshing at 12 Hz, correcting vehicle heading, and optimizing animation versus static map scenes.
This article chronicles Cheng Wei’s path from a modest upbringing and early setbacks through his tenure at Alibaba to founding Didi, detailing the company’s early struggles, rapid growth, fierce competition with Uber, strategic mergers, and its eventual dominance in China’s ride‑hailing industry.
The article introduces causal inference for ride‑hailing businesses, explaining the difference between causality and correlation, common misconceptions, and how randomized experiments and observational techniques like propensity‑score matching can quantify effects of actions such as coupons, driver assignments, and platform growth decisions.
The paper presents Didi’s three‑layer governance‑algorithm framework for O2O ride‑hailing, addressing high business complexity, limited labeled data, interpretability, and multimodal features through small‑sample, transfer, and multi‑task learning, achieving notable gains in dispute resolution, NPS and CPO while highlighting remaining data and robustness challenges.
This article examines how O2O ride‑hailing platforms such as Uber and Didi rely on layered system architectures and AI‑driven data middle platforms to enable intelligent dispatch, dynamic pricing, and safety mechanisms, detailing the core components, matching algorithms, and machine‑learning models that power these services.
This article examines how O2O ride‑hailing services like Uber and Didi rely on layered system architecture and algorithmic data platforms to enable intelligent matching, dynamic pricing, and safety mechanisms through machine‑learning models and real‑time data integration.
This article interprets a WWW 2020 research paper that proposes an AI-powered online driver repositioning system, detailing its three-stage framework, dispatch task design, optimization via minimum-cost flow, and experimental results showing improved driver efficiency and platform balance.
During the COVID‑19 outbreak, Amap’s engineers and partners created a free “Medical‑Staff Ride” service for Wuhan’s medical personnel, designing, coding, testing, and launching the app‑integrated ride‑hailing feature within three days and adding a reservation function the next day, thanks to prior platform upgrades, tight cross‑functional collaboration, and rapid decision‑making.
The article explains why Didi needs advanced dispatch algorithms, describes the complexities of order‑driver matching from simple one‑to‑one cases to large‑scale bipartite matching, and introduces batch matching, supply‑demand prediction, chain dispatch, and AI‑driven optimizations that together improve global efficiency and user experience.
This article walks through the evolution of a ride‑hailing platform’s dispatch system—from a single‑database prototype to a data‑driven, AI‑powered architecture—detailing architectural choices, big‑data pipelines, model training, real‑time scheduling strategies, and monitoring practices for handling 400,000 daily orders.
This article explains how Didi’s ride‑hailing platform evolved its dispatch system—from basic nearest‑driver assignment to sophisticated batch matching, demand‑prediction, chain dispatch, and reinforcement‑learning techniques—highlighting the operational challenges, algorithmic solutions, and the massive scale impact on user experience.
This article explains why modern ride‑hailing platforms need advanced dispatch algorithms, describes the underlying order‑allocation problem, explores simple and complex matching scenarios, and introduces batch matching, supply‑demand prediction, chain dispatch, and AI‑driven techniques used by Didi to improve efficiency and fairness.
Didi’s ride‑hailing dispatch system has progressed from a simple greedy, first‑come‑first‑served matcher to sophisticated batch, chain, and predictive algorithms that use deep‑learning demand forecasts and reinforcement‑learning optimization to assign drivers under complex business rules, boosting response rates and serving over 30 million daily requests.
The article explains how building a Neo4j‑based social graph of users, drivers, devices and other attributes enables detection of individual and group subsidy‑abuse fraud in ride‑hailing services through multi‑hop relationship analysis and targeted rule‑based alerts.
This article details Didi's 2016 full‑link load‑testing initiative, covering data‑isolation strategies, virtual driver/passenger tooling, trace‑based traffic marking, staged deployment tactics, and the operational insights gained from stress‑testing a massive ride‑hailing platform.
This article analyzes a ride‑hailing business case, using SPSS to fit discount‑order and discount‑revenue curves, derives optimal discount points (84% for fastest order growth and 86% for maximum revenue), and discusses how these insights guide subsidy and marketing strategies.
Using real‑time order and fuel‑point data from the Didi driver app, the author demonstrates a systematic, data‑driven approach to identify the most cost‑effective car models for ride‑hailing across major Chinese cities, complete with methodology, analysis, and city‑specific rankings.
Since its 2016 launch, Didi’s Driver‑Passenger Co‑Display (司乘同显) has evolved into version 2.0, delivering real‑time synchronized navigation, traffic and route data for both riders and drivers, boosting information freshness, reducing calls, and improving ride‑hailing efficiency across multiple service lines.
This article details the design and implementation of a Go‑based backend for a ride‑hailing app, covering real‑time vehicle tracking, route planning with OSRM, bandwidth‑saving UDP + ProtoBuf communication, in‑memory storage, R‑tree indexing, and the full API workflow.
This article describes how a ride‑hailing service built a memory‑based backend that animates cars on a map, covering challenges of sparse GPS updates, route planning with OSRM, protocol selection (UDP), data serialization (Protobuf), storage architecture, geospatial indexing with R‑tree, and the final API design.
Facing explosive traffic growth in 2014, Didi’s ride‑hailing platform tackled critical challenges by redesigning its LBS architecture, replacing unstable long‑connection services with an AIO‑based framework, partitioning databases, adopting Dubbo and RocketMQ for distributed processing, and building a real‑time monitoring and data center using Storm, HBase, and custom SQL‑to‑HBase translation.
This article details the rapid traffic growth challenges faced by Kuaidi Dache from 2013‑2014 and presents representative architectural bottlenecks and the engineering solutions—including LBS optimization, long‑connection redesign, distributed refactoring, a wireless open platform, real‑time monitoring, and data layer transformation—that enabled stable, scalable, high‑performance ride‑hailing services.
Facing explosive growth in 2013‑2014, Kuaidi Dache re‑engineered its ride‑hailing platform by partitioning MongoDB, replacing single‑queue NICs with multi‑queue, rewriting its long‑connection service with AIO, adopting Dubbo and RocketMQ for micro‑services, and building a Storm‑HBase real‑time monitoring and data synchronization pipeline.
