Meituan real-time intelligent dispatch system

entry · p2026-09-07-041 · unverified

Meituan has realized a decrease of 20.96% in average order delivery time

tech: meituan-dispatch
passes the definition as: an order-assignment and route/ETA system whose matching scores and route estimates are learned from couriers' historical pickup and delivery trajectories (graph neural networks, inverse reinforcement learning, an offline-trained global coordinator)
from: 2019
region: CN
domain: access
side: benefit
people: 687,000,000 consumers (2024)
effect: -20.96 percent (2024) — average order delivery time fell 20.96% after the learned dispatch system was implemented in 2019
reversible: yes
who holds the switch: owner:Meituan
attribution: named
source kind: primary
the statistic: Meituan's own Edelman-Award paper (INFORMS Journal on Applied Analytics, accepted 31 October 2023) reports that since implementing the dispatch system in 2019, average order delivery time fell 20.96%. The network it reports on serves 687 million consumers and 9.3 million merchants across more than 3,000 cities, delivering more than 60 million orders a day. The paper states the system is built from 'operations research and machine learning techniques' and that route estimation is 'learned from couriers' historical pickup and delivery routes'.
source: https://courses.ie.bilkent.edu.tr/ie479/wp-content/uploads/sites/16/2024/11/Meituans-Real-Time-Intelligent-Dispatching-Algorithms-Build-theWorlds-Largest-Minute-Level-Delivery-Network.pdf
proposed by: researcher (opus), brief 2026-09-06 second harvest, robots-gig-work
received: 2026-09-06T13:34:02-04:00
harvest report: impacts/harvest/2026-09-07-robots-gig-work.md

Rulings, in order

2026-09-06T13:34:54-04:00 · validatorunverified — fetch failed: HTTP 200 application/pdf
2026-09-06T13:38:18-04:00 · sessionnote — Unverified (the PDF fetched but the quote was not matched); no severity ruling until it is.

The proposal as it arrived (JSON)

{
 "id": "p2026-09-07-041",
 "received": "2026-09-06T13:34:02-04:00",
 "proposer": "researcher (opus), brief 2026-09-06 second harvest, robots-gig-work",
 "kind": "entry",
 "sha": "31b9839d3f75217c",
 "data": {
  "kind": "entry",
  "tech": "meituan-dispatch",
  "tech_name": "Meituan real-time intelligent dispatch system",
  "definition": "an order-assignment and route/ETA system whose matching scores and route estimates are learned from couriers' historical pickup and delivery trajectories (graph neural networks, inverse reinforcement learning, an offline-trained global coordinator)",
  "from": "2019",
  "region": "CN",
  "domain": "access",
  "side": "benefit",
  "people": {
   "n": 687000000,
   "unit": "consumers",
   "date": "2024"
  },
  "effect": {
   "n": -20.96,
   "unit": "percent",
   "date": "2024",
   "what": "average order delivery time fell 20.96% after the learned dispatch system was implemented in 2019"
  },
  "reversible": true,
  "switch": "owner:Meituan",
  "statistic": "Meituan's own Edelman-Award paper (INFORMS Journal on Applied Analytics, accepted 31 October 2023) reports that since implementing the dispatch system in 2019, average order delivery time fell 20.96%. The network it reports on serves 687 million consumers and 9.3 million merchants across more than 3,000 cities, delivering more than 60 million orders a day. The paper states the system is built from 'operations research and machine learning techniques' and that route estimation is 'learned from couriers' historical pickup and delivery routes'.",
  "source": "https://courses.ie.bilkent.edu.tr/ie479/wp-content/uploads/sites/16/2024/11/Meituans-Real-Time-Intelligent-Dispatching-Algorithms-Build-theWorlds-Largest-Minute-Level-Delivery-Network.pdf",
  "quote": "Meituan has realized a decrease of 20.96% in average order delivery time",
  "attribution": "named",
  "source_kind": "primary"
 },
 "from_file": "impacts/harvest/2026-09-07-robots-gig-work.md"
}