LIST OF IMPORTANT POINTS FOR

PRESENTTION

Kindly

a)read the paper twice and

b)discuss in your team and
c) answer the following points in your PPT
and report.
d)You will not be able to answer all the
points, but this will guide you.
The report should not be more than 0
pages , but should be more than 5 pages
( 5 page < pages of report<0 pages)

1. PRESENTING TEAM of AOR COURSE

a) Name of all team members,

b) their email Id
c) and contact number of leaders of the team and
another member of the team

2. Client Organization

. Find the details of the client company. Name, Country where

they are situated (USA, China, India)
No of employees, Revenue, Cost etc.
a) (Most companies Listed in Stock Exchanges, these
details are known figures

1
 b) No of people working in (OR/MS/ Analytics Team ) and
in this project.
c) Main Authors of the project , Write the first four authors.

Meituan, formerly Meituan Dianping, is a China-based e-

commerce platform providing life services. The Company
connects consumers and businesses to provide services satisfying
people's daily eating needs. The Company owns an instant food
ordering and delivery brand, Meituan, as well as provides
services through its mobile application, Meituan. The Company is
also engaged in the operation of a bike-sharing brand, Mobike.

CEO: Wang Xing

Founded: May 00
Website: about.meituan.com
Employees: 4,73
Market Capitalization: .7T HKD
Employees: 4,73
Revenue: 40.5 B (5.8% YoY growth)
Net Income: 3.86 (307.3% YoY growth)
https://www.hkex.com.hk/Market-Data/Securities-Prices/Equities/
Equities-Quote?sym=3690&sc_lang=en
https://www.google.com/finance/quote/3690:HKG?
sa=X&ved=ahUKEwjxxfTm09GJAxUiyDgGHZs6EUYQ3ecFegQIRBAY

3. Project history

a) How did the project get started and then evolve? (year )
 Phase 0 (Before 06): Meituan's on-demand
food delivery service started with a manual dispatch system

2
where experienced dispatchers assigned couriers to orders. As
order volume grew, this system became inefficient.
 Phase (06): Transitioned to an automated, area-level
sequential assignment system that used a greedy algorithm.
However, this approach led to degraded service quality and
unacceptable computational times as order volume increased.
 Phase (07): Upgraded to an area-level batch assignment
system using constructive heuristic methods. This version met
real-time requirements but could not maintain assignment
quality as daily order volumes continued to rise.
 Phase 3 (08 and onwards): Evolved to a citywide, optimal
order assignment system integrating operations research and
machine learning techniques. Launched in 09, this system can
support over 60 million orders daily.

b) What were the problems/issues faced by the organization

and why these issues were important for the organization.
 Dynamic and Sequential Decision-Making: The on-demand food
delivery market necessitates real-time responses to a constantly changing
environment. Meituan's system receives new orders and couriers come
online continuously, demanding dispatch decisions every seconds. The
challenge lies in making these decisions in a dynamic environment while
considering potential future events like changes in order volume, courier
availability, and traffic conditions. Effectively managing this challenge
was crucial for maintaining service quality and scaling the business.
 Balancing Multiple Objectives: Meituan's success depends on satisfying
multiple stakeholders, including consumers, merchants, couriers, and the
platform itself. This requires the simultaneous optimization of various
objectives, such as delivery time for consumers, meal freshness for
merchants, fair wages for couriers, and operational efficiency for the
platform. Failure to balance these needs could lead to dissatisfied
stakeholders and hinder Meituan's growth.
 Uncertainty: The real-world complexities of delivery operations
introduce significant uncertainties that the dispatch system must handle.
Unexpected events like traffic jams, variations in food preparation times,
and courier preferences can impact delivery times and service quality.
These uncertainties make it challenging to accurately estimate courier
behavior, impacting dispatch decisions and potentially leading to
conflicts between stakeholders. For instance, a delay in food preparation

3
 might cause subsequent delays in a courier's route, leading to
dissatisfaction among customers and potentially conflicts with merchants.
 Solving a Large-Scale NP-Hard Integer Programming Problem in
Real Time: Assigning orders to couriers is a complex computational
problem, further complicated by Meituan's scale and the need for rapid
solutions. With millions of orders and couriers, the problem becomes
computationally expensive, particularly because the matching scores for
assigning multiple orders to a courier are not additive. This means that
the score for assigning a batch of orders is not simply the sum of
individual assignment scores. Finding a good solution within a strict -
second timeframe was essential to maintain accurate courier information
and ensure efficient real-time dispatch. Meituan needed to overcome this
challenge to maintain a responsive and efficient delivery network, which
was fundamental to its growth and expansion.

c) Ask for the technical Details (Separately discussed in item

4)

d) Who developed the model? How many resources (machine

hours of computer time, person days of human (OR
Analyst) time) were used . Who was the project leader?
e) What was the size of the modelling team?
f)
What type of challenges, difficulties were encountered
g) Who were the members (stakeholders) in the development
team ?
h) Who were the members (stakeholders) in the
implementation team ?
.
i) Has the model been in the use ? If yes, How long has the
work been used? How frequently it was used (hourly, once
in every shift, daily, weekly, monthly, quarterly) . How
frequently it is updated ?

4
 c) Technical Details: (To be discussed separately)
 d) Model Developers and Resources: The sources
provide names of individuals involved in the research and
development of the dispatch system but don't specify
details about the resources used (machine hours, person-
days). Yile Liang is identified as the principal algorithm
expert leading the research and development of the
system.
 e) Modeling Team Size: The research and development
group for the dispatch system grew to a team of 30-40
people. This suggests an established OR team was
involved.
.
 g) Development Team Stakeholders: The sources name
individuals involved in the development, primarily from
Meituan Inc. and the Department of Automation at
Tsinghua University. However, specific roles within the
development team and details about stakeholder
involvement aren't explicitly mentioned.
 h) Implementation Team Stakeholders: The sources do
not provide information about the implementation team
stakeholders.
 i) Model Usage and Update Frequency: The real-time
intelligent dispatch system has been in use since 09. It
operates continuously, making dispatch decisions every
30 seconds at a city level. This suggests a real-time,
continuous operation rather than hourly, shift-based, or
periodic usage. The sources don't explicitly state the
frequency of model updates.

4. Describe the technical content:

• What managerial problem or question was addressed?

5
• What analysis, software or other product emerged from the
work?
• What type of software was used. Details, Version name,
Computer (like EXCEL SOLVER, AMPL, GAMS, CPLEX,
MATLAB, SAS, Minitab etc)
• What hardware/software/communication environment was
used and why?
• What analytical methods were used? (Optimization,
forecasting, Stochastic Process, Inventory Theory,
Dynamic programming, Stochastic Programming )
• Provide the formulation of the model
• Why those methods and not others? In the case of multiple
methods being used, why these methods were
selected ?
• Was the solution methodology appropriate for the problem
encountered?
• Is the work innovative, If yes, why ? if not , why not ?
• How was the complexity of the project (high medium, low)
• If possible, distinguish the work from what has been done
earlier by others. List one or two other papers that
you may have contacted.

5. Describe Project deployment:

• How was the solution technically deployed?

• Were the analytics embedded in an application accessible
to users and/or other stakeholders?
• How was the modelling work done by the team and results
of the OR/MS model were converted to implementable
decisions.
• What challenges were overcome in gaining an analytics-
based solution to the problem or question being
addressed?

6. Describe the impact:

• What are the quantitative and qualitative measures of
impact on the client?
• Write the impact of the work with respect to rt revenue
cost , profit figure of the company.
• What was the importance of the work to the client? Was it
critically important?
6
• Is there any major policy change happening in a State or
National Government, or International agencies
organization?
• How large are the pertinent client operations? How large is
the relative impact?
• How much of the benefit has been realized so far, and how
much is anticipated in the future?
• How was the impact measured? Can the measured impact
as per accounting norms?
(can be certified by Chartered Accounts /Certified Financial
accountants )
7. What about the portability of the work?
• Is this work or concept reproducible/usable elsewhere in
your industry or other industries/other countries (For
world bank, UNO project done by multi-lateral
organizations etc.) ?
• Is the generic (model data independent, model solver
independent like AMPL)
• If so, by whom?
o This could be by others in the same industry or in other
similar industries/ other

8. RELEVANCE TO INDIA

Discuss how the problem is relevant to

INDIA ?