Professional Documents
Culture Documents
Whitepaper - MO. What You Need To Know
Whitepaper - MO. What You Need To Know
Mathematical
Optimization:
What You Need
to Know
Introduction....................................................................................................... 1
Business Understanding........................................................................ 9
Data Understanding............................................................................... 9
Data Preparation.................................................................................... 9
Modeling................................................................................................ 9
Evaluation.............................................................................................. 9
Deployment............................................................................................ 9
AI Technology.................................................................................................... 13
Conclusion........................................................................................................ 19
With MO, organizations can define and In these and other examples, MO factors
embed the key features of their complex, in all the relevant variables, constraints,
real-world problems in MO programming and business objectives, instantly pro-
models. cesses and evaluates an astronomical
number of possible combinations of
MO programming models capture the
these factors (using a math programming
important decisions in business processes.
solver), and then suggests a plan of action
For example, should I:
that optimizes the overall efficiency of a
Use my resources to build product A or B? particular business process.
Send this truck to San Francisco or Dallas? In order to optimize the efficiency of a
Include this stock or a different one in my business process, MO optimally balances
investment portfolio? the trade-offs among numerous (and often
conflicting) business objectives, such as:
MO programming models capture the
Maximizing revenue
resources that are required to execute
these decisions. For example: Minimizing costs
AIRLINES & AIRPORTS METALS MATERIALS & MINING CONSUMER ELECTRONICS TRANSPORTATION TECHNOLOGY
ELECTRICAL EQUIPMENT
ENERGY
AUTOMOTIVE
RETAIL TELECOM
SPORTS SCHEDULING GOVERNMENT
FINANCIAL SERVICES
SEMICONDUCTORS
COSMETICS
CONGLOMERATE
HOSPITALITY
Success Stories
In this section, we share three To solve this “Tail Assignment Optimi-
success stories of organizations zation” problem and build the most efficient
utilizing MO to realize significant schedule for its entire fleet, its operations
business benefits. These stories are research department decided to apply MO
in order to save on fuel and operational
based on four case studies, which can costs, while reducing delay propagation.
be found here on the Gurobi website.
To build the most efficient schedule for
the entire fleet, trade-offs among different
criteria and the possible impact need to be
considered. This makes “Tail Assignment
Air France: Optimization” a multi-criteria mixed-integer
Tail Assignment Optimization programming (MIP) problem.
Air France-KLM is one of the world’s The operations research team built a
largest air carriers in terms of revenue and decision support tool, powered by the
passengers transported. The company’s Gurobi Optimizer, that recommends a
operations research department – which solution that internal users can use to fit
strives to optimize all the carrier’s oper- their specific needs. The tool provides an
ations by transforming data into smart optimal solution with respect to the given
decisions – was facing a “Tail Assignment constraints and the criteria, but the user
Optimization” problem, which consists still has the flexibility to make decisions
of assigning flights to aircraft while resp- based on experience, if needed.
ecting operational constraints. This means Solene Richard, Air France-KLM’s Data
building a schedule for each aircraft by Science and Operational Research Team
setting the sequence of flights that an Leader, commented: “With this new decision
individual aircraft will perform. support tool powered by the Gurobi Solver,
we estimate that we are saving around 1
percent of the fuel costs with huge volumes
for the entire fleet. We also estimate that
we save on delay propagation and on
operational costs. Using mathematical
optimization to solve the tail assignment
problem brings large savings on a yearly
basis.”
BUSINESS DATA
UNDERSTANDING UNDERSTANDING
DATA
PREPARATION
DEPLOYMENT
MODELING
EVALUATION
Figure 1: Data-driven
analytics (DDA) approach
In this step, available data is cleaned and, if MO technologies utilize this DDA problem-
the quality is low, new sources of data are solving methodology. To successfully
identified in an effort to improve quality. deploy an MO tool, it is imperative to
follow all the steps in the DDA approach
highlighted above.
Modeling By adhering to this DDA approach,
During modeling – the most critical step in companies can capture all the key features
the DDA approach (which will be discussed of a complex real-world business problem
in detail in the next section) – we identify as an MO model and build decision support
the modeling techniques that will enable us tools that enable them to optimize their
operational efficiency.
MO Models
In this section, we describe the overall structure of an MO
model in detail. The following figure illustrates this structure.
Resource
Allocation &
Satisfy Demand Scheduling
Business Objectives
Resource
Financial
Utilization
Reports
INPUTS Metrics
Mixed-Integer Programming
(MIP) Models
MIP models are a combination of IP/BIP
and LP models, where the set of decision
variables can be divided into two subsets:
one subset of decision variables that are
continuous and non-negative, another
subset of decision variables that are non-
negative integers or binary. For example,
let’s consider a production process that has
AI Technology
In many of the world’s leading companies, of society and revolutionize the way
AI is perceived as a force of good. AI has businesses operate.
the potential to make the world safer, Forrester reports that as of 2019, more
healthcare more accessible, education more than 50% of global enterprises have already
personal, and manufacturing more efficient. implemented AI technologies and that
In the coming years, AI technologies are around 80% of global enterprises will be
expected to have a profound (and primarily using AI in five years.
positive) impact on virtually every aspect
40%
No to AI
29%
25%
23%
19% 20% 20%
7% 8%
6%
Source: Forrester
Mathematical Optimization
as an AI Technology
The business world has been transformed Techniques such as analytics, ML, and
over the past 50 years by the introduction MO, which make it possible to extract
of computers and information technologies significant value from data
and continues to change due to: Computing resources that can handle
The continued growth in available real-word business problems of any size
data containing a wealth of valuable or complexity
information
Mathematical Optimization
and Heuristics
In problem solving, a heuristic technique In AI, a metaheuristic is a higher-level
is any approach that employs a practical procedure designed to find, generate, or
method not guaranteed to produce a global select a partial search algorithm that may
optimal solution. The idea of heuristic provide a sufficiently good solution to
methods in AI is based on cognitive an optimization problem. Metaheuristics
science, the study of how humans think. sample a set of solutions too large to
Indeed, humans use heuristics all the time be completely sampled. Metaheuristics
to make decisions and solve problems. may make a few assumptions about the
Likewise, heuristic algorithms are often optimization problem being solved and
used in AI to get a computer to find an different metaheuristics are inspired
approximate solution instead of an exact by phenomena occurring in nature.
solution.
MAXIMUM
BUSINESS VALUE
MA
YB
ER
EM
AB OD
BUSINESS VALUE
SO EL
LU
TE
LY
R ET
TIME DEPLOYED RA
IN
NEGATIVE
NEGATIVE
BUSINESS VALUE
= PROBABILISTIC MODEL
Maximize Contact us
Efficiency
Do you want to see how much faster
Gurobi is than your current solver?