You are on page 1of 63

Confidential.

This presentation is provided for the recipient only and cannot


be reproduced or shared without Fair Isaac Corporation's express consent.
2011 Fair Isaac Corporation. 1
FICO Xpress Optimization Suite
Oliver Bastert
Senior Manager
Xpress Product Management
September 22 2011
Webinar
2011 Fair Isaac Corporation. Confidential. 2 2011 Fair Isaac Corporation. Confidential. 2
Agenda
Introduction to FICO
Introduction to FICO Xpress Optimization Suite
Performance
Distributed Modelling and Solving
Case Q&A
2011 Fair Isaac Corporation. Confidential. 3 2011 Fair Isaac Corporation. Confidential. 3
Introduction to FICO
2011 Fair Isaac Corporation. Confidential. 4 2011 Fair Isaac Corporation. Confidential. 4
Profile
The leader in predictive analytics for decision management
Founded: 1956
NYSE: FICO
Revenues: $605 million (fiscal 2010)
Products
and Services
Scores and related analytic models
Analytic applications for risk management, fraud, marketing
Tools for decision management
Clients and
Markets
5,000+ clients in 80 countries
Industry focus: Banking, insurance, retail, health care
Recent
Rankings
#1 in services operations analytics (IDC)
#7 in worldwide business analytics software (IDC)
#26 in the FinTech 100 (American Banker)
Offices
20+ offices worldwide, HQ in Minneapolis, USA
2,200 employees
Regional Hubs: San Rafael (CA), New York, London, Birmingham (UK),
Munich, Madrid, Sao Paulo, Bangalore, Beijing, Singapore
FICO Snapshot
2011 Fair Isaac Corporation. Confidential. 5 2011 Fair Isaac Corporation. Confidential. 5
FICO delivers
superior predictive analytic solutions
that drive
smarter decisions.
Thousands of businesses worldwide,
including 9 of the top Fortune 10,
rely on FICO to
make every decision count.
2011 Fair Isaac Corporation. Confidential. 6 2011 Fair Isaac Corporation. Confidential. 6
Transforming Decision Management
Sharpen customer-centric
decisions
Predict customer needs and behavior
Pinpoint best offer and action
PREDICT
Increase customer
profitability
Reduce losses from fraud and risk
Connect all decisions about a customer
PROFIT
ADAPT
Change faster and
respond flexibly
Change business rules instantly
Create a test-and-learn culture
IMPROVE
Continually improve
strategy performance
Model decisions for greater control
Optimize strategies to grow faster
2011 Fair Isaac Corporation. Confidential. 7
IMPROVE Strategy Performance
Model decisions for greater control
Identify the decision drivers and the
effects of every action
Use the decision model as a planning
tool to test changes in the business
environment
Optimize strategies to grow faster
Create analytically derived strategies to
meet specified business objectives
Design strategies with millions of
variables instantly
Uses FICO optimization software to
create analytically driven decisions on
fleet distribution and utilization
Deployed across every key market in
continental Europe
Benefit estimated at $19 million
FICO Xpress tells us, for example:
On Friday morning, bring only four
cars from Heathrow to Mayfair, and
bring another four from Stansted
Airport. The utilization of our fleet has
gone up by one or two percentage
points.
2011 Fair Isaac Corporation. Confidential. 8
FICO: Game-Changing Analytics
First
commercially
available credit
scoring systems
First automated
origination
systems with
analytics
First cross-
bureau credit
scores
First small
business scoring
systems
First neural
network-based
fraud solutions
First cardholder
profiling for fraud
First insurance
underwriting
scoring systems
First adaptive
control systems
for managing
card accounts
First credit line
optimization
solutions
First predictive
systems for
insurance fraud
First analytic
systems for
retailers to
optimize offers
First adaptive
analytics for
fraud
First credit
capacity scores
First score for
prescription
adherence
1960s 1970s 1990s 1980s 2000s
FICO holds 100+ patents in analytic
and decision management technology,
with 150 more patents pending
2011 Fair Isaac Corporation. Confidential. 9 2011 Fair Isaac Corporation. Confidential. 9
Tools Solutions Scores Tools Solutions
Building an Analytic Advantage
Business
Intelligence
Descriptive
Analytics
Predictive
Analytics
Decision
Optimization
Summarize Past and Current Behavior
Automatically
take the ideal
action on each
individual
Target each
decision to a
customers future
behavior
Make different
offers to groups
of customers
Understand
the trends in
the business
Predict Future Behavior and Adapt
D
e
c
i
s
i
o
n

V
a
l
u
e
Analytic Capability
2011 Fair Isaac Corporation. Confidential. 10 2011 Fair Isaac Corporation. Confidential. 10
FICO Product Portfolio
For Lifecycle Specific Decision Processes
Marketing Origination
Customer
Management
Collections and
Recovery
Fraud
Management
Applications
FICO

Precision
Marketing
Manager
FICO

Retail
Action Manager
FICO

Origination
Manager
FICO

TRIAD

Customer Manager
FICO

Debt
Manager

FICO

Recovery
Management
System

FICO

Falcon

Fraud Manager
FICO

Insurance
Fraud Manager
For Any Decision Process
Scores
B2B: FICO

Score FICO

Credit Capacity Index

FICO

Insurance Risk Scores


B2C: myFICO

Tools
Business Rules Management: FICO

Blaze Advisor

Predictive Analytics: FICO

Model Builder
Optimization: FICO

Xpress Optimization Suite FICO

Decision Optimizer
Professional
Services
Custom Analytics
Operational Best Practices
Strategy Design and Optimization
2011 Fair Isaac Corporation. Confidential. 11 2011 Fair Isaac Corporation. Confidential. 11
Introduction to
FICO Xpress Optimization Suite
2011 Fair Isaac Corporation. Confidential. 12
Xpress Optimization Suite
Solvers
Modelling
Development
D
e
p
l
o
y
m
e
n
t
LP
MIP
QP
MIQP
QCQP
MIQCQP
SLP
MISLP
NLP
MINLP
CP
Mosel
MOdelling and Solving Environment Language
XAD
Graphical user interface development using Mosel
.NET/Java/C/C++/VB
IVE
Development Environment
IVE-XAD
GUI development
Programming
Interfaces
Solver API Mosel API BCL*
GUI
* Builder Component Library for modelling in a programming language
2011 Fair Isaac Corporation. Confidential. 13
Xpress-IVE: Mosel & Optimizer
Editor
Debugger
Profiler
Progress graphs
Visualization
Wizards
Mosel extensions
Deployment
2011 Fair Isaac Corporation. Confidential. 14
Production Planning
2011 Fair Isaac Corporation. Confidential. 15
Product Portfolio & Pricing Optimization
FICO Optimization Dashboard: Debt Consolidation Module
Confidential do not copy
2011 Fair Isaac Corporation. Confidential. 16
Portfolio Rebalancing Solution
2011 Fair Isaac Corporation. Confidential. 17
Facility Location with Google Maps integration
2011 Fair Isaac Corporation. Confidential. 18
Key Features and
Benefits of Xpress-Mosel
Features Benefits
Advanced programming languages:
Algebraic modeling language
Procedural programming language
Entire Mathematical Model can be stored
in one place for rapid development and
easy maintenance.
Utilize different solvers in the
same model
From Mosel you can solve LPs, MIPs, MIQPs,
Non-Linear problems, Stochastic problems,
and Constraint problems
Decompose & parallelize a model to
take advantage of multiple
CPUs/cores
Faster solve times
Make full use of your computing infrastructure
through distributed computing
Build a GUI exclusively within
Mosel code
Decreases development time, gets optimization
in front of business user quicker
Portable across operating systems Mosel Model compiled in one OS can be
deployed on all other supported Operating
Systems, decreasing development time
Open, modular architecture,
User extensible
User flexibility to solve the most complicated
optimization problems
not limited to/by predefined language features
Compiled Protects intellectual property
Offers a variety of APIs and data
connectors
Easy deployment and works in heterogeneous
environments
2011 Fair Isaac Corporation. Confidential. 19
Xpress History and Product Focus
26 years of experience in modelling and optimization
24 years of experience in mixed integer optimization
12 years of experience in nonlinear optimization
10 years Xpress-Mosel, modelling and solving environment
Integration of modelling and solving
Focus on (potentially) exact solution methods
Xpress-Solvers often can prove optimality of the solution
They always give you information on the quality of the
solution
2011 Fair Isaac Corporation. Confidential. 20
Xpress Innovations
Solving
1983: LP solver running on PCs
1992: parallel MIP (1997 on distributed PC/Linux networks)
1995/1996 : commercial branch and cut algorithm
1998: bound switching in dual simplex
2003: lift-and-project cuts
2009: parallel MIP heuristics
2010: LP/MIP solver crosses 64-bit coefficient indexing threshold
Modelling
1983: general purpose algebraic modelling language (mp-model)
2001: algebraic modelling language combining modelling, solving,
and programming (Mosel)
2005: profiler and debugger for a modelling language
2005: user-controlled parallelism at the model level
2010: algebraic modelling language supporting distributed computing
2011 Fair Isaac Corporation. Confidential. 21
Xpress differentiators
Unique capabilities for large scale optimization including ability to
solve ultra-large problems (true64bit capabilities) and support for
distributed modeling and optimization
Complete set of state-of-the-art optimization engines that are
robust, reliable and faster than competing solutions
An easy-to-learn, powerful modeling and programming language,
Xpress-Mosel
The premier visual development environment, IVE, for
developing mathematical models
An intuitive drag-and-drop editor for creating GUIs that
seamlessly integrate with the model for rapid prototyping and
deployment
A partner committed to solving all of your most difficult
optimization problems
2011 Fair Isaac Corporation. Confidential. 22
Xpress Optimization Suite Users
2011 Fair Isaac Corporation. Confidential. 23
Recent enhancements
Xpress 7.1 delivers (GA Nov 2010)
Solve much bigger problems
The possibilities are limitless with the enhanced optimization and modeling
support for solving ultra-large-scale models where the number of coefficients
can exceed 2 billion.
Solve large problems faster
Cut solution times dramatically by leveraging distributed execution of Mosel
models that can now be controlled from a master model across a
heterogeneous set of machines.
Get solutions faster with significantly improved solver
performance
Average increase of 50% arithmetic/50% geometric for multi-threaded MIP
Average increase of 50% arithmetic/25% geometric for single-threaded MIP
Average increase of 50% arithmetic/70% geometric for MIQP
Significant improvements to speed and stability of quadratic simplex and
SLP non-linear algorithms
2011 Fair Isaac Corporation. Confidential. 24
Recent enhancements
Improved developer usability
Developers will also enjoy greater productivity from usability
improvements to the development environment and enhanced
modeling functionality such as the MIIS automated modeling
error/infeasibility detection.
Easier to integrate with other applications
Optimization programmers will be pleased by the addition of
simplified and more robust data exchange capabilities between
Mosel and applications.
Xpress 7.2 delivers (GA April 2011)
Exceptional public benchmark performance
2011 Fair Isaac Corporation. Confidential. 25 2011 Fair Isaac Corporation. Confidential. 25
Performance
2011 Fair Isaac Corporation. Confidential. 26
Comparing Solver Performance
Solver performance is important but not the only decision
criterion
Selection of benchmark sets
Represent client mix of problems
Solvable instances but not too simple
Feasibility and optimality check of solution
Numerically stable problems are preferred for performance
benchmarks
The only public benchmarks for optimization solvers is run by
Hans Mittelmann. He frequently changes the benchmarking sets
The best known collection of MIP instances is currently updated
from version MIPLIB 2003 to version MIPLIB 2010 and will
contain for the first time an agreed benchmarking subset.
A benchmark comparison is always a snapshot of the
performance of the available software at a given point in time.
2011 Fair Isaac Corporation. Confidential. 27
The MIPLIB 2003 Experience
Problem
Old Best Known Obj.
Value (*)
Xpress Improved Obj.
Value (**)
GAIN
(|1-(**)/(*)|)
atlanta-ip 95.009549704 90.00987861 5.3%
msc98-ip 20980991.006 19839497.006 5.4%
protfold -30 -31 3.3%
rd-rplusc-21 171182 165395.2753 3.4%
sp97ar 664565103.76 660705646.5 0.6%
stp3d unknown 500.736 N/A
ds 283.4425 116.59 58.9%
momentum3 370177.036 236426.335 36.1%
t1717 193221 170195 11.9%
liu 1172 1102 5.9%
dano3mip 691.2 687.733333 0.5%
O
p
t
i
m
a
l
U
n
s
o
l
v
e
d
Solving Hard Mixed Integer Programming Problems with Xpress-MP:
A MIPLIB 2003 Case Study, Informs Journal on Computing, 2009
by Richard Laundy, Michael Perregaard, Gabriel Tavares, Horia Tipi, and Alkis Vazacopoulos
2011 Fair Isaac Corporation. Confidential. 28
Geometric Mean
is the comparison criterion of choice
Instead of comparing the overall runtime on a given matrix set (or
equivalently the average runtime or arithmetic mean on that set)
the accepted way of comparing optimization solver performance
is by comparing the Geometric Mean
The presence of a few extremely small or large values has no
considerable effect on geometric mean so it measures
performance more accurately than arithmetic mean which is
biased towards large outliers.
The geometric mean denotes the most likely runtime you will
observe for an instance of the test set.
2011 Fair Isaac Corporation. Confidential. 29
Standard LP Problems (Barrier, Simplex)
Public Benchmark by H. Mittelmann as of 29 Apr 2011, GeoMean
0
10
20
30
40
50
60
70
80
90
XPRESS CPLEX GUROBI MOSEK
Standard LP Problems (Barrier, Simplex)
2011 Fair Isaac Corporation. Confidential. 30
Barrier on Large LP Problems
Public Benchmark by H. Mittelmann as of 29 Apr 2011, GeoMean
0
100
200
300
400
500
600
700
XPRESS CPLEX GUROBI MOSEK
Barrier on Large LP Problems
2011 Fair Isaac Corporation. Confidential. 31
MIQP Problems
Public Benchmark by H. Mittelmann as of 29 Apr 2011, GeoMean
0
50
100
150
200
250
300
350
400
450
XPRESS CPLEX GUROBI SCIP
MIQP Problems
2011 Fair Isaac Corporation. Confidential. 32
MIPLIB 2010 Benchmark set
Geometric Mean, single threaded, MIPLIB 2010 paper
0.00
50.00
100.00
150.00
200.00
250.00
300.00
XPRESS CPLEX GUROBI
2011 Fair Isaac Corporation. Confidential. 33 2011 Fair Isaac Corporation. Confidential. 33
Distributed Modelling and Solving
c 2011 Fair Issac Corporation.
Schemes of parallelization
1. Simple submodel run
wait for
termination
compile/load/run
submodel
process results
results
start
results
Submodel
Master
User
c 2011 Fair Issac Corporation.
Schemes of parallelization
2. Iterative sequential submodel runs
(decomposition algorithms)
process results
wait for
termination
start
run submodel
compile/load
submodel
results
results
Submodel
Master
User
c 2011 Fair Issac Corporation.
Schemes of parallelization
3. Independent parallel submodels
wait for
termination
compile/load/run
submodels
process results
results
start
results
start
...
Master
User
Submodel 1 Submodel n
c 2011 Fair Issac Corporation.
Schemes of parallelization
4. Communicating concurrent submodels
wait for events
process events
process results
start
start
...
compile/load/run
submodels
results
broadcast updates/termination
events/results
Master
Submodel 1 Submodel n
User
c 2011 Fair Issac Corporation.
Distributed solving
Use all the computing power available in your
local network by solving (sub)models on
remote machines
run master model
Local
c 2011 Fair Issac Corporation.
Distributed solving
Use all the computing power available in your
local network by solving (sub)models on
remote machines
run master model run submodel
Local Remote
c 2011 Fair Issac Corporation.
Distributed solving
Use all the computing power available in your
local network by solving (sub)models on
remote machines
run master model run submodel
Local Remote
Physical location of model les, input and
result data depending on application
c 2011 Fair Issac Corporation.
Distributed solving:
Location of models and data
1. On local host
physical les or in memory (e.g. included in master
model or in calling host application)
a (master) model can recursively start new instances
of itself
submodel le
data, results
run master model run submodel
load
Local Remote
c 2011 Fair Issac Corporation.
Distributed solving:
Location of models and data
2. On remote host
congurable read/write access on remote machine
submodel le
data, results
run master model run submodel
load
Local Remote
c 2011 Fair Issac Corporation.
Distributed solving:
Location of models and data
3. Centralised repository
eases version control in multi-user environments
submodel le
data, results
run master model run submodel
load
Local Remote
Central repository
c 2011 Fair Issac Corporation.
Distributed applications
Multi-user application with Mosel model as
dispatcher
Mosel server
User
User
D
a
t
a
b
a
s
e
Production
machine
machine
Production
.
.
.
.
.
.
Example: optimization applications in nance
(solving large numbers of small to medium
size problems)
c 2011 Fair Issac Corporation.
Distributed applications
Decomposition with central data store
Database
User .
.
. Mosel
optimization master
Submodel
Submodel
Examples: Column generation in transport or
personnel planning; blockwise
(Dantzig-Wolfe) decomposition in production
planning
c 2011 Fair Issac Corporation.
Distributed applications
Decomposition with remote, distributed data
sources
D
a
t
a
D
a
t
a
User .
.
. Mosel
optimization master
Remote
model
Remote
model
Example: Large-scale planning in
heterogeneous computing environment
2011 Fair Isaac Corporation. Confidential. 8 2011 Fair Isaac Corporation. Confidential. 8
Case Studies
c 2010 Fair Issac Corporation.
Portfolio rebalancing:
Problem description
Modify the composition of an investment
portfolio as to achieve or approach a specied
investment prole.
c 2010 Fair Issac Corporation.
Optimization application in Mosel
Standalone
Data les Mosel model
IVE
Output les
start application
return results
c 2010 Fair Issac Corporation.
Optimization application in Mosel
XAD GUI
Data les Mosel model
XAD
Output les
output
Summary
ration le
Congu-
start application
return results
c 2010 Fair Issac Corporation.
Optimization application in Mosel
Embedded into host application
Mosel model
Output les
output
Summary start application
return results
Java Data les
c 2010 Fair Issac Corporation.
Optimization application in Mosel
Alternative interfaces
output
Summary
Data les
start application
Mosel model
XAD IVE Java
Output les
output
Summary
Data les
ration le
Congu-
return results
output
Summary
c 2010 Fair Issac Corporation.
XAD interface:
Detailed results
c 2010 Fair Issac Corporation.
XAD interface:
Parameter and version log
c 2010 Fair Issac Corporation.
XAD interface:
Multiple run summary
c 2010 Fair Issac Corporation.
Some highlights
Model:
easy maintenance through single model
deployment as BIM le: no changes to model by
end-user
language extensions according to specic needs
Interfaces:
several run modes adapted to different types of
usages
efcient data exchange with host application
through memory
parallel model runs (Java) or repeated sequential runs
(XAD)
c 2010 Fair Issac Corporation.
Aircraft routing:
Problem description
For given sets of ights and aircraft,
determine which aircraft services a ight.
Aircraft are not identical
they cannot all service every ight
a specic maintenance site must be used per plane
some scheduled long maintenance breaks
Starting condition: each aircraft has a starting
position and a specic amount of accumulated
ight minutes
c 2010 Fair Issac Corporation.
Aircraft routing:
Application architecture
Master problem: route selection
Subproblems: route generation (one instance
per plane)
parallel, possibly remote, execution of submodels
User interface (optional): XAD GUI
c 2010 Fair Issac Corporation.
Aircraft routing:
Application GUI
c 2010 Fair Issac Corporation.
Aircraft routing:
Visualization
c 2010 Fair Issac Corporation.
Aircraft routing:
User interaction
2011 Fair Isaac Corporation. Confidential. 34 2011 Fair Isaac Corporation. Confidential. 34
Q&A
Confidential. This presentation is provided for the recipient only and cannot
be reproduced or shared without Fair Isaac Corporation's express consent.
2011 Fair Isaac Corporation. 35
THANK YOU