Advanced Modeling Techniques and New Trends for
Stochastic Programming and Global Optimization
Maximal Software will be hosting a free full-day workshop, Optimization Modeling with MPL:
Integrating Optimization in Real-World Business Applications, on Saturday, the day before
INFORMS Annual Conference,
Saturday November 12, 2005. The purpose of the workshop is to provide participants
with an insightful overview and give step-by-step instructions for successfully
building optimization applications.
The speakers for the workshop, Bjarni Kristjansson (Maximal Software, Inc.),
Alan King (IBM Research), Janos D. Pinter (PCS, Inc.), who all have decades of experience in this field, will take you through all
the steps of an optimization project using powerful optimization tools from Maximal Software, ILOG,
IBM, and Pinter Consulting.
Optimization technologies have become key tools in making important business decisions that increase
competitive advantage. Optimization, through the use of advanced mathematics and computer science
techniques, is used to assist organizations with solving their complex business problems in areas such
as manufacturing, distribution, finance, and scheduling. The success of optimization projects depends
on many different factors such as which modeling tools are used, integration with corporate data, and
the selection of the most efficient solution algorithms available for the problem.
The workshop is designed to give the participants hands-on experience with industrial-strength
optimization and modeling tools including:
- MPL Modeling System and OptiMax 2000 from Maximal Software
- CPLEX mixed integer solver from ILOG
- COIN/SMI stochastic solver from the COIN Open Source Project
- LGO global nonlinear solver from Pinter Consulting
- Microsoft ACCESS and EXCEL
The purpose of the workshop is to show how optimization models can be brought together into business applications. At the end of the workshop, the participants will be able to develop their own optimization models, link them to data sources and solve the models using state-of-the-art industrial-strength solvers. Participants will also acquire a good working knowledge on how to embed optimization models into applications.
By attending this workshop you will be able to:
- Identify the best use of optimization techniques and how to deploy them.
- Prepare and import data from different sources for optimization
- Identify which solvers works best for your optimization problem
- Gain a realistic view on the use of optimization for business applications
- Build your own embedded optimization applications.
Attendance is free for registered conference participants but the cost for non-conference
participants is $750. To register, please check the appropriate box on the INFORMS registration
form or send an email to firstname.lastname@example.org.
Please register early as seating is limited and is on a first-come first-serve basis.
Early Morning Sessions: Optimization Modeling in MPL
8:40 - 9:00 Registration, software installation and coffee
9:00 - 9:50 Introduction to Optimization Modeling and Solving (Bjarni Kristjansson)
- Current state-of-the-art in optimization and modeling
- Introduction to the MPL Modeling System
- Fundamental modeling techniques and model development
- Formulating models with the MPL Modeling Language
- Solving models with CPLEX and CONOPT
- Tuning solvers for better performance
9:50 - 10:40 Advanced Modeling Techniques (Bjarni Kristjansson)
- Tricks of the trade: Formulate MIP models for efficient solution times
- Formulating logical conditions with binary variables
- Using preprocessing to improve performance on large and difficult models
10:40 - 11:00 Coffee break
Late Morning Sessions: Connecting and Deploying Optimization Models
11:00 - 11:30 Connecting to Databases and Spreadsheets (Bjarni Kristjansson)
- Advanced data modeling techniques
- Connecting MPL to Databases
- Connecting MPL to Excel spreadsheets
- Scalability and speed issues for models with very large data sets
- Advanced indexing techniques for sparse data
11:30 - 12:00 Embedding and Deploying Optimization in Applications (Bjarni Kristjansson)
- Tools and methods for embedding optimization in applications
- Demonstration of optimization applications
- Introduction to OptiMax 2000 Component Library with Excel and Visual Basic
- Creating your own embedded optimization application step-by-step
- Optimization on the Internet and Web services
12:00 - 13:00 Lunch break
Early Afternoon Sessions: Stochastic Programming
13:00 - 14:00 Incorporating Uncertainty into Mathematical Programs (Dr. Alan King)
- Stochastic Programming: Current state-of-the-art
- Practitioner Perspective: Resource Allocation Problem
- Robust Optimization in the style of Bertsimas and Sim
- Stochastic Programming with Recourse
- Real Options
- COIN/SMI - Stochastic Modeling Interface
14:00 - 14:50 Formulating Stochastic Models in MPL (Bjarni Kristjansson)
- Representing Stochastic Models
- Scenario-based Recourse Problems
- Examples of Stochastic Models in MPL
- Solving Stochastic Problems
14:50 - 15:10 Coffee break
Late Afternoon Sessions: Global Optimization
15:10 - 16:10 Global Optimization with MPL/LGO (Dr. Janos D. Pinter)
- Global Optimization: A Brief Review of the State-of-art
- The LGO solver suite
- Model development and solution using MPL/LGO
- Modeling tips and solver tuning for better performance
16:10 - 17:00 Formulating and Solving Global Optimization Models (Bjarni Kristjansson)
- Formulating Global Optimization Problems
- MPL/LGO Implementation
- Solving Global Optimization Problems with LGO
Bjarni Kristjansson, Maximal Software, Inc.
Bjarni Kristjansson is the president and founder of Maximal Software, Inc., a leading vendor of
optimization modeling software. Maximal Software is the developer of the MPL Modeling System;
an algebraic modeling language that allows the user to formulate complicated optimization models
quickly and easily. Maximal Software also provides the OptiMax 2000 Component Library, which
allows MPL models to be easily embedded into end-user applications. Bjarni Kristjansson has many
years of practical experience providing innovative solutions, especially in regards to integrating
optimization with wide range of information technologies and business applications.
Dr. Alan King, IBM Research
Alan King is a research staff member with the Mathematical Sciences department at IBM's
Thomas J Watson Research Center, which he joined in 1988. His research and development
activity focusses on optimization technologies for decisionmaking under uncertainty.
He has published many papers in scientific journals and has held appointments on the
editorial boards of Mathematics of Operations Research and SIAM Journal of Optimization.
He was the lead developer for the software product OSL Stochastic Extensions, and is
currently lead developer for the Stochastic Modeling Interface in the COIN-OR
open source project.
Alan joined IBM's Mathematics and Analytics Computation Center on Maiden Lane, NYC
in 1996-7, and the Research VP for Strategy technical staff in 1998-9. In the period
2000-2002, Alan lead a small team investigating algorithms and models for resource
management problems in utility computing. In 2003-4, Alan pursued research projects
applying optimization models to the pricing of contingent claims and taught a semester
at Columbia University. Since 2004, Alan has participated in the Analytics Infrastructure
Solutions high performance computation infrastructure team taking various roles as
client architect, developer, and technical consultant. Alan is also Research's campus
relationship manager for the University of Washington.
Dr. Janos D. Pinter, PCS Inc.
Janos Pinter (MSc, Budapest; PhD, Moscow; DSc, Budapest) is the founder and owner of
Pinter Consulting Services Inc. Dr. Pinter is a recognized technical author, algorithm
and software developer. His primary area of research and development is nonlinear
(global and local) optimization. He is the author of Global Optimization in Action
(winner of the 2000 INFORMS Computing Society Prize) as well as of numerous other
publications. He has over 15 years of experience in the practice of global optimization,
and he often presents lectures, workshops and short courses on the subject. For more
details, please visit www.pinterconsulting.com.