Übersicht
| Kursnr. | 9T350DE | Wissensstand | Basic |
|---|---|---|---|
| Dauer | 5.0 Tage | Kursform | Klassenraum
(Praxis-Training) |
| Kurstyp | Offen/Privater Onsite | ||
| Kurspreis |
EUR 3.000,00 exkl. USt.
EUR 3.570,00 inkl. USt. |
This course teaches students how to use the CPLEX Optimizers that come with the IBM ILOG CPLEX Optimization Studio to model and solve problems expressed with linear constraints (linear and mixed-integer programming problems) or certain nonlinear constraints (quadratic programming problems).
The CPLEX mathematical programming engine delivers the power to solve large-scale, real-world optimization problems, the reliability necessary for mission-critical decision making, and the speed required in today's interactive applications. It also comes with fast and,reliable implementations of the fundamental algorithms used for solving demanding mathematical optimization problems.
The course is split between core and optional units that allow for a more customized classroom experience by letting students request additional material on areas of specific relevance to their work. A combination of instructor-led presentations and lab exercises enables students to learn the concepts, architecture, components, processes, and procedures necessary to model and solve a variety of problem types using CPLEX. This course begins with an overview of linear programming (LP), and subsequent units teach students how to create and solve LP models using ILOG Concert Technology. Students also learn about mixed-integer programming (MIP), quadratic programming (QP), and debugging. The optional units that are available for students cover topics such as CPLEX performance tuning, working with LP callbacks, preprocessing, columnwise modeling, and parallel CPLEX.
For information on other related WebSphere courses, visit the WebSphere Education Training Paths website:
Zielgruppe
This course is designed for operations research practitioners who are new to CPLEX, or who have some experience with CPLEX and want to learn more about its functionality.
Voraussetzungen
Students should have:
- a working knowledge of C++ or Java
No experience in linear programming is necessary, but a background in mathematical programming is helpful.
Kursziele
- Use CPLEX to model and solve a variety of mathematical programming (MP) problem types
- Modify parameters to improve the quality of solutions
- Debug models and perform post-optimization analysis
- Experiment with a range of advanced CPLEX features
Themenübersicht
Course introduction
Duration: 15 minutes
Unit 1. Introduction
Duration: 30 minutes
Learning objectives:
- Identify scenarios in which optimization is useful
- Describe IBM ILOG CPLEX Optimization Studio and its components and functional groups
Unit 2. Linear programming
Duration: 1 hour
Learning objectives:
- Describe the basic concepts of linear programming (LP)
- Define a sample problem statement
- Formulate a simple LP model
- Describe the basic concepts of the Simplex algorithm
Unit 3. Modeling LPs with Concert
Duration: 1 hour 30 minutes
Learning objectives:
- Identify the main components of IBM ILOG Concert Technology
- Describe how to use Concert as a modeling technology
- Declare and populate a model using Concert
- Explain how Concert interacts with CPLEX
Unit 4. Solving LPs with Concert
Duration: 1 hour 30 minutes
Learning objectives:
- Identify the main algorithm classes in CPLEX and Concert
- Describe how CPLEX optimizes a model
- Extract a model, access the solution, and display the results
Unit 5. CPLEX LP algorithms
Duration: 1 hour
Learning objectives:
- Identify the various LP optimizers available to CPLEX
- Compare the optimizers and select the appropriate one for a given problem type
- Change the optimizer used by CPLEX
Unit 6. CPLEX parameters
Duration: 45 minutes
Learning objectives:
- Describe CPLEX parameter types and categories
- Use and modify CPLEX parameters
Unit 7. Post optimality analysis
Duration: 1 hour
Learning objectives:
- Identify the reasons to perform post optimality analysis
- Describe how shadow price and reduced costs are calculated
- Perform sensitivity analysis on a model
Unit 8. Debugging
Duration: 1 hour
Learning objectives:
- Manage and interpret CPLEX log files
- Export a model in a variety of different file formats
- Use the CPLEX Interactive Optimizer to examine a solution
Unit 9. Analyzing infeasibilities
Duration: 1 hour 30 minutes
Learning objectives:
- Analyze and resolve infeasibilities using the Conflict Refiner and FeasOpt
- Identify a "minimal conflict" within a model
- Explain the advantages and disadvantages of two approaches to analyzing infeasibilities
Unit 10. Mixed integer programming
Duration: 2 hours
Learning objectives:
- Describe the uses of mixed integer programming (MIP) and compare MIP with linear programming
- Define the linear relaxation of a mixed integer program
- Describe how CPLEX uses cuts when finding solutions
- Explain the use of the branch and bound algorithm
Unit 11. Solving MIPs with Concert
Duration: 2 hours
Learning objectives:
- Describe how Concert recognizes a MIP problem and adapts accordingly
- Access and interpret the node log
- Convert the type of variables in a model
- Manage the CPLEX branch and bound
- Use the CPLEX solution pool
Unit 12. Quadratic programming
Duration: 2 hours
Learning objectives:
- Identify the nonlinear problems CPLEX can solve
- Add a quadratic objective in Concert
- Formulate a mixed integer quadratic program
- Define a quadratically constrained program
Optional units
Optional Unit 1. CPLEX performance tuning
Duration: 1 hour 30 minutes
Learning objectives:
- Tune a model for speed
- Tune a model to deal with memory considerations
- Tune a model for numerical considerations
Optional Unit 2. Concert modeling constructs
Duration: 30 minutes
Learning objectives:
- Define a piecewise linear function in Concert
- Identify use cases for semi-continuous variables and semi-integer variables
- Explain the use of special ordered sets
Optional Unit 3. LP query callbacks
Duration: 45 minutes
Learning objectives:
- Describe the use of LP callbacks in CPLEX
- Use callbacks to control an LP optimization
Optional Unit 4. MIP callbacks
Duration: 30 minutes
Learning objectives:
- Describe the use of MIP callbacks in CPLEX
- Distinguish between informational, query, and control callbacks
Optional Unit 5. Lazy constraints and user cuts
Duration: 30 minutes
Learning objectives:
- Describe the use of lazy constraints and add them to a model
- Describe the use of user cut pools and add cuts to a model
- Compare the relative advantages of user cuts and lazy constraints
Optional Unit 6. Goals
Duration: 45 minutes
Learning objectives:
- Define the uses of goals in Concert
- Distinguish between the various types of goals available
Optional Unit 7. Preprocessing
Duration: 45 minutes
Learning objectives:
- Use and control CPLEX preprocessing features
- Describe the problem modifications that preprocessing can make
Optional Unit 8. Columnwise modeling
Duration: 1 hour
Learning objectives:
- Define columnwise modeling
- Set up a problem by columns
Optional Unit 9. Column generation
Duration: 1 hour
Learning objectives:
- Explain why column generation is useful for solving large-scale problems
- Apply the column generation method
Optional Unit 10. Parallel CPLEX
Duration: 30 minutes
Learning objectives:
- Explain the options available for Parallel CPLEX
- Identify practical considerations in the use of parallel optimizers
Optional Unit 11. CPLEX MIP cuts
Duration: 1 hour
Learning objectives:
- Define and give examples of the various types of MIP cuts that CPLEX automatically derives
Hinweise
The unit and exercise durations listed below are estimates, and may not reflect every class experience. If the course is customized or abbreviated, the duration of unchanged units will probably increase.
This course is an update of the following previous course:
- Mathematical Programming with IBM ILOG CPLEX V12 (WB349)