Examination arrangement

Course content

The course is an introduction to stochastic optimization. The course will cover the following topics:

• Motivation for stochastic optimization: why does uncertainty matter?
• Different modeling approaches, with a particular focus on recourse models
• Theoretical properties of recourse models
• Solution algorithms, among which: Benders' decomposition (L-shaped), stochastic dual dynamic programming (SDDP), and dual decomposition
• Scenario generation
• Applications of stochastic optimization (with a focus on energy)

Learning outcome

Position and function within the study programme:

The course is designed for PhD students of the Department of Industrial Economics and Technology Management (IØT) and other departments who work with theoretical and practical optimization problems in different branches of industry and services with substantial uncertainty about problem data and other elements of problem formulation. The course is built upon optimization courses in IØT's master programme and knowledge of probability theory.

The course will convey the following knowledge: The theoretical foundation necessary for formulation, analysis and solution of stochastic programming problems and relevant applications; The knowledge necessary to conduct research in the field of optimization under uncertainty.

The course will develop the following skills: modeling and solving practical problems as a stochastic optimization model.

Other important learning objectives: Recognize when explicit modeling of uncertainty is important; Validating models by using stability tests.

Learning methods and activities

Lectures and non-obligatory exercises. The course can be given in form of intensive lectures with several hours per day, several days per week, during a limited number of weeks in the semester.

Recommended previous knowledge

Knowledge of linear and nonlinear optimization is essential. Such knowledge can be obtained through courses TIØ4120 Operations Research, Introduction, TIØ4126 Optimization and Decision Support for Industrial Business Planning, or TIØ4130 Optimization Methods with Applications, or similar.

Required previous knowledge

Master of Science in Industrial Economics and Technology Management, or similar.

Course materials

Given at the beginning of the semester.

Credit reductions