Zum Inhalt springen
Technische Universität München

Operations Research in Hydrology ()

Vortragende/r (Mitwirkende/r)
Faizan Anwar [L]
Faizan Anwar [L]
Nummer 0000001348
Art Vorlesung mit integrierten Übungen
Umfang 2 SWS
Semester Sommersemester 2025
Unterrichtssprache Englisch
Stellung in Studienplänen Siehe TUMonline
Termine Siehe TUMonline

Teilnahmekriterien

Siehe TUMonline

Lernziele

1. understand the problem of unknown model parameters
2. understand what optimal model parameters are
3. efficiently traverse the model parameter space
4. develop computer code to search for an optimal set
5. identify model parameter over-fitting
6. use their search results to run models to simulate various best- and worst-case scenarios such as those in the future depending on climate change
7. give recommendations for future strategies based on their simulations
8. present and defend their calculations and findings against the questions of an expert audience

Beschreibung

1. Introduction to Operations Research in Hydrology
2. Various terms used, and their definitions
3. Brief introduction to conceptual rainfall-runoff modeling and a few models
4. Problems that need specialized algorithms to find optimal solutions
5. Global optimization
6. Optimization with constraints
7. The Differential Evolution optimization algorithm
8. The simulated Annealing optimization algorithm
9. Tricks to achieve results faster
10. Model parameter over-fitting
11. Determining parameter bounds for a global search
12. Diagnostics to verify the progress and validity of an optimization run
13. Scenario simulations using various model parameter vectors

Inhaltliche Voraussetzungen

• ED130088 Computer Programming in Hydrology

Lehr- und Lernmethoden

The module consists of lectures with integrated exercises. The students are required to bring their own laptops. Portable open-source software will be made available to all. The students should be able to code in the Python programming language. After the introduction and definition of essential terms, the problem of optimization is broken in to small chunks that the students are supposed to complete in each exercise. The aim is that the students are able to find a good set of model parameters that fulfill a given criteria using predefined models, and input data using the learned optimization schemes. Finally, the students write a report (as a group) where they demonstrate their ability to understand a given situation and provide the solution in the form of a running computer code that takes input data, fits a model to it by using the either the Differential Evolution and/or the Simulated Annealing algorithm, analyzes the resulting model parameters and outputs, and summarize them in the form of statistics and graphs.

Studien-, Prüfungsleistung

The examination is a project done in groups of up to three students. Through a project report (80% of the grade, ca. 25 pages), the students prove their ability to understand and model a given problem in hydrology (e.g., find a set of model parameters that fulfill some given criteria and then use it to make future predictions). They will analyze the problem and provide a solution using methods and tools that they learned in this course. In the project report, it should be possible to clearly identify the specific contribution of each group member. Through a final presentation (20% of the grade, individual group member evaluation, up to 60 minutes) they demonstrate their ability to present and defend their findings based on critical questions by an expert audience. The final grade depends on how detailed their analysis and solution is and the quality of its presentation.

Empfohlene Literatur

Advances in differential evolution, Chakraborty 2008

Links

To top