MS-E2147: Optimization in Public Transport
| Course name | Optimization in Public Transport |
|---|---|
| Course code | MS-E2147 |
| Abbreviation | “(Jokingly called the HSL-course)” |
| Period | I-II |
| Lecturer | Philine Schiewe |
Description
Optimization topics in the context of problems in public transportation. Highly specialized optimization course which builds on the topics of linear optimization, combinatorial optimization and graph theory concepts.
Course material
Official material
Lecture notes and slides provided by the professor. Quality is pretty good.
Extra material
Contents and workload
The course requires a bit more specialized understanding and surprisingly a lot more graph theoretical understanding than expected from an optimization (or master level OR) course.
The projects are a lot more like actual code program development than other university level projects as you take dataframes as inputs and you have to output certain output files as dataframes. The project really motivates the students to work on their own solution approaches. These problems follow the framework presented throughout the course (and discussed first fully during week 1).
Overall workload
Weekly contents
| Week | Topics |
|---|---|
| 1 | Infrastructure networks and passenger assignment - Soft introduction to the whole setting of the problems which will be considered during the course. |
| 2 | Background on network optimization - This may be given as a self-study. If you have not taken courses like MS-E2145: Combinatorial Optimization this may require a lot more attention. Understanding NP-completeness and reduction proofs are relatively important considering some theory. Do take time for this. |
| 3 | Line Planning: Feasibility and cost model - Considers how public transport models can be built with feasibility and different cost-based objective functions. |
| 4 | Line Planning: Direct travelers and travel time - Considers how public transport models can be built with feasibility and traveler “ease” as an objective function. |
| 5 | Event-activity networks and passenger routing - A new modeling method for building passenger behaviour connected to how travelling is built on existing public transport lines. |
| 6 | Aperiodic timetabling - Timetabling with a single set time period. |
| 7 | Periodic timetabling - Timetabling with single length time period with a possibility of having events “overflow” between different time period instances. |
| 8 | Periodic timetabling: Cycle base formulation - Graph-approach formulation of timetabling building on the concepts considered within previous weeks. |
| 9 | Periodic timetabling: Algorithmic approaches - Self-evident |
| 10 | Integrated timetabling and passenger routing - Attempting to formulate passenger routing within an existing timetable scheme |
| 11 | Vehicle scheduling - Utilizing existing public transport lines to build event-activities which optimize public transport network traversal. |
| 12 | Projects are presented. |
Practicalities
The final grade is given by a combination of points from the exercise sheets (up to 45 points), presentation of solutions in the exercise sessions (up to 10 points) and the project (up to 45 points, including its presentation).
Point thresholds for grades follow the generic 50 points for a grade 1 with 10 point intervals until 90 points for a grade 5. Some bonus points are available from each category.
Related courses
Official prerequisites
MS-C2110 Introduction to Operations Research, MS-E2121: Linear Optimization.
Additional prerequisites
- Recommended: Good understanding of programming for the project and setting up an environment (VS code with interpreter for example) for programming may be needed.
- Helpful: MS-E2145: Combinatorial Optimization. If MS-E2145 is not a possibility, some fundamental understanding of graphs and proofs on graphs, for example see MS-A040X Foundations of Discrete Mathematics or any other course on graph theory or network theory.
More like this
See possible research topics under Prof. Schiewe.