Self-organized train dispatching

From Master Projects
Revision as of 11:06, 27 November 2008 by Schut (talk | contribs) (New page: {{Masterproject |Master name=Computational Intelligence and Selforganisation |Student name=Sander Bouwhuis |Student number=0000000 |Project start date=2008/03/01 |Project end date=2008/02/...)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Self-organized train dispatching
status: ongoing
Master: project within::Computational Intelligence and Selforganisation
Student name: student name::Sander Bouwhuis
number: student number::0000000
Start start date:=2008/03/01
End end date:=2008/02/01
Supervisor: Martijn Schut
Second reader: has second reader::Gusz Eiben
Company: has company::NS Spoorwegen
Poster: has poster::Media:Media:Posternaam.pdf

Signature supervisor



While a good timetable in theory allows trains to move without being interrupted, in practice this hardly ever happens. Problems arise due to a variety of reasons, such as accidents or malfunctioning rolling stock. These incidents are very hard to predict, and while their direct effect is usually quite manageable, they are also the leading cause of so-called knock-on or secondary delays. Secondary delays happen when a train cannot reach its destination on time, affecting another train when for instance a train driver cannot make it to his next scheduled service in time. This causes the second train to be delayed because of the first one.

In my master project at the Dutch Railways (NS), I investigate the application of self organization in this field. Agents representing trains and stations will work out a solution to conflicting situations based on their local information, instead of relying on central dispatchers who are often poorly equipped for handling short and urgent delays. The objective is to minimize the effect of secondary delays in the system by being able to respond to smaller and therefore more manageable delays in a timely fashion.