W3C home > Mailing lists > Public > www-rdf-interest@w3.org > February 2003

RDF(S) in Railway Applications ? reasonable? feasible?

From: Roman Bischoff <romanix@netscape.net>
Date: Fri, 07 Feb 2003 07:04:48 -0500
To: www-rdf-interest@w3.org
Message-ID: <7FC8111E.60FFA7B0.0021BFBA@netscape.net>


My question is about using RDF(S),DAML, OIL, in railway infrastructure control and planning applications.

The context:

A european national railway company needs to renew some of their IT systems.
They need to share data ((resource, operation, scheduling) and functionality across divisions and companies. There are many interdependencies between different kinds of resources and rules.

Resource  examples: locomotives, wagons, railtrack segments, signals, people, points,... 
Agent     examples: signal, control systems, engineer, switches, displays, ...
statement examples: commandments, instructions, commands, rules, 


1.The basic clock of the railway-backbone main lines is 60 minutes.
2.We define the biggest railway station, (Vertex with highest degree, highest throughput)  (e.g. hub-1 in City-A) as starting point to distribute the basic clock.
3.Intercity trains in direction of city-B leave hub-1 every full hour (6:00, 7:00)
4.Intercity trains in direction of city-B leave hub-1 on departure platform No. 8.

My observation is:
The various kinds of timetables and schedules are results of (different types of) statements made by planners. 


There is a multitude complex conflicts / conditions:
1. fleet and train personnel circulation conflicts
2. train connection conflicts (e.g.
   train X, has to assure the connection with train Y, in city-Z, 
   If train Y, is more than 15min late but train X is not late, 
   the connection needs to be broken up)
3. sequence conflicts (train X has to enter into the station before train W)
4. train service conflicts (trains must not cross on single tracks)
This are again statements.

Some railway specialists (e.g. a mathematician) are convinced that many conflicts cannot be solved based on algoritms). Another important aspects is that the time to make a conflict-resolution decision is very short.

There are lot's of different kinds of statements which build the basis for the schedules and operational decisions.
So I thought, maybe one could use RDF triples to formulate such statements.
Then have a kind of version controlling, and maybe inference engine on them.
The schedules then could be created based on the statements.

Now I wonder whether and to what extend an approach based on URI, RDF(S), DAML, OIL could help to handle the complexity and interdepencies.
My initial intention was that a "resource- and statement-centric" approach for resource-allocation (reservation and usage) could be based on SemanticWeb techniques.

After browsing through "www-rdf-interest" I got the impression that the SemanticWeb concepts and constructs where general enough for being used in such situations.
But after reading some discussions (threads like  "XML Schema vs DAML/RDF/RDFS") I have the impression that in practice it's (today) better not to bet on these technologies for use outside Internet/SemanticWeb.

What is your opinion?

b) Is it worth to consider DAML/OIL/RDF/RDFS in more detail for applications of this kind?

c) What issues could be addressed with which technology (RDF(S), DAML, OIL)?

d) What is your experience with using URIs for real world physical objects like railway track segments (edges in a graph) or locomotives ?

Any opinions are welcome.


The NEW Netscape 7.0 browser is now available. Upgrade now! http://channels.netscape.com/ns/browsers/download.jsp 

Get your own FREE, personal Netscape Mail account today at http://webmail.netscape.com/
Received on Friday, 7 February 2003 07:27:31 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:07:44 UTC