Network optimization in railway transport planning

master's thesis

• Author: Matej Jusup
• Mentor: Marko Vrdoljak (mentor)
• Mentor: Andreas Dress (mentor); VIAF: 19759435
• Committee member: Marko Vrdoljak (predsjednik povjerenstva)
• Committee member: Vjekoslav Kovač (član povjerenstva)
• Committee member: Maja Starčević (član povjerenstva)
• Committee member: Miljenko Marušić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Science; (Department of Mathematics); Zagreb
• Defense date and country: 2017-02-24, Croatia
• Scientific / art field, discipline and subdiscipline: NATURAL SCIENCES; Mathematics

Abstract

This work is dealing with train timetabling problem. In the first chapter, one can find an introduction to network flows which is needed for understanding deeper concepts later on. Namely, basic graph theory definitions are stated as well as core problems like the minimum cost flow and shortest path problem. Furthermore, two equivalent representations of network flows are described, including some useful properties connected to each of them. At the end of the chapter, linear programming and simplex method are introduced into some detail. In the second chapter more complex theory is introduced. At the beginning, multi-commodity flow problem is stated and few solutions approaches are briefly described. Once we settled for one of them, the rest of the chapter is dealing with Lagrangian relaxation and column generation techniques. Since column generation is the main result needed for solving our problem, some finer results, like determining lower and upper bounds, are stated. In the last, third chapter, one can find a model for representing train timetabling problem for a single line network. That model was introduced by Valentina Cacchiani in her Ph.D. thesis. In this work, periodicity of timetable is assumed because it makes computations way quicker, as well as it has some other benefits. At the end, one can find an algorithm based on column generation technique for solving introduced model. That algorithm is based on 6 steps, and after reading this work, one should be able to fully understand each of them.

Abstract (croatian)

Ovaj rad bavi se problemom rasporeda vožnje u željezničkom prometu. U prvom poglavlju nalazi se uvod u mrežne tokove koji je potreban za razumijevanje naprednijih koncepata. Konkretno, iskazane su osnovne definicije teorije grafova kao i neki temeljni problemi poput problema najjeftinijeg toka i problema najkraćeg puta. Nadalje, opisana su dva ekvivalenta prikaza mrežnih tokova, uključujući neka korisna svojstva za svaki od njih. Na kraju poglavlja, linearno programiranje i simpleks metoda, objašnjeni su na razini razumijevanja. U drugom poglavlju nalazi se naprednija teorija koja se nadovezuje na prvo poglavlje. Na početku poglavlja prikazan je problem više dobara, kao i nekoliko pristupa rješavanju navedenog problema. Nakon što smo se odlučili za jedan od pristupa, ostatak poglavlja bavi se Lagrangeovom relaksacijom i metodom generacije stupaca. Kako je upravo metoda generacije stupaca najvažniji rezultat za rješavanje našega problema, napredniji rezultati vezani uz određivanje donjih i gornjih granica su detaljno objasnjeni. U posljednjem, trećem poglavlju, nalazi se model za prikazivanje problema rasporeda vožnje za mreže s jednom tračnicom. Navedeni model prvi puta je predstavljen u doktorskom radu Valentine Cacchiani. U ovom radu također pretpostavljamo periodičnost rasporeda vožnje kako bismo, između ostalih, ostvarili prednost poput bržeg vremena računanja. Na kraju rada nalazi se algoritam, temeljen na metodi generacije stupaca, za rješavanje predstavljenog modela. Navedeni algoritam sastoji se od 6 koraka, od kojih je svaki detaljno opisan u ovome radu.

• Keywords: train timetabling problem; network flows; minimum cost flow; shortest path problem; linear programming; simplex method; multi-commodity flow problem; Lagrangian relaxation; column generation techniques
• Keywords (croatian): problem rasporeda vožnje u željezničkom prometu; mrežni tokovi; problem najjeftinijeg toka; problem najkraćeg puta; linearno programiranje; simpleks metoda; problem više dobara; Lagrangeovom relaksacijom; metoda generacije stupaca
• Language: english
• URN:NBN: urn:nbn:hr:217:138780
• Study programme: Title: Mathematical Statistics; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra matematike (magistar/magistra matematike)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Open access
• Created on: 2018-02-27 13:01:58