Dvig učinkovitosti izvedbe javnega prevoza potnikov z uporabo algoritma za zaznavo sprememb v povpraševanju : doktorska disertacija

Hojski, Danijel

| | SLO | ENG | Cookies and privacy

Bigger font | Smaller font

First page > Show document

Show document

Title:	Dvig učinkovitosti izvedbe javnega prevoza potnikov z uporabo algoritma za zaznavo sprememb v povpraševanju : doktorska disertacija
Authors:	ID Hojski, Danijel (Author) ID Lep, Marjan (Mentor) More about this mentor... ID Erveš, Rija (Comentor)
Files:	DOK_Danijel_Hojski_2025.pdf (3,19 MB) MD5: 17EC62FB7F4748FAAD075AB5FD58A5BE
Language:	Slovenian
Work type:	Doctoral dissertation
Typology:	2.08 - Doctoral Dissertation
Organization:	FGPA - Faculty of Civil Engineering, Transportation Engineering and Architecture
Abstract:	Tradicionalni pristop k načrtovanju javnega potniškega prometa (JPP) temelji na dolgoročnem zbiranju podatkov in oblikovanju stalnih voznih redov. V praksi pa se pogosto soočamo z neučinkovitostjo: preveliko kapaciteto na podeželju in izven konic ter pomanjkanje zmogljivosti ob nenadnih povečanih potrebah. Ključni izziv je pravočasno odzivanje na kratkoročne (dnevne) in srednjeročne (tedenske, mesečne) spremembe v povpraševanju. Na to lahko vplivajo izredni dogodki in tarifne politike, kot je uvedba brezplačnih vozovnic za določene populacije. Takšen primer se je zgodil leta 2020 z uvedbo brezplačnega JPP za vse osebe, starejše od 65 let. Ukrep je povzročil nenaden in obsežen porast uporabe JPP ter postavil pred prometne planerje izziv spremljanja in odzivanja na novo oblikovane potovalne vzorce. Disertacija obravnava podatkovno usmerjen pristop, ki temelji na elektronskem sistemu IJPP za prodajo in validacijo vozovnic. Analizirani so bili podatki iz štirimesečnih obdobij v letih 2020, 2022 in 2023. V ospredju je bilo razumevanje vedenja starejše populacije ob možnosti brezplačnega potovanja ter sledenje trendom uvajanja in utrjevanja uporabe. Odziv na ukrep je bil hiter: že v prvem mesecu je brezplačno vozovnico pridobilo 30 % upravičencev (približno 250.000 upravičenih). Do konca leta 2023 je delež narasel na 60 %. Brezplačni potniki so predstavljali do polovico vseh validacij izven prometnih konic in do 20 % v konicah. Na tej podlagi smo razvili algoritem za napovedovanje popoldanskega ali večernega povpraševanja glede na jutranje validacije. Upoštevani so bili tudi zunanji dejavniki z dokazanim vplivom na mobilnost. Cilj je bila večja odzivnost in prilagodljivost JPP v realnem času. Verifikacija algoritma v retrospektivni analizi je v večini primerov izkazala nad 90 % natančnost napovedi povpraševanja, kar potrjuje hipotezo, da lahko podatki iz sistema e-ticketinga — tudi brez dodatnih nadgradenj, kompleksnih statističnih metod ali sledenja potnikov prek kartic — pomembno prispevajo k napovedovanju povpraševanja. Dognanja iz disertacije so pokazala, da lahko tehnologije ITS, kot je sodobni sistem e-ticketinga, kljub določenim omejitvam podatkov (zlasti pomanjkanju informacij o izstopih potnikov in kapacitetah vozil), učinkovito podpirajo naloge planiranja in upravljanja javnega prevoza. Algoritem odpira številne možnosti za nadgradnjo, zlasti z vključevanjem dodatnih, težje opredeljivih dejavnikov ter z bolj poglobljeno analizo različnih scenarijev. Nadaljnje urjenje algoritma z rekurzivno metodo strojnega učenja lahko izboljša njegovo prilagodljivost na spremembe v povpraševanju ter omogoči zanesljivejše napovedi za širši spekter scenarijev.
Keywords:	planiranje JPP, prometno povpraševanje, učinkovitost JPP, brezplačni JPP, e-ticketing
Place of publishing:	Maribor
Place of performance:	Maribor
Publisher:	[D. Hojski]
Year of publishing:	2025
Number of pages:	XVI, 141 str.
PID:	20.500.12556/DKUM-93024
UDC:	656.025.2:711.7(043.3)
COBISS.SI-ID:	259753987
Publication date in DKUM:	27.11.2025
Views:	138
Downloads:	39
Metadata:
Categories:	KTFMB - FG
:	Copy citation

Average score:	(0 votes)
Your score:	Voting is allowed only for logged in users.
Share:

Hover the mouse pointer over a document title to show the abstract or click on the title to get all document metadata.

Licences

License:	CC BY-NC-ND 4.0, Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International

Link:	http://creativecommons.org/licenses/by-nc-nd/4.0/
Description:	The most restrictive Creative Commons license. This only allows people to download and share the work for no commercial gain and for no other purposes.
Licensing start date:	02.06.2025

Secondary language

Language:	English
Title:	Increasing the efficiency of public transport service using the algorithm for prognosis of travel demand changes
Abstract:	The traditional approach to public transport (PT) planning relies on long-term data collection and the creation of fixed timetables. In practice, this often leads to inefficiencies: oversupply in rural areas and off-peak periods, and capacity shortages when demand suddenly increases. A key challenge is the ability to respond effectively to short-term (daily) and medium-term (weekly or monthly) changes in demand. These fluctuations can be triggered by extraordinary events or fare policy changes, such as the introduction of free travel for selected population groups. One such case occurred in 2020 with the introduction of free PT for all residents aged 65 and above. This policy led to an immediate and significant increase in PT use and presented transport planners with the challenge of monitoring and responding to newly emerging travel patterns. This dissertation takes a data-driven approach based on the IJPP electronic ticketing system, which records sales and validations of tickets. The study analyzed four-month periods in 2020, 2022, and 2023, focusing on understanding the travel behavior of the older population once they were granted free travel, and tracking usage patterns as they developed and stabilized. The response to the policy was rapid: within the first month, 30% of eligible users (around 250,000 pensioners in Slovenia, or about a quarter of the population) obtained a free pass. By the end of 2023, this figure rose to 60%. These users accounted for up to half of all off-peak validations and around 20% during peak periods. Based on these findings, an algorithm was developed to forecast afternoon or evening demand based on morning validations. External factors with a proven impact on mobility were also considered. The objective was to enable greater responsiveness and adaptability of the PT system in real time. The algorithm’s retrospective verification demonstrated over 90% accuracy in most cases, confirming the hypothesis that data from the e-ticketing system — even without additional upgrades, complex statistical methods, or tracking users via smart cards — can significantly contribute to demand forecasting. Findings from the dissertation show that ITS technologies, such as the modern e-ticketing system, despite certain limitations in IJPP data (especially the lack of information on passenger exits and vehicle capacities), can effectively support public transport planning and management tasks. The algorithm opens numerous possibilities for further upgrades, particularly through the inclusion of additional, less tangible factors and more in-depth scenario analysis. Further training of the algorithm using recursive machine learning methods could enhance its adaptability to demand fluctuations and enable more reliable predictions across a broader range of scenarios.
Keywords:	PT planning, travel demand, efficiency of PT, fare-free PT, e-ticketing

Comments

Leave comment

You must log in to leave a comment.

Comments (0)

0 - 0 / 0

There are no comments!

Back