Secondary language: |
English |
Secondary title: |
Mode choice model parameters estimation |
Secondary abstract: |
The present work focuses on parameter estimation of two mode choice models: multinomial
logit and EVA 2 model, where four different modes and five different trip purposes are taken
into account.
Mode choice model discusses the behavioral aspect of mode choice making and enables its
application to a traffic model. Mode choice model includes mode choice affecting trip factors
by using each mode and their relative importance to choice made. When trip factor values are
known, it enables calculation of trip mode shares.
In order to estimate mode choice model parameters input data are needed, which can be
obtained with a stated preference survey. Using maximum likelihood method, parameter
estimation of mode choice model using those data is performed.
Multinomial logit model consists of mode utility functions, which consist of mode specific
constants, mode choice affecting factors and their coefficients. Estimation of one parameter is
needed for each factor for each trip purpose, besides that, estimation of mode specific
constants is needed too.
Mode choice model EVA 2 consists of non-linear parameter utility functions in which
estimation of three parameters for each mode choice affecting factor for each trip purpose is
needed. |
Secondary keywords: |
graduation thesis;geodesy;mode coice model;stated preference survey;maximum likelihood method;utility functions; |
File type: |
application/pdf |
Type (COBISS): |
Undergraduate thesis |
Thesis comment: |
Univ. Ljubljana, Fak. za gradbeništvo in geodezijo |
Pages: |
XVI, 84 str., pril. |
Type (ePrints): |
thesis |
Title (ePrints): |
Mode choice model parameters estimation |
Keywords (ePrints): |
model izbire prometnega sredstva;anketa izraţene preference;metoda največjega verjetja;uporabnostne funkcije |
Keywords (ePrints, secondary language): |
mode choice model;stated preference survey;maximum likelihood method;utility functions |
Abstract (ePrints): |
V diplomski nalogi je predstavljena ocena parametrov dveh modelov izbire prometnega sredstva in sicer multimodalnega logit in EVA 2 modela. Pri tem so obravnavana štiri različna prometna sredstva in pet različnih namenov potovanja. Model izbire prometnega sredstva obravnava vedenjski vidik izbire prometnega sredstva in omogoča vgradnjo le-tega v model za napoved prometa. Model izbire prometnega sredstva zajema faktorje potovanj s posameznimi prometnimi sredstvi, ki vplivajo na izbiro prometnega sredstva in relativno pomembnost teh faktorjev za izbiro. Na podlagi znanih vrednosti faktorjev potovanj tako omogoča izračun deleţa potovanj, opravljenih s posameznim prometnim sredstvom. Za oceno parametrov modela izbire prometnega sredstva je potrebna baza podatkov, ki jo lahko pridobimo z izvedbo ustrezne ankete izraţene preference. Na podlagi teh podatkov z uporabo metode največjega verjetja ocenimo parametre modela izbire prometnega sredstva. Multimodalni logit model sestavljajo uporabnostne funkcije posameznega prometnega sredstva. Sestavljene so iz specifičnih konstant prometnih sredstev, faktorjev, ki vplivajo na izbiro prometnega sredstva, in njihovih koeficientov. Za vsak faktor je torej določen en parameter za vsak namen potovanja, poleg tega pa so določene še specifične konstante prometnih sredstev. Model izbire prometnega sredstva EVA 2 sestavljajo nelinearne uporabnostne funkcije posameznih faktorjev, ki vplivajo na izbiro prometnega sredstva. Funkcijski zapis zahteva za vsak faktor oceno treh parametrov za vsak namen potovanja. |
Abstract (ePrints, secondary language): |
The present work focuses on parameter estimation of two mode choice models: multinomial
logit and EVA 2 model, where four different modes and five different trip purposes are taken
into account.
Mode choice model discusses the behavioral aspect of mode choice making and enables its
application to a traffic model. Mode choice model includes mode choice affecting trip factors
by using each mode and their relative importance to choice made. When trip factor values are
known, it enables calculation of trip mode shares.
In order to estimate mode choice model parameters input data are needed, which can be
obtained with a stated preference survey. Using maximum likelihood method, parameter
estimation of mode choice model using those data is performed.
Multinomial logit model consists of mode utility functions, which consist of mode specific
constants, mode choice affecting factors and their coefficients. Estimation of one parameter is
needed for each factor for each trip purpose, besides that, estimation of mode specific
constants is needed too.
Mode choice model EVA 2 consists of non-linear parameter utility functions in which
estimation of three parameters for each mode choice affecting factor for each trip purpose is
needed. |
Keywords (ePrints, secondary language): |
mode choice model;stated preference survey;maximum likelihood method;utility functions |
ID: |
8310892 |