The following information was submitted:
Transactions: INTERNATIONAL JOURNAL of MATHEMATICAL MODELS AND METHODS IN APPLIED SCIENCES
Transactions ID Number: 19-905
Full Name: Irena Strnad
Position: Researcher
Age: ON
Sex: Female
Address: University of Ljubljana, Faculty of Civil and Geodetic Engineering, Traffic Technical Institute, Jamova cesta 2, SI-1000 Ljubljana
Country: SLOVENIA
Tel:
Tel prefix:
Fax:
E-mail address: irena.strnad@fgg.uni-lj.si
Other E-mails:
Title of the Paper: Genetic algorithms application to EVA mode choice model parameters estimation
Authors as they appear in the Paper: Irena Strnad, Marijan Žura
Email addresses of all the authors: irena.strnad@fgg.uni-lj.si,marijan.zura@fgg.uni-lj.si
Number of paper pages: 9
Abstract: This paper presents parameters estimation of EVA (EVA â€" German abbreviation for Erzeugung, Verteilung and Aufteilung meaning Production, Distribution, and Mode Choice) mode choice model of city of Ljubljana, Slovenia by using genetic algorithms software. First we present design of stated preference survey, then we briefly review EVA mode choice model, present different types of utility functions, Maximum likelihood method as the estimation method and application of genetic algorithms software. Probabilities of choosing each of four considered modes (private car, public transport, bike, walking) can be calculated by using estimated mode choice model parameters. A practical example of mode choice probabilities for an actual trip is shown at the end. Final log-likelihood enables comparison among different types of utility functions. Results show that absolute differences in final log-likelihood among most types of utility functions are not high in spite of differenc!
es in function shapes, which implies that different functions may best describe different variables. Log-likelihood function for most utility function types by using standard optimization tool only convergated to local maximum, what clearly states the need to use genetic algorithms software to find the best solution.
Keywords: Genetic algorithms, Maximum likelihood method, Mode choice model, Stated preference survey, Utility function
EXTENSION of the file: .doc
Special (Invited) Session: EVA mode choice model parameters estimation
Organizer of the Session: 202-274
How Did you learn about congress:
IP ADDRESS: 86.61.126.1
