This dissertation consists of seven papers. InTime and income constraints in discrete choice models withan application to mode choice, we discuss how variouseffects of time and income constraints can be included indiscrete choice models. Several cases of indirect utilityfunctions are estimated and compared, showing that the obtainedvalues of time and elasticities can be very sensitive to modelspecification.
Three papers deal with the development of TILT, Tool forIntegrated analysis of Location and Transport, an integratedland use-transportation model for Stockholm. In the first,The influence of accessibility on residential location,we develop a theoretical framework for modeling thesimultaneous choice of location and travel pattern, showing howthe joint choice of location and average travel pattern can bemodeled. With focus on the location choice, we presentmethodology and results for an estimation of the framework. Inthe second,Integrated travel pattern modeling, the choice of travelpattern in is explained in detail, and estimation methodologyand results are presented.
In the third,A model for integrated analysis of household location andtravel choices(co-authored with Lars-Göran Mattsson),a different derivation of the framework is presented, with amore rigorous treatment of the stochasticity in the travelpattern. Using numerical simulation, we investigate theresulting travel demand functions.
InLocation and transport effects of road pricing - asimulation approach(co-authored with Lars-GöranMattsson), we develop a model of a generic city, usingparameter values from models estimated on real data of locationand travel. The model is used to investigate the effects ontransport and location of a road pricing reform and of a tollring around the city center.
Two papers deal with road pricing in the presence ofunobserved heterogeneity in user preferences. In the first,Road pricing with limited information and heterogeneoususers: a successful case, we show that if all users havethe same choice set and all alternatives have the same monetarycost, any toll reform that reduces aggregate travel time andredistributes the toll revenues equally to all users makeseveryone better off. In the second,The use of average values of time in road pricing - a noteon a common misconception, it is argued that the difficultyto convert time delays into monetary fees when users areheterogeneous has been underestimated or neglected, and thatthere is a common misconception that it is possible to use anaggregate time value for this conversion.