D4.2Housing Location Model

Activity-based models of travel demand require a socio-demographic profile of individuals as input. The location of the house is a key variable as it is one of the pegs influencing the action space of individuals. These variables are constructed by creating a synthetic population. Current modelling approaches repeat this process on an annual basis, jeopardising the integrity of the data. This report describes an approach to creating a dynamic synthetic population that takesboth demographic processes andresidential choice behaviour into account.Residential choice behaviour is based on a housing model which first predictsthe intention to move house as a function of socio-demographics, housing attributes, key characteristics of activity-travel patterns, and the elapsed time since the previous move, adjusted according to empirical information about the relationship between the stated intention to move and actual moves.Monte Carlo simulationis then used to simulate actual housing choice of those intendingto move, calibrated on empirical distributions taking their housing history into account.The approach is illustrated for the City of Rotterdam, The NetherlandsThe model specification is limited due to publicly available information but can be extended if more data would be available.