Evolutionary Models of Migratory Systems. A Behavioural Approach with Application on Intraurban Mobility in Toronto CMA, 1966-1976

Abstract

<p>In this thesis a general and formal conceptualization, within any type of migration can be studied is proposed. This framework results from the merging of two different traditions of research. The first has its roots in statistical mechanics and is expressed by what is known as evolutionary models. In this respect we follow the master equation formulation of Weidlich and Haag(1983).</p> <p>The second tradition has its roots in probabilistic choice theory and is expressed by the nested logit model. Our particular formulation in this respect follows the derivation of Ben-Akiva and Lerman(1985).</p> <p>We demonstrate that, as evolutionary, our framework is more general than the existing dynamic frameworks. In addition, due to its connection to choice theory it has behavioural characteristics. The connection between the master equation and the nested logit model is provided by the migration rates or equivalently by the set of probabilities of individuals moving between any two zones of the system. We argue that to a set of observed migration rates and population distribution corresponds, through a set of linear equations derived from the master equation, a unique equilibrium population distribution which indicates the tendency of the system at the moment of observation. Thus, the dynamic equilibrium is dependent on the socioeconomic conditions prevailing in the system through the migration rates.</p> <p>Due to the level of aggregation of the available intrametropolitan household head migration data for the Toronto CMA and for the two time periods; 1966-71 and 1971-76, only one part of the nested logit model, the destination choice model, can be meaningfully calibrated. The results show that distance, the percentage of houses built before 1946, the number of new houses built during the study period and the density of the population are statistically significant predictor variables. We also undertaken a preliminary analysis of the impact of new house completions in Toronto City in the 1971-76 time period on the tendency of the system expressed by the dynamic equilibrium discussed above.</p>