A bid-rent network equilibrium model
This thesis discusses the development of a new model for investigating the relationship between transport and the location of activities. The research consists of three stages. In the first phase, the structural features of the relationship, which are the characteristics of locations, the decision-making processes of households, and the interaction between transport and land-use, are identified. Existing approaches are reviewed using these three components. The review shows that no existing framework satisfactorily represents the requirements in modelling the relationship. Secondly, a bid-rent network equilibrium model is developed. The modelling is considered in terms of competition for locations. Difficulties in analysing the unique characteristics of locations, namely heterogeneity and indivisibility, are discussed. A hedonic interpretation is incorporated to overcome the difficulties. The model represents the decision-making processes of households using the framework of an n-player non-cooperative game. The Nash equilibrium for this game is defined. The game is accompanied by the systematic interactions between transport and land-use. A mutual adjustment process addresses these interactions. A bi-level mathematical programme is suggested to embody the three components. The resulting formulation is interpreted as an oligopolistic Cournot game, which is an approximation of the n-player non- cooperative game. The functional relationship between the decision variables of the upper and lower levels in the bi-level model produces endogenously-determined transport impedance and locational attractiveness. The model incorporates a multiclass framework to consider interclass interactions, which establishes a multiclass bid-rent network equilibrium model. A heuristic algorithm is provided for the solution-finding technique of the bid-rent network equilibrium model. The algorithm combines a path-based routine for calculating the equilibrium solution to the lower level with the Newton-Raphson procedure for estimating the parameters of the hedonic-based stochastic bid-rent function in the upper level. The operation of the algorithm is examined using simple numerical examples. The final stage is an application of the bid-rent network equilibrium model to a real network. A medium sized city is chosen for the case study. The objective of the third stage is to demonstrate the ability of the model to investigate the relationship between transport and the location of activities. A base run and two policy runs are simulated. The base run means a simulation conducted using surveyed data. The policy runs represents the introduction of a congestion charge and the release of land for housing development. Class specific spatial behaviour is obtained. The behaviour is demonstrated using network performance indices representing transport impedance and locational attractiveness. Some policy implications of the simulation are presented.