The purpose of this thesis is to focus on some of the different decisions taken by taxi drivers. The three main areas considered are when a taxi driver chooses to work, where they choose to work and having chosen when and where, the strategy employed by the taxi driver. The chapter examining the decision on when a taxi driver chooses to work considers whether drivers are consistent with the neoclassical model of labour supply. Existing literature on taxi drivers has suggested taxi drivers behave in different ways. This thesis finds that taxi drivers respond to earning opportunities and are more likely to work when the earning opportunities are better than usual. The decision on where a taxi driver chooses to look for fares uses an experimental approach, putting subjects in the position of drivers choosing locations in a square grid. The location choice is a two dimensional extension of Hotelling’s model. In this particular experiment, Nash equilibrium is not obtained through minimum nor maximum differentiation and the learning of subjects and subsequent performance improvement is slow. Simulation through agent-based computational economics is used to investigate the different strategies taxi drivers can use. The simulations show that drivers can increase their own earnings and reduce the waiting time of potential customers by adopting a strategy which makes use of all the available information. The simulations also show that the effectiveness of a strategy is dependent on the choices of other drivers. This thesis suggests a different approach to the analysis of the labour supply decisions of taxi drivers and makes some recommendations regarding regulation of taxi services. The experimental and simulation chapters contribute to the literature through making use of these approaches in the context of looking at taxi drivers. The experiment and simulation could also be extended into other areas.