Introduction

The focus of this whole text is to emphasize the modular structure of transportation simulation packages, and in particular that besides the movement of the cars through the system considerable effort needs to be spent on modules which model human learning and decision-making, and on mechanisms which couple those modules. In consequence, we have started (in Chap. 7) with a simple micro-simulation which is able to support our approach, which means that it has individual vehicles which follow individual plans. However, the simple approach of Chap. 7 neither looks at correct vehicle speed not at correct link flow capacities.

In this chapter, it will be discussed how the CA traffic simulation from Chap. 7 can be made more realistic. In fact, this type of simulation is used in the Transims simulation package for transportation planning. Ultimately, also the CA approach has its limits and is better replaced by an approach where the spatial coordinates are continuous (Chap. ). The CA approach has however the advantage that its implementation is rather straightforward. This is due to the simple spatial structure, in which the existence of a vehicle at a specific location can be checked via a simple direct lookup at the corresponding cell. Techniques with continuous coordinates typcally store the position of the particle together with the particle, i.e. not together with the spatial substrate, so that the existence of vehicles at specific locations needs to me made computationally efficient via other methods. These problems can be overcome, and the resulting models are as efficient as CA models, but they represent some conceptual and programming overhead that needs to be recognized.