Microscopic traffic simulation models are an important tool for transport system analysis and management. Within these models lane-changing is an important component, because lane changing is an important factor according to traffic instabilities. Two types of lane changes can be distinguished, namely mandatory and discretionary lane changes. Mandatory lane changes need to be performed to keep the route, e.g. merging at a freeway on-ramp. Discretionary lane changes are performed to improve driving conditions, e.g. to overtake a slower driving vehicle to increase speed. Because this study is about the merging behaviour on a freeway on-ramp, the focus will be on mandatory lane changing.
Most of the micro simulation software use a gap acceptance model to model the merging behaviour on a freeway on-ramp. A driver who wants to merge must evaluate the available space between a potentially conflicting vehicle and itself. If the gap is large enough he decides to merge, otherwise he will wait for a next gap. With gap we mean the time between two vehicles a driver needs to merge safely between these vehicles. A gap acceptance model describes how a driver determines to accept a gap or not.
A recent research on merging behaviour have showed that this gap acceptance models are not able to represent real observed merging behaviour. According to this research a new theory on merging behaviour can be proposed. This theory will be based on gap selection. Every merging vehicle is able to find a suitable gap without being overtaken on the main road and without coming to a standstill at the end of the acceleration lane. A set of reachable gaps is defined for every merging vehicle. Using a choice model a choice will be made from this set of gaps.