In my opinion, it should be thought of from a traffic assignment point of view. Since the expanded capacity of the road may ease the traffic, therefore more people may be attracted to a path that uses this road as a link in hope to reduce their travel time. Thus it will create more an additional flow volume on this road. By how much? We will need to solve a network equilibrium problem.

This is very interesting question. I am interested in it as well, from the perspective of road safety impact assessment, according to European Directive of road safety infrastructure management. It is process of assessing safety impacts of new roads or re-design of existing roads. Safety consequences may be assessed through crash prediction models (safety performance functions), but these contain traffic volume - which may change, as mentioned in your question. I wonder whether there are some standardized procedures in this regards.

We are exploring this right now with one of the state road agencies. There are many conventional transport models available at macro-, mezzo- and micro levels which will estimate changes in traffic flow due to a road upgrade. There are commercial products used by specialists. Depending on the modelling method they use land-use based induction of traffic, route choice assignment based on origin-destination (very simplified description). Some examples are: AimSUN. VISSUM/VISSIM, Paramics. You can hire someone to model induced demand for you using the right tool. We are now looking at developing road safety models which work with input/output variables of these models, rather than more traditional AADT & infrastructure variables. Early days still, and very interesting.

Hi Chris, it sounds very interesting. Looking forward to see some of your results!