Public transport supply optimization in an activity-based model: Impacts of activity scheduling decisions and dynamic congestion

This paper is concerned about optimal supply levels for a public transport route within an activity-based simulation framework. In a joint optimization of bus headway and fare, we simulate the interaction between users — who choose mode and departure time according to their activities, timetabling and convenience of the public transport service — and a public transport service provider. Differences in optimal headway and fare are identified when departure time decisions of users are endogenously taken into account (in addition to mode choice), in comparison to the traditional approach that assumes given departure time patterns by users over a period. Additionally, impacts of dynamic congestion within private and public transport on optimal supply are analyzed. We find that accounting for departure time choice leads to larger welfare optimal headway and to a higher welfare optimal fare. In all cases, welfare maximization implies shorter headways and lower fares than operator profit maximization. For an uncongested car mode as substitute to bus, we find that congestion pricing for the public transport service is working properly: increasing the fare first leads to welfare gains; raising the fare beyond the welfare optimal price, however, leads to welfare losses. The steepest slope of the welfare function is found when the bus service is working at maximum capacity. For a congested car mode as substitute to bus, welfare maximizing parameters are the same since, at the optimum, car travel times are not subject to congestion any more. For large headways of the bus, we identify car congestion. In this case, congestion pricing for the bus service remains ineffective since both transport modes work at maximum capacity. However, profit maximization yields distinctly higher fares than for the uncongested car mode.