Investigating boundary effects of congestion charging in a single bottleneck scenario

Many congestion charging projects charge traffic only within part of a day with predetermined congestion tolls. Demand peaks have been witnessed just around the time when the charge jumps up or down. Such peaks may not be desirable, in particular (a) when the resulting peaks are much higher than available capacities; (b) traffic speeding up to get into the charging zone causes more incidents just before the toll rises up to a higher level; or (c) traffic slowing down or parking on the roadside decreases road traffic throughput just before the toll falls sharply. We term these types of demand peaks ‘boundary effects’ of congestion charging. This paper investigates these effects in a bottleneck scenario and aims to design charging schemes that reduce undesired demand peaks. For this purpose, we observe and analyse the boundary effects utilising a bottleneck model under three types of toll profiles that are indicative of real charging schemes. The first type maintains a constant toll across the charging period, the second type allows the toll to increase from zero to a given maximum level and then decrease back to zero and the third type allows the toll to rise from zero to a given maximum level, remain at this level for a fixed period and then fall down to zero. This investigation shows that all three types of toll profiles can produce greater boundary peak demands than the bottleneck capacity. A significant contribution of this work is that instead of designing an optimal traffic congestion pricing scheme we analyse how existing sub-optimal congestion pricing schemes could be improved and suggest how observed problems may be overcome. Hence, we propose a set of extra requirements to supplement existing principles or requirements for design and implementation of congestion charging, which aim to reduce the adverse consequences of boundary effects. Concluding remarks are made on implications of this investigation for the improvement of existing congestion charging projects and for future research.