Holding Control of Bus Bunching without Explicit Service Headways

Holding-based control methods for bus operation are examined to point out that allowing greater variance in headways between consecutive buses leads to possible gains in total delay, as compared to strict adherence to a service headway. This result, obtained empirically, indicates that optimal operation is not necessarily attained with even headways. Such finding is related to the well-known fact that there should not be too many control points for headway corrections when operating under the traditional method of scheduled departures from bus stations. Current feedback and predictive methods, however, can be productively applied at all stations, hence the importance of studying the effects of frequent control actions. Several feedback schemes are tested, as well as a rolling horizon predictive control method that seeks to minimize onboard and at, station delays. The latter has 110 headway reference and hence yields larger headway variations. The scenario is a BET corridor modeled in a microsimulation environment,. Simulation results indicate gains of 29% in total delay for predictive control in relation to open loop operation, and superior performance when compared to the tested proportional feedback control methods. (C) 2016, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.