Multiresponse optimization of dispatch rules for public bus services

A multiple response optimization approach based on an integrated simulation model and a service reliability measure is proposed for optimizing the dispatch rules of public bus services. The simulation model is based on two novel Markov models for modeling passenger loads and headway fluctuations. In consideration of the importance of passenger loads for bus services of high frequency, a service reliability measure based on the odds of seeing a full bus plying the route is designed. Such a simulation based multiresponse optimization approach can easily consider sampling variations and is scalable to deal with a larger public bus transit system with other objectives, such as minimizing the cost of operations. Practical data collection, analysis and estimation procedures for the load and headway distributions are also presented. A case study based on an actual public bus route in Singapore is used to demonstrate the usefulness of such an integrated simulation framework.