Mind the Gap – Passenger Arrival Patterns in Multi-agent Simulations

In most studies mathematical models are developed finding the expected waiting time to be a function of the headway. These models have in common that the proportion of passengers that arrive randomly at a public transport stop is less as headway increases. Since there are several factors of influence, such as social demographic or regional aspects, the reliability of public transport service and the level of passenger information, the threshold headway for the transition from random to coordinated passenger arrivals vary from study to study. This study’s objective is to investigate if an agent-based model exhibits realistic passenger arrival behavior at transit stops. This objective is approached by exploring the sensitivity of the agents’ arrival behavior towards (1) the degree of learning, (2) the reliability of the experienced transit service, and (3) the service headway. The simulation experiments for a simple transit corridor indicate that the applied model is capable of representing the complex passenger arrival behavior observed in reality. (1) For higher degrees of learning, the agents tend to over-optimize, i.e. they try to obtain the latest possible departure time exact to the second. An approach is presented which increases the diversity in the agents’ travel alternatives and results in a more realistic behavior. (2) For a less reliable service the agents’ time adaptation changes in that a buffer time is added between their arrival at the stop and the actual departure of the vehicle. (3) For the modification of the headway the simulation outcome is consistent with the literature on arrival patterns. Smaller headways yield a more equally distributed arrival pattern whereas larger headways result in more coordinated arrival patterns.

Language

  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 01601015
  • Record Type: Publication
  • Files: TRIS
  • Created Date: May 2 2016 3:38PM