Introducing and testing a game-theoretic model for a lottery-based metering system in Minneapolis, United States

This paper is part of an ongoing research that analyzes the feasibility of implementing “buyout” auctions for metering traffic inflow at special use lanes (such as high-occupancy/toll lanes). The term “buyout” refers to the fact that these games allow skipping the auction by paying a price. The ongoing research suggests that high consideration needs to be put on the road design before developing a game-theoretic model. It also suggests that from an operations point of view, if a “buyout lottery” system can be implemented successfully, then its conversion to a buyout auction system becomes straightforward. Therefore, this paper aims at: (1) testing a buyout lottery-based metering (LBM) system in a real case scenario, and (2) introducing a game-theoretic model that recommends to its users the strategy to adopt when deciding between paying the toll or playing the lottery. This paper introduces a microsimulation model of an LBM system that is applied to the city of Minneapolis. This model is then used to evaluate the system’s congestion and revenue effects, always assuming that drivers adopt what is referred as the “safest strategy”. In the evaluation, a set of most likely lottery scenarios is also obtained. Finally, a game-theoretic model is introduced for obtaining the best strategy. This strategy is compared with the safest strategy, in the context of the most likely lottery scenarios. The results from the microsimulation model suggest that the LBM system reduces congestion significantly and increases revenue. The best strategy obtained from the game-theoretic model presents challenges for its calculation. But in the context of the most likely scenarios, this paper suggests that drivers should adopt the safest strategy. The safest strategy may coincide with what are commonly known as “minimax regret” and “maximin” strategies.

Language

  • English

Media Info

Subject/Index Terms

Filing Info

  • Accession Number: 01661108
  • Record Type: Publication
  • Files: TRIS
  • Created Date: Feb 6 2018 4:13PM