Document Type

Article

Abstract

This paper develops a modeling framework for stochastic multi-agent systems and applies it to equilibrium and pricing analysis in urban taxi markets. Travel demand is represented as a trip network and embedded in a Markov chain that captures both locational and in transit taxi states, with transition dynamics reflecting trip durations, search frictions, spatial competition, and drivers’ perceptions of long-term value. The framework features a parametric Markov chain with endogenous transition probabilities and a behavioral model in which agents’ decisions depend on anticipated long-term rewards. We establish equilibrium existence and examine two locational pricing schemes that align individual incentives with system-level service objectives. The equilibrium conditions are formulated as a nonlinear program and solved using successive approximation algorithms. A case study based on Chicago taxi data illustrates the model’s use for regulatory and fleet analysis.

Disciplines

Industrial Engineering | Operational Research | Systems Science | Transportation Engineering

Share

COinS