Analysing taxi customer-search behaviour using Copula-based joint model

Abstract The drivers of vacant taxis tend to cruise the road network searching new passengers, which leads additional traffic congestion, air pollution and other problems. This study introduces a Copula-based joint model analyse destination selection route choice behaviour in customer-search process. A multinomial logit is used behaviour, path size explore routes behaviour. Accordingly, applied Copula function then established correlation between these two behaviours. customer generation rate, distance, travel time, cumulative intersection delay, length are selected as explanatory variables. trajectory data were collected by global positioning system Xidan District Beijing City from September 2014 February 2015. Using log-likelihood, Bayesian information criterion evaluation indexes measure fitting result, has highest goodness-of-fit. effect variables on search discussed based parameter estimation results. results this helpful understand taxi reduce operating costs improve efficiency operation system.