Simulation Of The Traditional Taxi Service And Online Taxi-hailing Service’s Operation Based On Agent-based Modeling

This paper presents an agent based model for simulating the demand development of taxi services in urban areas.The two operation modes are modeled and tested: hailing(the traditional taxi companies)and dispatching from the mobile application through the Internet(Didi,Uber).The model represents the real life demand,where population of the passengers(agents)every day decides what taxi company to choose.The technology of creating a taxi transport simulation model of the city,including the existing volumes of traffic and the tendency to increase the market,is presented.The tools for developing models,methods and sources for collecting initial information,methods for calculating parameters necessary for modeling,and analysis of the dependence of the distribution of passengers on two groups of taxis on the price and time of arrival are described.The simulation model is formulated in the graphic language of the Anylogic program.Simulation allows optimizing the system before its implementation.Simulation involves mapping the problem from the real world to the world of models(the process of abstraction),analyzing and optimizing the model,finding a solution,and mapping the solution back to the real world.Agent modeling is considered not as a replacement for traditional approaches,but as a very useful addition to system dynamics and discrete-event modeling.First,a basic model of the taxi market development was created.It reflects the nowadays passengers’ distribution between the traditional taxi services and the internet taxi services and the dynamics of the growth of the taxi market in the city.After this the model analyzes the changes in distribution due to changes in the initial parameters that affect the demand of passengers-this is the price of fare and waiting time of taxi’s arrival.The result of the experiments is analytical information about how the distribution of passengers is changing relative to the basic version of the taxi market development in Harbin.The results of the cases reflect a change in passengers demand,influenced not only by the main parameters(cost of the fare,waiting time of taxi arrival),but also by special parameter “attractiveness”,which has described by the Elaboration Likelihood Model theory and behavior economics.It allows to model customer loyalty,in which the client continues using this company even if the cost of fare of this company is higher than the competitor’s fare.