Calibration and validation of matching functions for ride-sourcing markets

Shuqing Wei, Siyuan Feng, Jintao Ke, Hai Yang

Research output: Chapter in Book/Conference Proceeding/ReportBook Chapterpeer-review

Abstract

A variety of matching functions have been devised for ride-sourcing markets, such as perfect matching functions, Cobb–Douglas-type matching functions, queuing models, and physical models. However, less is known about the applicability and performance of these matching functions; that is, under what conditions each function effectively characterises a real market. This chapter conducts a series of simulation-based sensitivity analyses to calibrate, validate, and compare the most common matching functions, and to ascertain under what conditions they are applicable. Thus, a simulator is established to generate 420 scenarios of a ride-sourcing market with various levels of supply and demand, and a few widely used matching functions are compared in terms of their key performance metrics—matching rate, passengers' average matching time, passengers' average pick-up time, and passengers' average total waiting time—under various market scenarios.

Original languageEnglish
Title of host publicationSupply and Demand Management in Ride-Sourcing Markets
PublisherElsevier
Pages55-85
Number of pages31
ISBN (Electronic)9780443189371
ISBN (Print)9780443189388
DOIs
Publication statusPublished - 1 Jan 2023

Bibliographical note

Publisher Copyright:
© 2023 Elsevier Inc. All rights reserved.

Keywords

  • Calibration
  • Cobb–douglas-type matching function
  • First-come-first-serve model
  • Matching functions
  • Queuing model
  • Simulation
  • Validation

Fingerprint

Dive into the research topics of 'Calibration and validation of matching functions for ride-sourcing markets'. Together they form a unique fingerprint.

Cite this