Investigating machine learning for simulating urban transport patterns: A comparison with traditional macro-models

Aug 3, 2026

Upload by Himbo

Predicting passenger flow within a city is central to intelligent transportation management, particularly under urban development pressures, post-pandemic policy shifts, and infrastructure upgrades. This study compares a selective machine learning approach with traditional simulation methods using Oslo, Norway, as the test case. Sensitivity and scenario analyses assess model parameters and city characteristics. Results show the traditional spatial interaction model is interpretable with fewer parameters but less able to reproduce real flow structures and with higher variability. Statistical evidence questions the ML model’s robustness yet indicates potential for scenario-based trend identification and decision support for planners.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages16

Reading Time40 min

AccessFree to download

Abstract

Introduction

Literature review

Investigating machine learning for simulating urban transport patterns: A comparison with traditional macro-models Downloaded from: [https://research.chalmers.se](https://research.chalmers.se), 2026-08-02 19:46 UTC Citation for the original published paper (version of record): Parishwad, O., Jiang, S., Gao, K. (2023). Investigating machine learning for simulating urban transport patterns: A comparison with traditional macro-models. Multimodal Transportation, 2(3) . [http://dx.doi.org/10.1016/j.multra.2023.100085](http://dx.doi.org/10.1016/j.multra.2023.100085) N. B. When citing this work, cite the original published paper. research.chalmers.se offers the possibility of retrieving research publications produced at Chalmers University of Technology. It covers all kind of research output: articles, dissertations, conference papers, reports etc. since 2004. research.chalmers.se is administrated and maintained by Chalmers Library (article starts on next page) Multimodal Transportation 2 (2023) 100085 Contents lists available at ScienceDirect Multimodal Transportation journal [homepage: www.elsevier.com/locate/multra](homepage: www.elsevier.com/locate/multra) | Full Length Article
Investigating machine learning for simulating urban transport patterns: A comparison with traditional macro-models
Omkar Parishwad a,∗, Sida Jiangb, Kun Gao a
a Chalmers Institute of Technology, Architecture and Civil Engineering, Gothenburg, 41296, Sweden b WSP,Transport Advisory, Globen, Stockholm, 12188, Sweden | | | | | --- | --- | --- | --- | | a r t i c l e i n f o | | a b s t r a c t | | | Keywords:
Intelligent transportation systems Passenger ﬂow prediction Spatial interaction model Traﬃc simulation
Sensitivity analysis Transport planning Norway | | Predicting passenger ﬂow within a city is crucial for intelligent transportation management systems, especially in the context of urban development, post-pandemic policy changes, and infrastructure improvements. Traditional macro models have limitations in accurately capturing the complex structure of real traﬃc ﬂows, and recent advancements in machine learning oﬀer promising approaches for improving transportation simulations. This research aims to compare the eﬀectiveness of traditional simulation models with a selective machine learning (ML) model for traﬃc ﬂow prediction in Oslo, Norway. Sensitivity and scenario analyses are conducted to examine the models’ parameters and derive the city’s characteristics.
Results substantiate that the traditional Spatial Interaction model (SIM), although interpretable and requiring fewer parameters, has limitations in accurately capturing real ﬂow structures and exhibits greater variability compared to the ML model. Statistical analyses support these ﬁndings and raise questions about the validity of the ML model’s results over the SIM. The research highlights the potential of ML models to identify trends in passenger ﬂows and simulate traﬃc ﬂows in diﬀerent scenarios related to city development. Overall, the research presents a decision support system for planners and policymakers to predict traﬃc ﬂow accurately and eﬃciently. It highlights the beneﬁts and drawbacks of both the traditional SIM and ML models, contributing to the ongoing discussion of the role of machine learning in transportation modeling. | | 1. Introduction The Land Use Transport Interaction (LUTI) models have been theorized upon correlating factors by transport and city planners to explain the existing ﬂows within the city (Wegener, 2004). Various endogenous phenomena metrics, such as measurement of workplace accessibility (Waddell, 2002), long-term implications of policy decisions, transportation infrastructure, and trip production activity distribution, real-estate and housing supply, goods and passenger transport distribution, incorporated into urban transportation or public services, were based on the classic four-stage sequential approach

FAQs

What is the main goal of this research on urban transport modeling?

To compare the effectiveness of traditional simulation models with a selective machine learning model for traffic flow prediction, using passenger flow in Oslo, Norway as the application case.

What modeling methods are compared in the study?

The study contrasts a traditional Spatial Interaction Model (SIM) with a selective machine learning model for passenger-flow and traffic-flow prediction.

What do the results suggest about the Spatial Interaction Model and the ML model?

The traditional SIM is interpretable and uses fewer parameters but captures real flow structures less accurately and shows greater variability. The statistical analyses also raise concerns about the ML model’s validity, while highlighting ML potential for identifying trends and simulating scenarios.

Aug 3, 2026

Upload by Himbo