Understanding how and why people traverse a city is vital to successful transport planning. Policy makers need to formulate policy, anticipate demand on their travel networks and identify the right investments.
Arup’s City Modelling Lab brings together our experts in transport, energy, climate change and economics with data scientists, software engineers and designers to answer these challenges. We build models at city, regional and national levels that provide the insights decision makers need.
In modelling future demand, we run multiple simulations to produce a spectrum of likely scenarios, to capture uncertainties and highlight where sensitivities may lie. Our models are rigorous, enabling authorities to plan those future policies and investments with new levels of confidence and clarity. We have built our City Modelling Lab using open source data, drawing on the technical expertise of the community, and publish the open source software tools we develop. This allows public authorities to remain transparent and accountable.
How we can help
Agent-based modelling
Traditional modelling fails to mirror the complexities of life and the rapidly changing transport sector. Assumptions on how people move and behave have grown increasingly inaccurate over in the past decade as choices have expanded in number and complexity.
The ambition for our City Modelling Lab is simple. To simulate everyone’s day more accurately and rapidly than previously possible. We capture that individuality by using Agent Based Modelling (ABM). We model each individual, each with a plan, each making decisions on where to travel, when to leave and what mode of transport to use. Each influenced by the choices of other individuals in the model and the evolving choices available to them.
Unlike traditional simulation, we model the ‘why’ as well as the ‘what’, gaining insights into complex behaviours. This allows us to better understand shifts in travel behaviour, establishing a truer perspective on the impacts of policy decisions, city planning and new transport schemes enabling more fair, equitable and sustainable decisions and investments for our society’s future.
Modelling city life
Cloud computing has unlocked the ability to model the complexities of city life at massive scale. As Transport for London’s ABM research partner, our model of the city uses over 1,000,000 agents each making 750 decisions over the course of their day. The acceleration in computing power now allows these high-fidelity models to run in the same timescales as less refined, aggregate models. It also opens the way for greater equity in transport planning, reflecting the needs of a wider range of people.
We build our models ground up, using travel diaries and census details to populate networks with individual profiles. Faster computing will also allow us to add more data sets and increase the accuracy. We are exploring how to overlay mobile phone and credit card data to create richer profiles, modelling individual behaviour at the micro level.
Dashboard simulates possible ways to improve the air in Berlin
Agent based model for mobility and emission scenarios, Germany
Simulating changing transport behaviours in just six weeks
Birmingham's Agent Based Model, United Kingdom
Getting ready for future road users
Transport Infrastructure Ireland, Ireland