The use of scenario development is a common and useful way to imagine new futures in for example urban planning, policy making, but also in mobility planning. By creating hypothetical plans which depict a set of plausible futures, often expressed in (policy) measures or actions, uncertainties can be accounted for effectively. Instead of focusing on fixed ideas or plans, a set of plausible options is explored.

We do this for the implementation of street experiments and the combined deployment of shared mobility options.

Street experiments and shared mobility options are two increasingly popular measures cities are considering to move toward a more sustainable future. Either by enhancing a shift from cars to more active transport modes or by enhancing accessibility and liveability. What is not so clear, and often under-explored in research, is the impacts these measures might have on mobility. In this dashboard, we do so by simulating scenarios through agent-based modelling.

But why creating scenarios through agent-based modelling?

Some advantages:

• Agent-based modelling is an easy and cost- and time-efficient way to simulate alternative futures
• It allows to make hypothetical changes to the transport system, including changes to the street network and to available transport modes. Examples include: creating or removing streets or bridges, introducing circulation plans or traffic-calming zones, introducing new public transportation lines, experimenting with new forms of shared mobility, etc.
• It allows to simulate mobility situations with a sample population. This can be done with both real-life sample data (such as travel diaries) as with census data (building a synthetic population)
• Scenarios can be tailored to the needs and interests of cities, including mobility practitioners

Some disadvantages:

• An agent-based model is an abstraction of the reality and consequently works with assumptions. This means that some situation will not be as “realistic”. An example is vehicle ownership such as bicycles. When not having data on bicycle ownership, the model assumes that every person can use a bicycle. See assumptions for more information on this.
• Constructing scenarios in MATSim is flexible but at the same time static. This means that various scenarios can be developed according to different parameters, but creating an interactive environment where changes can be implemented instantly is not possible. Each scenario must be modelled individually.

Based on the outcomes of the ABM approach, some key findings and recommendations for policy makers can be formulated in terms of

• the implementation of street experiments in the form of street closures in residential city streets
• the deployment of shared mobility options in combination with street experiments

What are the wins for implementing street experiments across the city?

• Implementing street experiments in residential areas on a small scale (low amount) has a more localised impact, but when extended, the impact reaches to a wider city level.
• When street experiments are combined with shared mobility options, the impact on traffic volumes and modal shifts is even larger.
• Implementing a considerable amount of street experiments tends to lower overall car traffic in the city.
  • Decrease is most prominent in the vicinity of the street experiment, but re-routing needs to be considered. In some cases, traffic may be redirect to adjacent streets, leading to a surge in car traffic. In other cases, overall car traffic diminishes in the direct vicinity of the intervention. Good knowledge and monitoring of the traffic situation is necessary.
  • Traffic can be re-routed to ring roads, but when a considerable amount of experiments are implemented, an overall decrease of car traffic can also be seen on these roads.
  • Implementing street experiments in city centres considerably reduces car traffic in these areas. This holds a huge potential for moving toward a low traffic city when it goes along with a broader set of car traffic calming measures, such as improved bicycles lanes or pedestrianisation schemes.
• It has the potential to enhance a modal shift from car use towards more active uses, particularly bicycles use.
• Combining street experiments with shared bicycle options can be more beneficial then providing shared car options. Providing more bicycles also increases general bicycle uptake and reduces car uptake, and thus also car traffic.

What are some issues to take into account?

• Street experiments tend to redirect car traffic (and thus increase car traffic in some areas).
  • Redirection occurs toward arterial roads and key traffic points yielding already high levels of car traffic, causing possible traffic jams - when extending the expansion of street experiments to a higher number, this effect is tampered.
  • Redirection occurs toward neighbouring streets. This might lead to protest of inhabitants experiencing a sudden increase in car traffic in their street.
• Combining street experiments with shared car options increases car uptake and consequently also car traffic volumes. The provision of shared cars can induce an unintended modal shift from bicycle and public transport use toward car use.

Some final considerations

Efforts to reduce car usage in numerous cities frequently encounter protest. It is a very sensitive topic that tends to polarise society further. It is therefore key to stay in touch and engage with communities and residents. Communication is vital.

Also, traffic outcomes and travel behavioural effects are not the only aspect to take into consideration when implementing street experiments. There is a number of other benefits worth considering, which are difficult to quantify with an agent based model. Street experiments also bring about social interaction among citizens and have the potential to enhance sense of community, mental health and wellbeing. For detailed guidance on how to implement street experiments, please visit streetexperiments.com, where you can access a comprehensive guideline kit!

In agent-based modelling, assumptions play a pivotal role in shaping policy insights and decision-making processes. For policy-makers, understanding the underlying assumptions is crucial for interpreting model outcomes and crafting effective strategies. Assumptions in agent-based modelling include various facets, including agent behaviour, interactions, environmental dynamics, and system feedback loops. These assumptions often reflect simplified representations of complex real-world phenomena, enabling policy-makers to simulate and explore diverse scenarios. However, it's essential for policy-makers to critically evaluate these assumptions, considering their implications on the model reliability and the robustness of policy recommendations. Transparent documentation and sensitivity analyses can aid in understanding the effects of assumptions, empowering policy-makers to navigate uncertainties and leverage agent-based modelling effectively in policy formulation and evaluation.

In general, it should be clear for policy makers that no model will mimic all facets of reality correctly. Therefore, all results should be interpreted accordingly; there will always be some room for error. Overcoming certain assumptions is possible but can proof to be hard, yet it can be worthwhile to see which direction the agent-based model will shift if certain assumptions are altered.