Samenvatting

Traditional non-linear models offer Dutch municipalities limited insight into the social behaviour of local stakeholders, even though this behaviour largely determines the success of policies in complex sociotechnical systems. This study addresses the following research question: “How can an agent-based modelling simulation prototype be designed to facilitate dialogue in the decision-making process for a neighbourhood's energy transition?”

The aim is to develop a minimal exploratory agent-based modelling (ABM) prototype of the ‘De Heuvel/Amstelwijk’ neighbourhood in Leidschendam-Voorburg. This prototype will enable municipalities to explore and discuss the impact of basic behavioural principles on physical interventions, such as the sustainable development of homes. This will be achieved through literature research, interviews with experts and local authorities, and the step-by-step development and evaluation of the prototype with policymakers.

Chapter 2 demonstrates that the energy transition at neighbourhood level poses a socio-technical challenge, with municipalities operating within a complex arena of interdependent stakeholders. Although municipalities have formal responsibility, they cannot simply enforce measures at the household level. Instead, they rely on bottom-up change, subsidies, communication and participation to influence behaviour and energy use. Their core task is to transfer power, resources and knowledge from the city level to the neighbourhood level, thereby supporting more sustainable energy consumption.

Chapter 3 argues that equilibrium-based models are unsuitable for the nonlinear, path-dependent and actor-driven nature of the energy transition. In contrast, agent-based modelling (ABM) models can represent individuals and infrastructure as interacting agents and support scenario exploration. However, existing ABMs are often overly complex and poorly aligned with current municipal workflows.

Chapter 4 demonstrates that simulation alone does not generate insight. Insight emerges when results are visualised in a way that facilitates anticipation and shared interpretation. Key elements include GISbased maps to anchor results in a recognisable neighbourhood, dashboards displaying a limited number of clear KPIs, time trends and spatial patterns, as well as careful consideration of the audience, content, context, dynamics and purpose of the data visualisation. The design is based on Nielsen’s usability heuristics, Few’s dashboard principles, and Rams' good design principles.

Chapter 5 introduces the Reasoned Action Approach (RAA) as the behavioural foundation of household agents. Each household is characterised by its attitude, perceived norm and perceived behavioural control, which together form behavioural intentions. When combined with actual control behaviour resulting from environmental factors, this intention determines whether reasoned action takes place.

Chapter 6 translates these choices into minimum implementation requirements, which include: a simplified neighbourhood of dwellings and households based on GIS and attribute data; a behavioural core based on RAA with threshold decisions; a small set of policy levers (subsidies and municipal effort) that mainly affect perceived behavioural control and attitudes; one clear outcome; a dashboard–map interface that can be used by novice and advanced users alike; exportable results that include coordinates and agent states; and implementation in AnyLogic with web deployment.

Chapter 7 demonstrates how these requirements were realised in a working prototype and tested with the municipality of Leidschendam-Voorburg. Users can configure scenarios, view a map displaying dynamic dwelling and household icons, observe time-based dynamics in charts and KPIs, examine the internal RAA status of individual households and export results to Excel alongside GIS coordinates for further analysis and mapping.

The findings indicate that such a prototype can be designed by combining a simple, transparent behavioural model with spatially grounded visualisation and a workflow that fits municipal evidencebased practice. In this configuration, the prototype does not replace participation but supports dialogue by turning assumptions, scenarios and spatial effects into shared, discussable artefacts for decisionmakers and stakeholders. dialogue by turning assumptions, scenarios and spatial effects into shared, discussable artefacts.