The Humat: The Architecture for a Data Driven Simulation of Social Innovations

Wander Jager, Patryja Antosz, Groningen Center for Social Complexity Studies, University of Groningen

The EU Horizon2020 SMARTEES project models cases of social innovations implemented in European cities. Investigated innovations promote low-carbon energy sources, ranging from communities insulating houses to cycling for urban transportation. The aim is to support local governments of cities in transitioning to energy efficiency and sustainability through simulating plausible effects of implementing similar social innovations in new contexts. As success or failure of urban innovations depends on the social dynamics emerging in a community, we developed the HUMAT agent-based modelling architecture that captures key drivers of social action.

Each HUMAT in the model chooses weather to act innovatively (e.g. bike to work). Chosen behaviour maximizes the individual level of satisfaction on three groups of needs: (1) subsistence, i.e. immediate consequences of action, (2) social, i.e. group (non)conformity, and (3) personal, i.e. individual values. Each HUMAT is also equipped with memory of past experiences and can inquire about perceptions of relevant others in ego-networks.

Decision-making strategies and subsequent behavioural changes of HUMATs are driven by cognitive dissonances that occur between cognitions in the three groups of needs. When a HUMAT is satisfied with the results of action, it behaves habitually. However, introducing a socially innovative alternative may evoke dissonances and decrease satisfaction from routines (e.g. a HUMAT becomes less satisfied as the only driver in a group of bikers). Strategies reducing cognitive dissonance include a change of behaviour (e.g. switching to riding a bike), and/or pressuring others to change their actions (e.g. convincing friends to drive), and/or re-evaluation of the importance of needs (e.g. decreasing the importance of identifying with a group of friends).

The HUMAT framework is parametrizable with empirical data to various cases (e.g. cities, innovations), provides possibilities of systematically testing various scenarios in-silico, and sheds light on probable dynamics of social processes.