Ph.D. Defence Notice: Social-Influence based Control of Voluntary-Behaviour Adoption in Small Groups using Agent-Based Modelling

Thursday, July 11, 2024 1:00 pm - 3:00 pm EDT (GMT -04:00)

Candidate: Keegan Aaron Fernandes

Title: Social-Influence based Control of Voluntary-Behaviour Adoption in Small Groups using Agent-Based Modelling

Date: July 11, 2024

Time: 1:00 PM

Place: EIT 3145

Supervisor(s): Davison, Dan - Wang, David

Abstract:

We take a systems control approach to explore various control influence techniques and investigate how effective they are at changing behaviour. We develop a psychology-based discrete-time behavioural change model of individuals in a group environment. The developed model incorporates work on individual behavioural change, namely the reasoned action approach, to represent how individual beliefs affect behaviour. Additionally, the model uses ideas from the study of social influence to account for social effects on individual beliefs such as conformity and internal biases. Feedback mechanisms are used to integrate the two models. Simulations of the developed model display characteristics of clustering and polarisation, consistent with empirical evidence. The impact of internal bias in driving this phenomenon is noted and developed into an external influence strategy. This strategy along with several others is evaluated using simulations of the model. The results of these simulations show that taking advantage of internal biases is an effective influence strategy. Additionally, results show that an influence strategy of rotating to different individuals can lead to more individuals performing a behaviour when compared to a ``one to all" influence technique. The developed control strategies are deployed on a real-world test scenario and the behavioural change results are compared to those of simulation. Results show that the developed model produces behavioural outcomes whose trends are consistent with real-world results. These results provide evidence that the model can be used to predict the effects of different influence techniques and to investigate which influence strategies are effective at changing behaviour and which are not. The results also provide insight into how behaviours propagate through a group environment.