Formal Analysis of an Agent Support Model for Behaviour Change Intervention

Abstract

Agent applications have been widely used in behaviour change intervention nowadays. This is due to the four features of agents: proactive, reactivity, social ability and autonomy. However, psychological reactance is one of the major limiting causes of agent interventions. Although, many studies have investigated into both psychological reactance and behaviour change nevertheless how reactive intervention can be supported to obtain an improved behaviour change intervention is still lacking in most previous studies. Therefore, this paper describes the formal analysis of agent support model for behaviour change intervention. The analysis made use of two widely accepted approaches in agent formal evaluation namely mathematical analysis and automated verification. The mathematical analysis examined the correctness of the formal model representation and formalization that aimed to ensure that all syntax and semantic representations used in the formal model is consistent. The mathematical analysis used equilibrium property to explore the formal model consistency. Likewise, automated verification depicts the checking of the model properties against its specifications and theoretical traces. The automated verification used Temporal Trace Language (TTL), which verifies the model properties and states against generated traces. The paper presents an agent support model that allows building agent-based software and applications that deflect psychological reactance and enhance an improved behavioural change intervention.