Adaptive Multi-Agent System for Dynamic Preference Learning: Application to Mobility

Mobility-as-a-Service (MaaS) is a promising approach to promote sustainable modes of transport and increase the attractiveness of public and shared multimodal mobility services. The long-term objective of MaaS is to change people’s travel behaviour by nudging them to make sustainable choices. However, changing people’s travel behaviour is not an easy task. A MaaS can support this change by providing personalised mobility services, tailored to the needs of each individual user. This paper presents AMAS4PL, an adaptive agent model for dynamic learning of mobility preferences. AMAS4PL aims at learning for each user, their preferences at the level of each regular move they perform. This work takes advantage of the adaptive multi-agent system approach to efficiently distribute the preference learning process at the level of the user’s regular moves. Experiments carried out highlight the system’s ability to compete with well-known preference learning methods while efficiently adapting to changes in user behaviour. The conclusion underlines the ability of AMAS4PL to be extended to other domains.