Agent-based Persuasive Route Recommendation for Public Goods
Over many decades, the transport sector has played a significant role in contributing to economic growth. Unfortunately, this sector has not only provided positive effects, but also has produced a number of negative impacts on society. These impacts are known as the external costs, and include traffic pollution, congestion and accident costs. Transport users rarely take these costs into consideration when they make travel decisions. As a result, the number of external costs is growing and is likely to continue to increase in parallel with the increase of urban mobility.
This thesis proposes a novel recommendation system, known as the Agent-based Public-Friendly Route Recommendation (APF2R). The APF2R can help commuters make green, safe and less congested travel decisions, while supporting society to mitigate the external costs. A novel persuasive reward algorithm is introduced, which can be used by other researchers to balance two conflicted parties. This study demonstrates an agent-based model, which was used to evaluate the persuasiveness of recommendation systems. The result of the proposed system shows potential in addressing the problem of external costs. An analysis of the experimental results undertaken here, captures the evolution of the distance of users’ ranks. These results indicate a means of persuasion in connection with behavioural change.