Activity-based Analysis of Travel Demand Using Cognitive Agents

This work reports on the use of cognitive agents to support the activity-based analysis of demand generation in urban traffic networks. Albeit motorists show reactive behaviour in many situations, for instance when they are effectively driving, interacting with others and with the traffic environment, in this work demand is regarded as the result of the cognitive process carried out by each individual of the population. The approach proposed allows for the diversity of activity parameters that may influence daily journeys decision-makings. The preliminary experiments are aimed at demonstrating the expressiveness of the cognitive model proposed in practically designing and implementing activity-based travel demand. Emphasis is initially given to analysing scheduled delay at activity destinations accounting for arrival time constraints, while different selection behaviours for departure time are simulated within a framework that combines a multi-agent demand model with the microscopic representation of the movement. Simulation results have shown that drivers are more likely to meet activity arrival constraints with an absolute earliness-lateness tolerance window than when tolerance thresholds are relative to travel time.