Modelling Context-Sensitive Dynamic Activity-Travel Behavior Under Conditions of Uncertainty Incorporating Reinforcement Learning, Habit Formation, And Behavioral and Cognitive Adaptation Strategies

Abstract

This paper develops a framework for modelling the dynamic formation of location choicesets. The proposed framework integrates three key concepts, namely aspiration, activation and expected utility. Aspirations are defined at the level of attributes of choice alternatives and represent an individual’s beliefs about performance levels that potentially can be achieved. Activation levels are defined at the level of choice alternatives and represent the ease with which an alternative can be retrieved from memory and, hence, the degree of awareness of an alternative. Finally, expected utility represents an individual’s evaluation of a choice alternative based on his/her current beliefs about attributes of the alternative. In the proposed system, all these cognitions - aspirations, activations and beliefs – are conditional upon context variables and subject to cognitive and social learning under uncertainty. We describe the core of the model and illustrate its properties using numerical simulation.