PhD Dissertation Proposal by Larry LIN | Modelling Movement Decisions in Networks : A Two-Stage Conceptual Framework

Movement is part and parcel of everyday life. People move for multiple reasons, whether it's for getting from an origin to destination (e.g. home to workplace), or for livelihood purposes (e.g. taxi drivers in search of passengers). On top of that, movement can occur at different frequencies, whether its continuous movement in matter of seconds (when driving), to ad-hoc movement which requires a contemplation period of years (e.g. migration). In this dissertation proposal, we propose a two-stage conceptual framework for modeling movement decisions of agents in a network environment. First, we represent user movement decisions in networks via a utility-driven decision model. This is essentially a probabilistic representation of the propensity to move from one node to another given several network-level observations. The second stage of the framework then involves taking the utility-driven decision model as an input to an agent-based simulation platform. The generalizability of the conceptual framework is demonstrated with concrete use cases in the problem domains of (1) leisure, (2) transportation and (3) migration. The proposed framework paves the way for development of high quality agent-based models for capturing the movement decisions of agents in networks, and thereafter serving as evaluation tools for various recommendation systems and policies.

About the Speaker