Modeling user behavior and infrastructure level of service: An agent-based simulation approach

Traditional modeling frameworks used for infrastructure asset management have suffered from two main shortcomings. Most approaches have focused their modeling efforts on the infrastructure asset itself, thereby ignoring the multitude of interactions that occur between other entities. In addition, an a-priori behavior of all elements in the modeling environment has always been assumed. This paper argues that these shortcomings have significantly limited the decision-making capabilities of infrastructure asset management systems by limiting their ability to simulate emergent behavior that is typical of all complex systems. One of the less-understood dynamics that is well known to asset managers is how infrastructure users react to changing Level of Service (LOS). This paper presents an agent-based model that simulates the behavior of various user categories to variations in the LOS they receive. User socio-economic characteristics that are analyzed include user expectations and tolerance levels. Political decision-makers are modeled as agents that have the ability to change infrastructure funding levels according to the satisfaction levels perceived by users. The models are formalized using the AnyLogic 6.5 simulation toolkit. A hypothetical example is used to demonstrate the how the interactions among agents in the model results in emergent system behavior. A sensitivity analysis is conducted to demonstrate how changing user socio-economic and psychological characteristics will influence potential intervention strategies by system operators.