top of page

Human Decision-Making Dynamics and its Impact on Infrastructure Systems

Research Goals:

1.  To investigate the dynamics of an integrated system comprised of infrastructure systems such as power networks, communication       networks, with a dynamic model of the social interaction with these. This will build on previous work on the complex system dynamics of    power transmission networks and of communication networks.

2.  To develop a hierarchy of simple models to represent the key human reaction and decision-making dynamics identified in the observations of the real systems, and coupled to models of the complex engineering infrastructure. These models will also allow the investigation of the effects of the topology of human networks on the overall system dynamics. 3. To develop analytical tools to quantify and predict regimes of behavior.

Methodologies:

A hierarchy of simple agent based models to represent the key human reaction and decision-making dynamics coupled to a hierarchy of models of complex engineered infrastructure systems is used to investigate the impact of the human interactions with the complex infrastructure systems. This interaction dynamics can then be used to look for feedbacks and vulnerabilities in the coupled systems This research is supported by NSF and is done in collaboration with: David Newman, lead P. I., Kara Nance, and Brian Hay, University Alaska, Fairbanks Ian Dobson, University Wisconsin, Madison Matt Zeidenberg, Columbia University.

Recent Research Findings:

Using a simple dynamic Cascade model representing the operation of an infrastructure system we have begun investigations of the impact of different classes of decision making on the infrastructure dynamics.. The dynamical evolution of the system in the long time scale is governed by a daily increase on consumer demand that raises the overall load on the system and the engineering response to failures that involves the upgrading of the components. The system is controlled through two parameters and two agents who operate the system by selecting those parameters control those two parameters. The utility functions used by the agents to optimize performance incorporate some perception of the events that affect the decision making of the agents. In this preliminary work we look at three social aspects characterizing the agents:

->  Risk averse and risk taking attitudes in the operation of the system.

->  The response to large events triggering a change in behavior on the part of the agents

->  The effect of the learning time in adapting to new conditions.

These three social characteristics have an impact on the performance of the infrastructure system. In going from a risk-taking operation to a risk-averse operation there is a reduction in the frequency of failures and in the number of failures per unit of time. However, risk-aversion brings an increase in the probability of extreme events.

During risk-averse operation, the PDF falls off with a lower power than in normal operation. When risk-averse operation is triggered in response to extreme events, we obtain similar results as in the case of continuous risk-averse operation, but the probability of extreme events can be higher than the continuous operation if this reaction is triggered too often, that is if the threshold for jumping into risk averse operation is relatively low.

bottom of page