Using Networks to Survive Disasters
This article by postdoctoral fellow Cristina Ruiz Martin and professor Gabriel Wainer, both at Carleton University, was about communications during disasters and how concepts in networks can be leveraged to create an improved model for communications during these times. The story first starts by citing several examples of previous disasters including Hurricanes Dorian and Irma, forest fires in British Columbia and Australia, earthquakes in Japan and New Zealand, and the Fukushima Daiichi nuclear disaster. This is done to convey to the reader that natural and manmade disasters are becoming more prevalent.
The article then focuses on resiliency, which is how well a government can repair damages from a disaster, and how it is crucial for a government to maintain a high resiliency. It mentions that many new emergency plans are structured as a series of commands rather than as flexible plans that account for unanticipated situations and the unexpected behavior by people in said situations. This inability to adapt is similar to the discussion of the diffusion of innovations in lecture and the textbook, specifically whether there exists direct-benefit effects such as having the correct technology to communicate and the ability to integrate new features into existing infrastructure. Without direct-benefit effects, it can be difficult to adopt innovations in networks.
The authors particularly noticed that while it is important to continually improve the reliability of communications, not much emphasis is put towards actually testing and improving these networks especially when they have to work during complex situations. They proposed a new model that improves these networks and then used an information diffusion process to show how the communications spread in their model.
This process allowed the authors to test their model under combinations of different scenarios regarding individuals, their behaviors, and how existing communications were structured. Their tests concluded that individuals’ behavior has a significant impact on resilience. One test cited was when information did not pass properly because multiple attempts to contact someone failed.
This relates to what was studied in lecture regarding the adoption of a behavior or a product. A node adopts a behavior/product if the payoff for switching is higher than their current internal payoff. A cascading behavior can occur from simply one action, which is similar to the study’s phone call test collapsing an information network. Clusters can also form from how individuals react to such situations and prevent a complete cascade from happening.
Overall, the article brought up several critical points. It is important for companies and organizations relying upon communication infrastructure to study their networks and make adjustments to their emergency plans as needed. Resources such as time and money are conserved for real time scenarios when simply testing models. Most importantly, this particular model will allow companies to provide reliable service to customers and organizations to improve the resiliency of mass communications during disasters.