From the Experts: Carolyn Beck, Ph.D.

Carolyn Beck, Ph.D., Professor in the Department of Industrial and Enterprise Systems Engineering at the UIUC, responds to a question she received during the COVID-19 Virtual Summit panel discussion on Monday, April 6, 2020. This question was asked in response to Professor Beck, Professor Sergei Maslov, and Professor Nigel Goldenfeld’s separate talks on epidemiological compartment modeling.

Question: Could you explain the difference in the possible uses of your models? Since they use different methodology, would they be useful for different purposes?

Answer from Professor Beck:

We are actually starting from the same place in our modeling: epidemiological compartment models. The models I discussed were what we call SIS models (susceptible-infected-susceptible), as the simplest example of compartment models. I then discussed incorporating network structures into these models, which means that you need to think about the dynamics of the disease over a mathematical graph. So, now we think about the nodes as their own little compartments, with some probability of infection, as interconnected to other nodes via edges in the graph, which may have varying weights or strengths.

Professor Maslov and Professor Goldenfeld were also discussing epidemiological compartment models. They discussed SEIR models (susceptible-exposed-infected-recovered). These are more specifically relevant to COVID-19, and they are looking at estimating the model parameters from data taken in different geographical areas (although I believe they have mostly focused on Illinois and the Chicago area).

To connect these models, we might take the distinct geographical area models that Professors Maslov and Goldenfeld have estimated, then look at human mobility and/or traffic patterns between these areas and connect them with network-like structures to arrive at the networked epidemiological models I discussed.

To view Professor Beck and Maslov’s talks, see the COVID-19 Virtual Summit here.