Agent-based modeling in JavaScript in the browser or on the server. [v0.5.19]

Edge Weight Adaptation in Response to Contagion

In this model, designed and presented at the Complex Networks Winter Workshop 2021 by Herzog, Johnson, Stone, Gao, & Donaldson, a contagion is released on a network. Agents who are infected might infect others, but they might also self-isolate, reducing their likelihood of infecting others to 0. If enough agents become infected, a global ‘lockdown’ is imposed, and all agents may choose to self-isolate (with a higher likelihood if one is infected or has a neighbor who is infected).

In a typical simulation run, we can see dynamics similar to the spread of COVID-19 in the United States and other countries in 2020. Harsh measures are imposed, and the contagion spread dies down. However, once lockdowns are lifted and typical behavior resumes, the rate of spread begins increasing again, and more become infected. Since there is no immunity in this model, these cyclical dynamics could continue indefinitely.