Name of the code
Box models
Individual based models
Differential equations averaging over the source population
\frac{dS}{d t} = - \beta S I
Susceptible
Infected
Follow individuals in the population
S
I
R
Introduce regional differences
As opposed to box models, can simulate the effect of interventions
(20-th century spanish flu)
(2000's influenza)
Examples: Imperial College (UK), Flute (US)
Individual based models
- Infection transmitted through social interactions
- Interactions take place in social groupings: household / school / work place / wider community
- Interactions take place in pre-defined daily timesteps of fixed duration and strength
- Infection transmitted probabilistically
A COVID-19 model of the 2020's
- Modular: Parameters changed easily, can be extracted from different distributions A tool for epidemiologists
- Allow time evolution of parameters (infectionsness evolution)
- Most important is mixing different scales and age groups: Realistic people that does realistic things
- Include uncertainties (e.g. combining data from different countries)
- All written in Python, by O(10) people
Individual based models can focus on how to get us out of lockdown!
- Test, track and trace (with a certain accuracy)
- Building closures
- Individual's complacency
- How hospitals manage their occupation and the availability of ICU beds