Want to keep learning?

This content is taken from the University of Twente's online course, Geohealth: Improving Public Health through Geographic Information. Join the course to learn more.

What type of simulation to use?

When we create a simulation model we can choose from a wide range of modelling techniques. The most important decision you will have to take is if you want to model a population or an individual.

In epidemiology many models exist that are population models. These models are often mathematical models and they represent all people living in an area as a population (group). This group is often divided into three types of individuals (susceptible individuals, infected individuals and recovered individuals).

The only thing we know about this group is how many people are in the three categories, namely Susceptible (S), Infected (I) or Recovered (R). Via modelling we calculate in every time step the transition between the three groups. Perhaps at the beginning we have 990 susceptible individuals and 10 infected and 0 recovered. After a week the 10 infected individuals are now recovered, but 20 new individuals are infected leaving us with 970 susceptible.

We do not know however if these individuals belong to a household living together, if they work or go to school, if they meet at all. This can be a problem and therefore another way of modelling developed with is individual based.

In individual based modelling we model each individual separately. This technique is often called agent-based modelling as we model one agent per person. This agent, has certain characteristics, like a place to live, an age, a family, a school to go to, and is also either susceptible, infected or recovered (has a health status). The agent also has behavior (let us say, activities) and can change behavior over time.

Where population models try to model general dynamics, agent-based models assume that when modelling each individual separately, general patterns will emerge describing the total system.

Both of these techniques have their advantages and disadvantages.

Advantages Population Modelling Advantages Agent-Based Modelling
Less computationally demanding Heterogeneity – can include different characteristics of people that may be important to capture the dynamics of a disease.
More suitable for modelling large areas and populations Easy to model a social network - contacts between people
Better in simulating global patterns Better in simulating local dynamics
Disadvantages Population Modelling Disadvantages Agent-Based Modelling
Assumes that populations are homogeneous which in fact they are not Difficult when your population is large because you will need many agents
Difficult to implement dynamical elements like behavior change. May require behavioral data

The type of simulation you choose is not always a rational choice. You may be familiar with one modelling technique and not with another.

We have described an example of agent-based modeling to confirm the hypothesis of Dr. Frank Osei on the influence of dump sites on the cholera diffusion. You can read more about this in the following article:

Reference: Augustijn, E.-W., T. Doldersum, et al. (2016). “Agent-based modelling of cholera diffusion.” Stochastic Environmental Research and Risk Assessment: 1-17. This article is available as an Open Access article for everyone to read.

Share this article:

This article is from the free online course:

Geohealth: Improving Public Health through Geographic Information

University of Twente

Get a taste of this course

Find out what this course is like by previewing some of the course steps before you join: