Royal Society Summer Science Exhibition

One of the most common tasks for the exhibitors may be to reload the model before visitors start. If this is not done visitors will be building on models created by others. To reload:

  1. Click the bookmark in the browser titled ' RELOAD EPIDEMIC GAME MAKER '
  2. Click the shortcut icon on the desktop titled ' RELOAD EPIDEMIC GAME MAKER '

When visitors finish with the exhibit please remind them they can continue at home or school by going to http://modelling4all.org. They will be able to continue building the same model they created at the exhibition by entering their serial number.

1 minute visit

  1. Press the 'Make Your Game' button to generate a model, then 'Go to your Game'
  2. A new browser page will open (and an Applet will load). Press the 'Go' button to start running the simulation
  3. The visitor should notice how the virus spreads from the yellow (index) individual to the rest of the population as time progresses and they move between home and school.
  4. The visitor could run the simulation a few more times (by pressing 'Setup' then 'Go') to learn how stochastic events (especially early on) can have a large effect on the rate of spread of the virus. They can use the graph at the bottom of the page to help make some inferences.

3 minute visit

Close the web page tab where the simulation is running and go back to the Epidemic Game Maker page to change the model. There are many possibilities here:

  1. Check the box to allow School Closing, press 'Make your Game' and 'Go to your game' then the 'Go' button. The visitor can now press a 'Close School' button at any stage during the run of the model. There are many things to think about here e.g. can you stop the spread of the infection, can you slow it, when is the best time to start closing schools. While an infection takes 100 hours to recover, there is only enough money to close the schools for 50 hours so it is very unlikely that one can stop the epidemic this way. See the 5 minute plus section for how one can alter these parameters.
  2. Check the box to add Voluntary quarantine ad campaign to the model. Those influenced by these adverts will stay home once they become aware that they are infected. There is a delay of 24 hours from becoming infectious and having symptoms. It costs to send a message out and if everyone did what the message said the spread of the virus would be halted with minimal cost BUT not everyone does what the message says! With the default parameters this approach will only slow the spread of the infection.
  3. Running Hand Washing or Catch It, Bin It, Kill It advertising campaigns cost money. Hand washing reduces the odds of picking up an infection from a virus trail while the Bin It behaviour reduces the frequency that one leaves viruses behind. As in real life these advertising campaigns only effect the behaviour of a fraction of the overall population.
  4. Visitors can also tick boxes to add 'sliders' for infection odds (the probability that an encounter will result in an infection) and encounter rate (the number of interactions per hour between individuals at the same location that could result in an infection). With lower values for these parameters the interventions have a greater chance of stopping an epidemic.

Visitors can add sliders to their model so they can easily experiment with the following paramters:

  • Infection odds: Odds of becoming infected after encountering someone who is infected.
  • Encounter rate: Number of encounters per hour at the same location. Each encounter with an infected person will cause an infection with the-infection-odds.
  • Infection duration: Average of the number of hours an infection lasts.
  • Symptoms delay: Defines the delay in hours between becoming infected and becoming aware of it.
  • Virus duration: Number of hours a virus lasts on surfaces in the environment.
  • Trail reduction factor: The fraction of the trail of viruses eliminated by those following the Catch It, Bin It, Kill It advertising campaign. The maximum is .99 to avoid a division by zero in leave trails less often.
  • Infection from trails odds: This is the odds of being infected by encountering viruses in the environment. It can be reduced by hand washing.
  • Hand washing odds factor: Applies only to infections due to encountering virus trails (viruses on the surface of things). This is the factor that the odds of infection from viruses is multiplied by.
  • School closing cost: Measured in how time units (hours) for one money unit.
  • Reach of hand washing ad: The fraction of the population that due to changes in their hand washing behaviour reduce their odds of being infected by viruses on the surfaces of things.
  • Reach of stay at home ad: Fraction of the population that will stay home when they become aware that they are infected as a result of running an advert.
  • Reach of catch it ad: The fraction of the population that due to an Catch It, Bin It, Kill it advertisement change their behaviour to catch their viruses thus reducing the trail of viruses left behind.

5+ minute visit

Some visitors may want to look at the underlying code. They can do this by ticking the advanced mode check box at the bottom of the Composer web page and clicking on the agents to 'Show details'. They will then be able to open behaviours to try to follow the programming logic and change values. A simple code change would be to alter the shape of the agents e.g. Person, School, Home.

Changing the code as outlined above can cause the model to behave in unexpected ways or even cause the model to compile with errors. If this happens you can either revert the change (browser back button should work) or simply restart.

Troubleshooting

  1. If you lose the internet connection: there is an offline version of the model. Just go to the desktop and click on the icon labelled 'Offline Epidemic Game Maker'. Don't use this by default because the visitors won't be able to save their model serial numbers.
  2. Error messages: first read the message because it might give you a clue how to fix the problem e.g. change a value in a behaviour. But if this doesn't work you can always start from scratch again simply by clicking the bookmark labelled ' RELOAD EPIDEMIC GAME MAKER '
  3. If the browser crashes: just relaunch a new browser and start again
  4. If things get slow: try not to run too many models at the same time - it is best to close the applet pages after finishing then relaunch another (applets can use a lot of CPU and memory). You may have a poor internet connection (see point 1 above). Wait a bit - applets load slowly the first time you run them in a browser.
  5. The networks works well about 80 to 90% of time. The local host version has some mysterious bugs that I'll look into tomorrow. Mostly agents ignore things like school closing or going home when ill.
  6. After an hour or two the Java applets can get very slow. Closing the browser and starting again fixes it.
  7. If a check box is checked that adds money and then it is unchecked and no other money-related boxes checked that money will not be visible.
  8. If the same problem keeps happening tell or email Ken Kahn: moc.liamg|klatnoot#moc.liamg|klatnoot

The laptops

  • Function F8 toggles between external display, internal display, and both. We'll only use external display.
  • Each laptop has two icons on it: "EGM with Internet" and "EGM without Internet". Always use the first if possible. The second does not have serial numbers.
  • There is a switch on the side for wireless. Probably best to leave it off otherwise it'll try to connect and could slow things down.
  • I'm not sure what screen resolution will look best. I'll experiment when I set up on Wednesday but we could have press visitors on Tuesday (3pm onwards).
  • Remember to use the security cables (I hope that Zoology has ones with combination numbers rather than keys but I guess we'll do OK with keys). Remember to plug in the laptops.

Teachers/exhibitors

Generally we hope visitors will gain a better understanding of why computer models are used to understand epidemics and appreciate the skills that are needed to create models. For instance:

  • Models are safe ways to explore large numbers of possible scenarios e.g. how a virus might spread through a population.
  • The spread of a virus through a population can be chaotic but like other models of complex systems (e.g. the climate) some scenarios are more likely than others.
  • This model is agent-based (ABM) and differs significantly from so-called aggregate models. ABM focus on representing individuals and the way they interact with others. Aggregate models are typically based upon differential equations and can predict the dynamics of an epidemic but are based upon simplifying assumptions e.g. perfect mixing of people. Aggregate models work best when modelling large populations.
  • Models should not be treated as oracles - they are computer programs with assumptions built into them that help us explore and understand systems e.g. how people and pathogens interact. We can develop a better understanding of models by systematic experimenting with parameters, inspecting and changing the code.
  • This model has some things missing that could be important e.g. it does not model death, there are no hospitals, there is only one type of virus and it doesn't change, etc.
  • While this model refers to viruses it can be used to simulate other infectious outbreaks. Indeed this model can be used to think about the spread of more abstract things e.g. rumours and innovative ideas.
  • With some creative thought models can be turned into an educational game. By adding interaction and maybe a competitive or collaborative play element it might be possible to motivate learners to get more involved in the modelling process, perhaps enough to read up on other pathogens or develop some programming skills.
Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License