# Agent-Based modelling - notes

## 'Funadamentals of NetLogo' youtube course from 'Complexity Explorer' - Santa Fe Inst.
- discussion of what models are
- Agent based model -
- agent: autonomous individual element with 'properties' and 'actions' in a simulation
- model the world using 'agents', an 'environment', a description of 'agent-agent interactions' and 'agent-environment interactions'
- not good at understanding 'black swan' events, but can look at *responses* of a statistical simulation to such events.
- Tools
- NetLogo - subject of the tutorials
- Swarm
- Repast (recursive, porous agents) different implementations - java most popular, well-used
- Mason - fast, efficient
- NetLogo (based on Logo - thus agents knowns as turtles)
- Low Threshold
- most widely used, easy to learn!
- 'news and Media' to include models as argument evidence
- High Ceiling
- publishable, expressive models
- verifiable, challengeable models
- ['fire' model](http://netlogoweb.org/launch#http://netlogoweb.org/assets/modelslib/Sample%20Models/Earth%20Science/Fire.nlogo) using netlogo web
- cell colour represents tree / no tree/ fire / burned.
- press 'setup' to get random tree/no tree landscape. Set percentage before setup. try to predict whether fire will spread to opposite side of map. What's the 'critical point' for a 'phase transition' / tipping point?

- [Segregation model](http://netlogoweb.org/launch#http://netlogoweb.org/assets/modelslib/Sample%20Models/Social%20Science/Segregation.nlogo)
- people in an urban landscape - different types (coloured squares) - 50/50 of each type
- problem explored - why might segregation exist?
- rules: as long as 35% of ppl around you are like you, then you are 'happy', if less than this you are 'unhappy'. Unhappy ppl have an X on them, and move to a random open square
- explores 'micromotives' and 'macrobehaviour'
- [Flocking model](http://netlogoweb.org/launch#http://netlogoweb.org/assets/modelslib/Sample%20Models/Biology/Flocking.nlogo)
- now getting into the coding
- components:
-