Chaos and Instability
This lecture will discuss Chaos and instability. Students should be able to:
- Define chaos and differentiate it from randomness
- Understand pseudo random numbers and their role of in Agent Based Models
- Understand model bifurcations
- Understand the notion of attractors and attractor maps.
ToDo: Parameter sweeps in models are used to test for chaos and instability.
- Use the learning model that you are developing to test, if you have it done already. Alternatively, choose any model from NetLogo that you like, that has at least 2 parameters that can be varied.
- Think of a interesting experiment where you need to analyze a 2D parameter space.
- Decide which output metrics you will measure. Use one or two, to limit the amount of graphing.
- Think about the stop criterion, or limit the maximum number of ticks.
- Perform the experiments and report them as usual.
- Be aware that a parameter sweep can quickly get out of had in therms of computational time
- First, calculate the number of parameter points that you want to test.
- Then, time a single run to get a sense of the overall duration of the experiment
- Adapt your parameter space as necessary. You can either reduce the bounds, or increase the coarseness.
- Be aware that laptops can get very warm after hours of crunching numbers.
Below is a link to the lecture slides.
Agent Based Modeling of Complex Adaptive Systems (Basic) by TU Delft OpenCourseWare is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
Based on a work at https://ocw.tudelft.nl/courses/agent-based-modeling-of-complex-adaptive-systems-basic/.