EMBER - Emergent and Macroscopic Behavioural ExtRaction:
an advanced model analysis tool
EMBER systematically analyses the emergent (macro-level) behaviour of any model of a complex system through the use of sophisticated mathematical techniques. The design of EMBER enables this investigation of a model without requiring knowledge of the underlying techniques through a simple user-friendly interface. By coupling innovative and advanced mathematics with well-established tools, EMBER can uncover properties such as path dependencies, and tipping points.
The tool interfaces with a model to uncover how, and why, the system exhibits certain behaviour. EMBER is a unique analysis tool that can provide insight which is unobtainable using standard tools. The development of a user interface allows the investigator to set up and run the analysis without requiring a background in programming. Furthermore, all the outputs from EMBER are tab delimited text files and can be visualised using the investigator's preferred software.
Two examples of EMBER's analysis
Each of the diagrams below shows the interface of the model in the top left, and the outputs uncovered by EMBER in the bottom right.
1. Fire spreading in a forest and how it is affected by the initial density of trees. As the density of trees rises, we see a non-linear increase proportion of trees burnt.
2. Lake management policies regarding residential pollution and the effect on fish and plant populations. This is a model of government legislation relating to levels of pollution by local residents and its effect in a nearby lake. Increasing pollution causes an increase in nutrients that promote algae growth and ultimately affect the population of fish and plant life in the ecosystem. EMBER can be used to compare different policy scenarios to evaluate how effective they are at reaching a desired goal (such as increased fish levels), but also what the associated costs are (such as financial, time, government popularity, etc). The results show the levels of Bream, Pike and Vegetation are dependent on the level of nutrients (and therefore pollutants) in the lake. Here we can see bistable states, multiple tipping points and path dependence, causing an abrupt transition from one regime to another as the pollution levels are varied.
EMBER is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.