FAQ
Last updated on 16th July 2024
- It seems like my model is being run when the server starts before requesting for it to be run?
The server runs some health checks on server startup. Read more about this here. - Why do I not see variables newly defined in the console?
When working with iterative development of your models you will likely change the inputs and variables multiple times rapidy when building and working with the console or your own code via the REST API. Modern browsers however are likely to cache changes from your previous builds or changes which could cause issues where code doesn't render, or variables are missing. It's recommend to either use a private/incongnito mode when testing, or to modify your browsers settings such that on close cache is cleared. You are free to keep things like history, passwords as those won't be affected by model changes. - How can I speed up the running of my model?
There are many reasons why your model might be running slowly. Some of these reasons could be the way the model has been programed. If there is a central agent doing most of the compuation, this will be a bottelneck slowing down the entire simulation. If there is any way of splitting it into multiple agents with a final aggregation of sorts, the work can be parrallelised. Another reason the model could be running slowly is that you are outputing too much data. If you do not need agent/link data (if you are not using the console, and don't need the individual agent data output to file), the agent and link data should be turned off.
To stop serialising agent/link data the config fieldscore-abm.serialize.agentsandcore-abm.serialize.linksshould be set to false.
When running a multirun, if the console is not being used, the in memory data should be turned off so that the Java Garbage Collector can work quickly to improve the performance of the model. Read more