====The Structure of Hosted Models====

ARDI ModelHost actually allows you to run //several// models concurrently.

To make it easy to work with and write, we normally suggest spreading complex systems into several distinct models. Having your compressed-air system distinct from your electrical system, or your boilers distinct from your chillers, will help you compartmentalise your work. 

The models can communicate with one-another, so the //output// of one model can then be the //input// to another, letting you combine your individual pieces into a complete system. 

Each individual **model** includes four types of [[Data Points|data point]].

^Type Name^Used For^
|**[[input|Inputs]]**|Inputs are the values that are fed into your model. The values are expected to vary and may be linked to ARDI properties or data points in other models|
|**[[constant|Constants]]**|These are the assumptions, the 'fixed' values in your model. These will only vary occasionally, and are more likely to be set once and then forgotten|
|**[[output|Outputs]]**|These are the results of your calculations, and are the values most likely to be interesting to the user|
|**[[calc|Calculations]]**|These are very similar to Outputs, but aren't made available to end-users. These are used for intermediate steps in your calculation that might not be interesting to users or are hidden for IP reasons.|

[[Creating a Model in Python]]