====Optimisation====

**Optimisation** is designed to help you find the ideal combination of attributes (that you are Capturing) to achieve a specific target outcome.

For example, you might have a Capture that records your cupcake manufacturing line and has the **oven temperature**, **fan speed**, **conveyor speed** from the ovens, and the **internal cake temperature** from the inspection area.

This system could then be used to try and calculate the ideal oven temp, fan speed and line speed to achieve the perfect temperature.

===Optimising===

First, filter the list for the type of product you're hoping to make. 

Then press the **Optimise** button in the [[the capture list|capture list]] - this will allow you to perform an optimisation.

{{optimise.png}}

You'll need to choose the following...

1) The attribute you're trying to optimise (ie. //Internal Cake Temp//). \\
2) The value you're trying to achieve (ie. //98 Deg C//). \\
3) The inputs you want to set (ie. //Line Speed, Oven Temp, Fan Speed//). \\

Then press **Optimise** to continue.

===Results===

After a few moments, the results will appear.

There will be a list of all of the inputs you chose (ie. //Line Speed, Oven Temp// etc.), followed by the suggested values.

These can be used to guide tests and experiments, refine control or estimate values.

//NOTE: If results from Bayesian Optimisation aren't very useful, you may want to consider using Machine Learning. It's more capable of working with complex, interconnected systems.//

===Usage Notes===

It's important to remember that the 'input' and 'output' of your optimisation don't have to be the inputs and outputs of your **process**.

If you have both //vibration// and //machine speed//, you can use optimisation to both...

* Estimate the **vibration** based on **speed**, or vice-versa, \\
* Estimate the **speed** based on **vibration**.