Difference between revisions of "Custom Criteria Optimization"

From MultiCharts
Jump to navigation Jump to search
Line 1: Line 1:
If the standard set of optimization parameters does not suit you, you can create your own custom criteria for optimization.
+
If the standard set of optimization parameters do not suit you, you can create your own custom criteria for optimization.
  
 
There are two ways to create custom criteria for optimization:  
 
There are two ways to create custom criteria for optimization:  
  
 
#Write custom criteria in java on the '''Optimization Criteria''' tab of the '''Exhaustive Search Properties''' window or on the '''Algorithm-Specific Properties''' tab of the '''Genetic Algorith Properties''' window  (see [[Performing Optimization]]); or:  
 
#Write custom criteria in java on the '''Optimization Criteria''' tab of the '''Exhaustive Search Properties''' window or on the '''Algorithm-Specific Properties''' tab of the '''Genetic Algorith Properties''' window  (see [[Performing Optimization]]); or:  
#Create a custom signal using [[SetCustomFitnessValue]] keyword in PowerLanguage Editor specifying your custom criteria.
+
#Create a custom signal using the [[SetCustomFitnessValue]] keyword in the PowerLanguage Editor specifying your custom criteria.
 
<p>
 
<p>
<div style="background-color: #E5F6FF;">Example: You want to optimize your signal by Sharpe Ratio. This parameter is not among standard criteria.  
+
<div style="background-color: #E5F6FF;">Example: You want to optimize your signal by SharpeRatio. This parameter is not among the standard criteria.  
  
One possible solution would be to create a function which calculates SharpeRatio in Power Language (see default SharpeRatio function) and use it as a [[SetCustomFitnessValue]] keyword parameter. See the '''!optimize by sharp ratio!''' default signal as an example. Such signal should be applied to your chart with the signals that you want to optimize applied before performing optimization.  
+
One possible solution would be to create a function which calculates SharpeRatio in Power Language (see default SharpeRatio function) and use it as a [[SetCustomFitnessValue]] keyword parameter. See the '''!optimize by sharp ratio!''' default signal as an example. Such a signal should be applied to your chart with the signals that you want to optimize ''before'' performing optimization.  
  
 
Make sure that you select the '''Custom Fitness Value''' from the list under the '''Standard Criteria''' section on the '''Optimization Criteria''' tab of the '''Exhaustive Search Properties''' window or on the '''Algorithm-Specific Properties''' tab of the '''Genetic Algorith Properties''' window  (see [[Performing Optimization]]).
 
Make sure that you select the '''Custom Fitness Value''' from the list under the '''Standard Criteria''' section on the '''Optimization Criteria''' tab of the '''Exhaustive Search Properties''' window or on the '''Algorithm-Specific Properties''' tab of the '''Genetic Algorith Properties''' window  (see [[Performing Optimization]]).

Revision as of 16:46, 19 May 2014

If the standard set of optimization parameters do not suit you, you can create your own custom criteria for optimization.

There are two ways to create custom criteria for optimization:

  1. Write custom criteria in java on the Optimization Criteria tab of the Exhaustive Search Properties window or on the Algorithm-Specific Properties tab of the Genetic Algorith Properties window (see Performing Optimization); or:
  2. Create a custom signal using the SetCustomFitnessValue keyword in the PowerLanguage Editor specifying your custom criteria.

Example: You want to optimize your signal by SharpeRatio. This parameter is not among the standard criteria.

One possible solution would be to create a function which calculates SharpeRatio in Power Language (see default SharpeRatio function) and use it as a SetCustomFitnessValue keyword parameter. See the !optimize by sharp ratio! default signal as an example. Such a signal should be applied to your chart with the signals that you want to optimize before performing optimization.

Make sure that you select the Custom Fitness Value from the list under the Standard Criteria section on the Optimization Criteria tab of the Exhaustive Search Properties window or on the Algorithm-Specific Properties tab of the Genetic Algorith Properties window (see Performing Optimization).

Note: !optimize by sarp ratio! default signal can be used as an example showing how to find the best parameters using custom criteria.