You can use basic models as influences in basic models. In the following example we have a basic model named opengauss that uses another basic model gauss at location distributions to define its influences g1 and g2.
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE model SYSTEM "../../../xml/model.dtd">
<model name="doublegauss" targetid="ag">
<influence id="ag" name="doublegauss">
<formula>g1 + g2</formula>
<influences>
<influence id="g1" name="1st Gauss" model="distributions\gauss"/>
<influence id="g2" name="2nd Gauss" model="distributions\gauss"/>
</influences>
</influence>
</model>
The basic model gauss can for instance look as follows.
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE model SYSTEM "../../../../xml/model.dtd">
<model name="gauss" targetid="og">
<influence id="og" name="gauss">
<formula>myg</formula>
<influences>
<influence id="myg" name="mygauss">
<distribution>
<gauss>
<mu parameter="#mu">0</mu>
<sigma parameter="#sigma">1</sigma>
</gauss>
</distribution>
</influence>
</influences>
</influence>
</model>
For parametrization it is important to see that IDs may not be unique anymore. Therefore you have to provide additional information for the ID if you use basic models in basic models. In the example an initialization would for example look like this.
<initialization>
<!-- Normal distributions for first example -->
<instance name="g" model="doublegauss">
<parameters>
<parameter id="g1.#mu">3</parameter>
<parameter id="g1.#sigma">1</parameter>
<parameter id="g2.#mu">5</parameter>
<parameter id="g2.#sigma">2</parameter>
</parameters>
</instance>
</initialization>
Here we get an instance of basic model doublegauss. As it uses the basic model gauss internally for subinfluences and we have to set the parameters of the distributions there using the id followed by a dot as prefix to the parameter name.
The same holds for fade out defintion.