Calculation section

Calculation section

In the calculation section you define everything that belongs to the calculation like the number of Monte Carlo simulations, the type of random number generator to be used, the number of output digits and so on. Beside the parameters for the calculation you define here the summarizing formulas for your measurend(s) of your simulation. MUSE is able to handle multivariate output by defining more than one output formula. The formulas are encapsulated in an element called measurand. A very simple example with one output formula is the following.

<calculation mcsimulations="100">
  ...
  <measurand>g1 + g2</measurand>
</calculation>

Here the variables g1 and g2 are added. The result is a one dimensional output file. A more Advanced example follows where we have three output quantities.

<calculation mcsimulations="100">
  ...
  <measurand>
    <formula name="s1">m - mt / V1 + V2</formula>
    <formula name="s2">m + mt / V3 + V4</formula>
    <formula name="s3">m + mt / V5 + V6</formula>
  </measurand>
</calculation>

The calculation section above defines three formulas named s1, s2 and s3. The formulas are defined in infix notation. MUSE transforms this formulas into postfix notation internally. To avoid this conversion you can define the formulas directely in postfix notation. For details see the section formula format of MUSE. MUSE comes with several mathematical functions like the square root, trigonometrical functions and also some predefined constants like pi. For details see the list of keywords.

Attributes for calculation section

The following attributes can be set in the calculation section:

See the section adaptive MC for setting the attributes for adaptive MC.

Variables and variation loops

[Variables] and variation loops are used to set global parameters used in the calculation of the measurement uncertainty. They can be used for shared values in processes and instances.

Variables

As variables can be defined in the calculation section as well as in the process section, we refer to the section [Variables].

Simple for-loop

Often you want to compare different outputs with different values for a parameter/variable or you want to define sequential measurements. In MUSE you use variation variables for that.

<variation name="varname" from="1" to="10" step="2"/>

With this simple line you will simulate five different scenarios. Each variation represents one value of the variable varname. The values for varname will be (1, 3, 5, 7, 9).

Variation list

Sometimes you are not interested in fixed interval variations, but you want to define for example when your different measurements take place. Therefore we have a construct called variation list, which lets you define exactly this scenario.

<variationlist name="varname">
  <value>8.0</value>
  <value>8.2</value>
  <value>8.5</value>
  <value>9.0</value>
</variationlist>

This can be an abstract definition of time when you did the measurements for 8:00, 8:12, 8:30 and 9:00.

Variation set

The variation list allows you to define almost any scenario. But it can be annoying if you have to use one variation variable, when you want to use a set of different variation variables and think about formulas how to express the dependence of such variables. What you can use in this case are the variation sets. They allow you to define different sets of variables for example for different dates.

<variationset name="varset">
  <set>
    <variable name="time"><value>8.0</value></variable>
    <variable name="temp"><value>18.0</value></variable>
    <variable name="staff"><value>0.4</value></variable>
  </set>
  ...
  <set>
    <variable name="time"><value>9.0</value></variable>
    <variable name="temp"><value>20.0</value></variable>
    <variable name="staff"><value>0.9</value></variable>
  </set>
</variationset>

You can interpret the example like this: at 8:00am you have a temperature of 18°C in the laboratory, the staff is still a bit sleepy and therefore gets just a reliability factor of 0.4; at 9:00am you do your last measurement, the temperature rises to 20°C and the staff is nearly at full concentration with a reliability factor of 0.9. The big advantage of this kind of definition of variation sets is, that you have all your information compressed at one point and the comparison between different sets should be very easy.

Related Topics

Please refer to this pages for more information:

• [Scope]

== More advanced options in calculation section ==
There are more advanced options that can be set and used in the calculation section. Please refer to the following pages for more information.

Settings:

• [Adaptive Monte Carlo]
• [Random number generator]

Parameter sections:

• Fade out of parts of the calculation
• [Analyzing results]
• Validating before simulating
• Defining and working with measurement series