Distributions

The Evacuationz model has several in-built distribution functions that can be used for selecting model parameters. Where selecting a value less than 0 is not logical (e.g. a negative pre-evacuation time), distributions will be truncated at zero. The general form of a distribution element is

<Distribution type="type">
    parameter element/s
</Distribution>

---

The following distributions are available:

Uniform

<Distribution type="enz_uniform">
    <Min>value</Min>
    <Max>value</Max>
</Distribution>

Normal

<Distribution type="enz_normal">
    <Mean>value</Mean>
    <StandardDeviation>value</StandardDeviation>
</Distribution>

Log-normal

<Distribution type="enz_lognormal">
    <Mean>value</Mean>
    <StandardDeviation>value</StandardDeviation>
</Distribution>

Triangular

<Distribution type="enz_triangular">
    <Min>value</Min>
    <Max>value</Max>
    <MostLikely>value</MostLikely>
</Distribution>

Weibull

<Distribution type="enz_weibull">
    <Alpha>value</Alpha>
    <Beta>value</Beta>
</Distribution>

Fixed

The fixed ‘distribution’ is mainly included for consistency and completeness.

<Distribution type="enz_fixed">
    <Value>value</Value>
</Distribution>

Discrete

This function allows the creation of a distribution from a series of discrete values with associated probabilities of their selection and an optional interval range. The csv value list for each element must have the same number of items and the items in the <Percent> element must add up to 100.

<Distribution type="enz_discrete">
    <Bins>csv value list</Bins>
    <Percent>csv value list</Percent>
    <Intervals>csv value list</Intervals>
</Distribution>

---

Truncation

Any of the above distributions can be forced to have an upper and/or lower truncated limit by including the <Upper>value</Upper> and <Lower>value</Lower> elements in the parameter list. For example:

<Distribution type="enz_weibull">
    <Alpha>value</Alpha>
    <Beta>value</Beta>
    <Upper>value</Upper>
    <Lower>value</Lower>
</Distribution>