Stochastic Tools

The stochastic tools module is a toolbox designed for performing stochastic analysis for MOOSE-based applications. The following sections detail the various aspects of this module that can be used independently or in combination to meet the needs of the application developer.



Distribution objects in MOOSE are function-like in that they have methods that are called on-demand by other objects and do not maintain any state. A custom Distribution object is created in the typical fashion, by creating a C++ class that inherits from the Distribution base class. Three functions are required to be overridden: "pdf", "cdf", and "quantile".

The "pdf" method must return the value of the probability density function (PDF) of the distribution. Similarly, the "cdf" method must return the value of the cumulative distribution function (CDF). Finally, the "quantile" method must return the inverse of the CDF, which is commonly referred to as the quantile function.

For example Listing 1 is the header for the UniformDistribution, which overrides the aforementioned methods.

Listing 1: Header for the UniformDistribution object that includes the three required method overrides for creating a distribution.


#include "Distribution.h"

class UniformDistribution;

template <>
InputParameters validParams<UniformDistribution>();
 * A class used to generate uniform distribution
class UniformDistribution : public Distribution
  UniformDistribution(const InputParameters & parameters);

  Real pdf(const Real & x, const Real & lower_bound, const Real & upper_bound) const;
  Real cdf(const Real & x, const Real & lower_bound, const Real & upper_bound) const;
  Real quantile(const Real & y, const Real & lower_bound, const Real & upper_bound) const;

  virtual Real pdf(const Real & x) const override;
  virtual Real cdf(const Real & x) const override;
  virtual Real quantile(const Real & y) const override;

  /// The lower bound for the uniform distribution
  const Real & _lower_bound;

  /// The upper bound for the uniform distribution
  const Real & _upper_bound;


To utilize a Distribution object within an input file, first the object must be created and secondly an object must be defined to use the distribution. Distribution objects may be created in the input file within the Distributions block, as shown below.

    type = UniformDistribution
    lower_bound = 5
    upper_bound = 10

To use a distribution an object must inherit from the DistributionInterface, which provides to methods:

  • getDistribution<br> This method accepts the name of an input parameter added via a call with the addParam<DistributionName> method. In general, application developers will use this method.

  • getDistributionByName<br> This method accepts the explicitly defined name of a distribution. In general, application developers will not utilize this method.

Each of these methods return a reference to Distribution object, from which you call the various methods on the object as discussed previously.


Sampler objects in MOOSE are designed to generate an arbitrary set of data sampled from any number of Distribution objects.

The sampler operators by returning a vector of matrices (std::vector<DenseMatrix>) from the getSamples method. The application developer is responsible for creating this output as needed depending on the type of sampler.

However, in general, the system is designed such that each row in the matrices represents a complete set of samples that could be passed to sub-applications via the SamplerTransientMultiApp.

Objects, Actions, and Syntax







  • Stochastic Tools App
  • SamplerDataTool for extracting Sampler object data and storing in VectorPostprocessor vectors.
  • StochasticResultsStorage container for stochastic simulation results coming from a Postprocessor.