RandomNumberGenerator

RandomNumberGenerator is used for generating the individual data points. It supports a wide variety of configurations, which we describe here.

manpy.simulation.RandomNumberGenerator.py

Created on 14 Feb 2013

@author: George

class manpy.simulation.RandomNumberGenerator.RandomNumberGenerator(obj, distribution)

Bases: object

holds methods for generations of numbers from different distributions

Parameters:

• obj – The object that uses RandomNumberGenerator. Only used internally for logging in case of exceptions.

• distribution – Dictionary describing the probability distribution from which the random number should be drawn. See Docs for more details.

Configuring the distribution

The configuration of the RNG’s internal distribution is usually done via several parameters, e.g. of Feature, Machine or Source. As of now, the distribution is described using dictionaries. There exist plans to improve this mechanism, but we don’t have a concrete roadmap.

Generally speaking, the config dictionary describes the distribution name and its parameters. It always has the following format:

distribution = {"DistributionName": {"Parameter1": value1,
                                     "Parameter2": value2,
                                     ...
                                     }
               }

When configuring the underlying probability distribution of a feature, you add such a dictionary as a value to “Feature”.

distribution = {"Feature": {"DistributionName": {"Parameter1": value1,
                                                 "Parameter2": value2,
                                                 ...
                                                }
                           }
               }

The following probability distributions and parameters exist:

Distributions and their respective parameters

Distribution	Parameters
Fixed	mean: specifies the value
Normal	mean and stdev
Exp	mean
Gamma	alpha/shape, beta/rate
Logistic	location, scale
Erlang	alpha/shape, beta/rate
Lognormal	mean, stdev
Weibull	shape
Cauchy	location, scale
Triangular	min, max, mean
Categorical	See tutorial below
General Params	min: minimum output value max: maximum output value

Besides of numbers, RandomNumberGenerator can also return categorical values, which can be useful in industrial settings. To use this functionality, you need to pass “Categorical” as distribution name. In the parameter dictionary, you specify the probabilities of the individual categorical feature values:

distribution = {"Feature": {"Categorical": {"A": 0.2,
                                            "B": 0.7,
                                            "C": 0.1
                                           }
                           }
               }