Try OpenEdge Now
skip to main content
User's Guide
Configuring and running simulation : Performing other operations : Choosing a randomization method
 

Choosing a randomization method

Business Process Modeler supports two types of randomization in the Simulation perspective; namely, randomizing the time of intervals between process instances; or randomizing the duration of work time for a specific workstep in the process.
*To randomize the time interval between process instances, use the Instances tab of the Properties view for a process, as described in Step 5 in Configuring simulation parameters for a process section.
*To randomize the work time for a specific workstep instance, use the Task tab of the Properties view for a Activity, Adapter, or Subprocess workstep, as described in Step 3 in For worksteps section.
Business Process Modeler provides four probability distribution types, which are described below:
Table 45. Probability distribution types
Distribution Type
Description
Constant
Use this option to maintain the same interval between instances in the simulation. You can define the mean interval between instances in the adjoining boxes in terms of hours, minutes, and seconds. This is used when the duration is fixed, and does not change over time.
Negative Exponential
A continuous probability distribution that is often used to characterize the time between events or the durations of activities. It does not assume a predetermined pattern of distribution. A Negative Exponential distribution is used when the probability decreases over time.
Normal
Use this option to define the symmetrical, bell-shaped distribution pattern that is often used for a simulation. This distribution pattern is determined by its mean and the standard deviation. You can use the adjoining Mean and Std-dev boxes to set the mean and standard deviation in terms of hours, minutes, and seconds. In a typical normal distribution, 70% of results fall between one standard deviation above and one standard deviation below the mean. The standard deviation is a measure of how tightly items are clustered around the mean in a set of data. A low standard deviation means that the results are tightly clustered; a high number that the distribution is widely spread.
Note: If the Standard Deviation setting is more than the Work Time value for a workstep and results in the range of the deviation being a negative value, then only the absolute value of the deviation (not the negative value) is used for running the simulation.
Uniform
A discrete uniform distribution option that represents a uniform distribution pattern within a finite set of possible values. The possible values are set within the minimum and the maximum value specified in hours, minutes, or seconds. This is used when the process instance count (or the worktime duration of a workstep instance) is fixed over a specified range of values.