]
Thomas Diesler resolved JBPM-1122.
----------------------------------
Resolution: Won't Fix
Future jBPM3 releases can only contain bug fixes. All effort goes into jBPM4.
Suport multiple simulation runs and apply statistical evaluation of
them
------------------------------------------------------------------------
Key: JBPM-1122
URL:
https://jira.jboss.org/jira/browse/JBPM-1122
Project: JBoss jBPM
Issue Type: Feature Request
Components: Simulation
Reporter: Bernd Ruecker
Priority: Optional
Currently every scenario is only executed once during an experiment. If you want to draw
valid conclusions you need more than one simulation run (as a rule of thumb 30 is a good
number).
If running the same scenario multiple times (with different seeds for the random number
generator) the results of the single runs is no more interesting, the means of the figures
are important.
The tool should support configuring multiple runs in the configuration and provide an
additional result with the means.
These means allow to draw conclusions about probabilities for a special result (for
example "cycle time < 2 hours") with a easy normal distribution (because of
the central limit theorem).
have a look in my master thesis for details
(
http://www.camunda.com/content/publikationen/bernd-ruecker-business-proce...),
there I also implemented it manually for the prototype (source code on page 112).
Very cool would be an automatic detection of the required number of runs if JBPM-1121 is
implemented (
http://jira.jboss.org/jira/browse/JBPM-1121).
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: