OMNeT++/OMNEST Bug Tracker - OMNeT++
View Issue Details
0000100OMNeT++simulation kernelpublic2009-09-01 09:562011-07-15 15:15
baumgart 
 
normalfeaturealways
newopen 
4.0 
 
0000100: Reordering of run numbers would help to support variable number of repetitions
Currently repetitions of a parameter set lead to consecutive run numbers. This makes it difficult to realize the following workflow:

- Run a parameter study with a small number of repetitions
- Check the results and confidence intervals
- Run the same parameter study with an increased number of repetitions without having to discard previous simulation results.

An easy solution would be to reorder run numbers, so that first all parameter combinations with repetition=0 and after that all parameter combinations with repetition=1 are simulated

 an ini file
No tags attached.
Issue History
2009-09-01 09:56baumgartNew Issue
2009-09-03 17:11andrasNote Added: 0000169
2011-07-15 15:15rhornigNote Added: 0000524

Notes
(0000169)
andras   
2009-09-03 17:11   
Makes sense :) We'll do it.
(0000524)
rhornig   
2011-07-15 15:15   
There is a workaround for this:
- instead of using e.g.
repeat=15

define a dummy iteration variable and set it to act as your repetion counter and set also the measurement-label and repetition-label. Provided that the iteration variabler are numHosts and mean, put the followring BEFORE the definition of these variables (and remove the repeat= line)


**.dummy=${myRep=0..10}
measurement-label = $numHosts=${numHosts}, $mean=${mean}
replication-label = "#${myRep}"
# rest of iteration variables
Aloha.numHosts = ${numHosts=10,20}
Aloha.host[*].iaTime = exponential(${mean=1,4,5..9 step 2}s)