Simulation was disrupted halfway using forward-in-time rmetasim

[paiklinsawat]

Dear Eric,

After your help in fixing the previous issues on rmetasim, we have no problem running skelesim 0.9.8 for 100 reps on the Window10 OS with the dependent packages upgraded. However, in the past month, we've encountered another issue. After calling the library and skeleSimGUI(), there was a pop up saying Windows couldn't find R. We tried to solve this problem by going to "System Properties" --> Advanced --> Environment Variables --> Edit the System variable "Path" by adding the path to R (C:\Program Files\R\R-3.4.3\bin\x64). This temporary solved the problem because R.exe showed up and script.Rout keep storing the data, then finally the file "skeleSim.out" showed up and the whole simulation process was completed.

However, R.exe kept disappearing after running for 5-10 mins, therefore the simulation was always disrupted halfway. We've changed several Window computers, but the same issue remains. Although, this issue was solved once we set the permutation test to 0, it would be better if we could run under the significance test.

Thank you very much for your help.

Best, Pai

[christianparobek]

Hi Pai - So sorry I didn't respond sooner. Are you still having this problem?

[paiklinsawat]

Hello Christian,

We haven’t solved the problem when running skelesim using rmetasim forward-time simulation in Window7OS. R.exe still keeps disappearing after running reps>20 with or without permutation test. All simulation was disrupted halfway either when running reps=100 permutation=10 or reps=20 and change the permutation ranging from permu=0 to permu=10. The file “script.Rout” said error in significance test, so we roughly guess that the problem might involved with inappropriate setting for permutation test, but because we’ve tried a wide range of permutation values, we are not sure whether we are solving the problem in the right direction.

Our first goal is to run 100 repetitions with 10 permutation before diving into 1,000 repetitions 100 permutation for the final round.

Thank you very much for your help. Best, Pai

On Mon, Jan 29, 2018 at 5:29 AM, Christian Parobek <notifications@github.com

wrote:

Hi Pai - So sorry I didn't respond sooner. Are you still having this problem?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/christianparobek/skeleSim/issues/72#issuecomment-361102202, or mute the thread https://github.com/notifications/unsubscribe-auth/AgF2ofN9pL1tQ_seSkKO7PNr4jQvDouEks5tPPTWgaJpZM4Rb8UL .

--

Worata (Pai) Klinsawat (Lecturer)

Conservation Ecology Program, King Mongkut's University of Technology Thonburi (Bang Khun Thian)

49 Soi Thian Thale 25, Bang Khun Thian Chai Thale Road, Tha Kham, Bang Khun Thian, Bangkok 10150 - Thailand

E-mail address: worata@gmail.com, worata.kli@kmutt.ac.th worata@gmail.com Tel: (+66)97-270-1709, (+66)2-470-7559

[paiklinsawat]

Here is the message from script.Rout:

Global analysis:

Error in overallTest(g, nrep = num.perm.reps, stats= stats, quietly = TRUE, :

‘g’ must have more than one stratum defined.

Calls: runSim … globalAnalysis -> lapply ->FUN -> formatOverallStats -> overallTest

Execution halted

Pai

On Mon, Feb 12, 2018 at 3:27 PM, PaiWorata Klinsawat worata@gmail.com wrote:

Hello Christian,

We haven’t solved the problem when running skelesim using rmetasim forward-time simulation in Window7OS. R.exe still keeps disappearing after running reps>20 with or without permutation test. All simulation was disrupted halfway either when running reps=100 permutation=10 or reps=20 and change the permutation ranging from permu=0 to permu=10. The file “script.Rout” said error in significance test, so we roughly guess that the problem might involved with inappropriate setting for permutation test, but because we’ve tried a wide range of permutation values, we are not sure whether we are solving the problem in the right direction.

Our first goal is to run 100 repetitions with 10 permutation before diving into 1,000 repetitions 100 permutation for the final round.

Thank you very much for your help. Best, Pai

On Mon, Jan 29, 2018 at 5:29 AM, Christian Parobek < notifications@github.com> wrote:

Hi Pai - So sorry I didn't respond sooner. Are you still having this problem?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/christianparobek/skeleSim/issues/72#issuecomment-361102202, or mute the thread https://github.com/notifications/unsubscribe-auth/AgF2ofN9pL1tQ_seSkKO7PNr4jQvDouEks5tPPTWgaJpZM4Rb8UL .

--

Worata (Pai) Klinsawat (Lecturer)

Conservation Ecology Program, King Mongkut's University of Technology Thonburi (Bang Khun Thian)

49 Soi Thian Thale 25, Bang Khun Thian Chai Thale Road, Tha Kham, Bang Khun Thian, Bangkok 10150 - Thailand

E-mail address: worata@gmail.com, worata.kli@kmutt.ac.th worata@gmail.com Tel: (+66)97-270-1709, (+66)2-470-7559

--

Worata (Pai) Klinsawat (Lecturer)

Conservation Ecology Program, King Mongkut's University of Technology Thonburi (Bang Khun Thian)

49 Soi Thian Thale 25, Bang Khun Thian Chai Thale Road, Tha Kham, Bang Khun Thian, Bangkok 10150 - Thailand

E-mail address: worata@gmail.com, worata.kli@kmutt.ac.th worata@gmail.com Tel: (+66)97-270-1709, (+66)2-470-7559

[EricArcher]

The error is being generated because the input data for the population structure tests only have one stratum defined, therefore the test statistics can't be conducted. Verify that you have 2 or more strata in each simulation replicate. Are some going extinct?