search

Open-Systems-Pharmacology / Forum

Discussion forum for the Open Systems Pharmacology Project
65 stars 19 forks source link

problem on multi-dose simulution #441

Closed xixi-1995 closed 4 years ago

xixi-1995 commented 4 years ago

hi, everyone I met one problem when I did multi-dose simulation, image I have tried to simulate lower multi-dose ,it goes well, when I tried to simulate higher multi-dose, it shows error like the picture described, I'm wondering how to solve this problem and how to understand the error I have met? I really need your help

Yuri05 commented 4 years ago

The error means that the underlying ODE system could not be numerically solved. You can try to increase the number of internal solver steps in the solver settings (make sure that Advancedparameter mode is selected). However the predefined value of 100000 is already pretty high; increasing it will with high probability result in the same error after longer computation time grafik

The problem occurs between the last output time point before 1560 min and 1560min. E.g. if you simulate with 4 points/h in this time interval, the problem begins somewhere in the time interval [1545..1560] min. Try to find a time bound within this time interval where the numerical problem begins.

E.g. do something like in the screen below. You will get a new location of the problematic output time (e.g. 1550.1234 min). After that set a simulation time to the last output time point before before the new problematic time (e.g. for my example it would be 1550.1225 min), simulate and try to investigate the simulation results to get a hint what might go wrong. grafik

xixi-1995 commented 4 years ago

@Yuri05 thanks for your patiently reply , I have tried to adjust resolution as you mentioned above, it seems that I need try something more as the end-time is up to 168h. it is noted that I have tried to do a multi-dose simulation with dosing interval is12h,it seems go well, while dosing interval=24h seems to have errors, I'm still confused.

Yuri05 commented 4 years ago

Can you upload your project?

xixi-1995 commented 4 years ago

hi@Yuri05 ,thanks for your help. Here is my PK-Sim file, the simulation of po-60mg-multi-144h-bid goes well.while po-60mg-multi-144h-qd seems to have something wrong. [Uploading Duloxetine.zip…]()

Yuri05 commented 4 years ago

@xixi-1995 you didn't upload the project

xixi-1995 commented 4 years ago

oh sorry, Duloxetine.zip

Yuri05 commented 4 years ago

ok, that's a numerical problem with Weibull formulations with lag time > 0. As a workaround: approximate your Weibull dissolution function with a table. The easiest way to do this:

  1. Right click on the formulation graph and select "Export to Excel" grafik
  2. Create new table formulation and import it from the exported Weibull formulation file grafik
  3. Configure your simulation and replace your Weibull formulation with the new table formulation Duloxetine_Table.zip

@AndreDlm FYI

xixi-1995 commented 4 years ago

@Yuri05 That's quite helpful!! Thank you very much! I'm curious about why and will it be suit for situation similar, I will try more later~