Taguchi Appliaction in Pharmacy

[Ranjit Roy]

Dear Kinan,

I understand, by some means, you have been able to run DOE prescribed
experiments and determine the optimum condition. Now you wish to verify the
predicted performance by running confirmation experiment, but cannot do so
experimentally because of tight production schedule. Your current approach
is to build a simulation model to verify optimum performance for which you
would need performance data at a few design configurations (historical
performance data) which are missing.

Needless to say, any simulation must be validated before you can accept its
output. Conformance (predicted value matching experimental results) with 3
or more actual performance data is ideal. However, in your situation, if the
simulation model (Neural Network) is easily developed, you at least have the
option to verify its performance at the current production configuration. It
is better than nothing, at this stage, but you must try to find other ways
to validate the model by comparing results at other configurations.

I do not prefer any other way that is reliable. Perhaps some of our DOE
discussion group (https://groups.google.com/forum/?hl=en#!forum/TDA-DG )
members will point out other possible means to do so.

Good luck with your research.

- Ranjit Roy 12/11/2013

Ranjit K. Roy, Ph.D., P.E., PMP
3829 Quarton Road, Bloomfield Hills, MI 48302-4059, USA.
www.Nutek-US.com <http://www.nutek-us.com/> Inf...@Nutek-us.com Tel:
248-812-2071
Visit http://nutek-us.com/wp-free.html for FREE literature on the Taguchi
DOE.


From: Kinanti Rakayantias <kinanti_r...@yahoo.com>
Subject: [Thesis Questions] Taguchi Appliaction in Pharmacy
To: "inf...@nutek-us.com" <inf...@nutek-us.com>

Dear Dr. Ranjit K. Roy,

My name is Kinanti Rakayantias, I'm a student of University of Indonesia
majoring Industrial Engineering. Right now, Im working on my thesis, which
talks about the parameter design using Taguchi method in the production of
medicine tablet.

I have a question about confirmation experiment. Due to the production
schedule, I couldn't do the confirmation experiment to verify the result.
I'm thinking about building the model using neural network but unfortunately
the manufacturer I took as my study case dont have any historical data
either. In this kind of case, is there any other thing I can do to verify
the suggested combinations of parameters?

Thank you in advance for any help you might be able to provide.

Kind regards,
Kinanti Rakayantias (Kinan)
+6287-887-519317 | kinanti_r...@yahoo.com