Memory of a Pulse

karankhosla · Post by **karankhosla** » Mon Jun 01, 2009 11:20 am

Hello!

This is based on a problem in Prof Coyle's book (Practical Approach to SD, Chapter 10, Freezing of forecasts).

Lets say I take a sample ("samfcast") of a product demand forecast "fcast" in March every year, this sample is held for 3 months, till June, for accounting purposes, after which, "samfcast" returns to zero.

I am trying to create a variable "DPR" which will take the last observed value of "samfcast", before it returns to zero at the end of June, and use it as the value for DPR carried forward at the start of a new year for the next 12 months. I have tried various approaches, levels, cascading fixed delays, but these are either too complicated or do not produce the desired effect. Any suggestions?

Thank you!

Karan

Lee Jones · Post by **Lee Jones** » Mon Jun 01, 2009 2:29 pm

Have you tried the SAMPLE_IF_TRUE function?

karankhosla · Post by **karankhosla** » Mon Jun 01, 2009 4:48 pm

Just read about it in the manual - it's what I'm looking for! Unfortunately, I haven't bought DSS as yet, but I will in the coming month. Thank you Lee.

LAUJJL · Post by **LAUJJL** » Tue Jun 02, 2009 10:24 am

Hi Karan

I think that sample if true works even in PLE, although being documented as not working. I am not sure.
But the function can be easily simulated.
See the model joined.
But you wil still need DSS version if you want to do really serious modelling.
Regards.
JJ

karankhosla · Post by **karankhosla** » Tue Jun 02, 2009 6:07 pm

Thank you JJ!

Best. Karan