Error in Units: Smooth Function

Use this forum to post Vensim related questions.
Post Reply
DanielH
Junior Member
Posts: 7
Joined: Wed Jul 21, 2010 5:13 pm

Error in Units: Smooth Function

Post by DanielH »

Dear All,

I am using the smooth function to portray a process where a firm projects the expected market share via an exponential smoothing of the actual market share variable.

Although the model is behaving/simulating normally, I keep getting a "error in units" message when doing the units check.

Input and output (market share) are set as dimensionless, but, Vensim indicates that one of the sides is in 'months'.

Has anyone faced this error before? I am afraid there may be a conceptual issue in the model.

Thanks! regards, Daniel


Analysis of units error:
Right hand and left hand units do not match
#Projected Market Share of Firm 2>SMOOTHI#
Has Units: Dimensionless
INTEG( ( Market Share Firm 2
- #Projected Market Share of Firm 2>SMOOTHI# )
/ ( 3) ,
( 0.5) )
Has Units: Month
Administrator
Super Administrator
Posts: 4846
Joined: Wed Mar 05, 2003 3:10 am

Post by Administrator »

Can you post the actual equation as well?
DanielH
Junior Member
Posts: 7
Joined: Wed Jul 21, 2010 5:13 pm

Post by DanielH »

Thank you; please find the equation below.

Projected Market Share of Firm 2=SMOOTHI(Market Share Firm 2,3,0.5)
Units: Dmnl
DanielH
Junior Member
Posts: 7
Joined: Wed Jul 21, 2010 5:13 pm

Post by DanielH »

Just meant to circle back and say that when inserting the variable as a Level the error in units problem no longer appears. I inserted as a level and then changed the type to auxiliary, (just like the documentation tells to do ;-).

Thanks.
bob@vensim.com
Senior Member
Posts: 1107
Joined: Wed Mar 12, 2003 2:46 pm

Post by bob@vensim.com »

The source of the problems is the use of numbers, instead of named constant variables, in the SMOOTHI function. The explicit equations can more readily accommodate this, so you don't get the error message. You might consider naming the adjustment time and initial value as explicit constants. This both makes the equation clearer, and allows you to easily change the constants when experimenting with the model.
DanielH
Junior Member
Posts: 7
Joined: Wed Jul 21, 2010 5:13 pm

Post by DanielH »

Thanks; I have included the two constant parameters to make the model clearer as per suggested.
Post Reply