Endogenizing subscript values

[Louison]

Dear all,

I am using subscripts for a model and I would like to make the values of the subscript elements endogenous. However, when I replace the numbers with the expression of an equation, I get the message "expecting an operator" and I can't figure tout how to solve this problem. Would you have any suggestion ?

Please find enclosed a small model I made to learn how to use subscript in Vensim. Subscript values are defined in the "future consumption" variable. I'd like to replace them by expressions to make them endogenous.

Many thanks for your help.

Louison

[Administrator]

If I understand you correctly, you want the values for "FUTURE CONSUMPTION" to be in an external file? If yes, there is a whole section in the documentation on data (time series)
http://www.vensim.com/documentation/ind ... f_data.htm

If you have constants, you can still use the data way to get values into your model, or use CIN files.
http://www.vensim.com/documentation/ind ... _files.htm
http://www.vensim.com/documentation/ind ... hanges.htm
http://www.vensim.com/documentation/ind ... ?23320.htm

[Louison]

Thank you very much for your reply, I will dig into the documentation you pointed out to.

To be more precise: I don't want the values to be in an external file, I want them to be determined by an equation relating to other variables in the model (hence making them endogenous). In other terms I want the values of the subscript elements to be determined by the model itself and not to be numbers fixed ex ante. Is there a way to do that ? Simply writing an equation instead of the numbers doesn't work (error message "expecting an operator").

[tomfid]

What does the subscript dimension represent? (If it's time, time is already implicit; you don't need a subscript.)

[Administrator]

Thank you very much for your reply, I will dig into the documentation you pointed out to.

To be more precise: I don't want the values to be in an external file, I want them to be determined by an equation relating to other variables in the model (hence making them endogenous). In other terms I want the values of the subscript elements to be determined by the model itself and not to be numbers fixed ex ante. Is there a way to do that ? Simply writing an equation instead of the numbers doesn't work (error message "expecting an operator").

Can you upload the model that has the "expecting an operator" error?

[Louison]

Tom: it represents expected values of consumption at different point of time in the future. We need these expected values to change endogenously and then to feed in other variables and we are struggling finding a way to do it with Vensim. We thought subscripts could do.

Admin: it is the model I attached with my first post, with any expression instead of the numbers in the future consumption variable (I apologize I write from my personal computer and I do not have Vensim DSS on this one to replicate the error).

[Administrator]

Admin: it is the model I attached with my first post, with any expression instead of the numbers in the future consumption variable (I apologize I write from my personal computer and I do not have Vensim DSS on this one to replicate the error).

It works fine for me, I really need to see what you are doing to help.

[tomfid]

I think you need VECTOR LOOKUP then.

[Louison]

I apologize not to be able to upload an example that replicate the error at the moment, I won't be back on my office computer equipped with DSS for a week.

In the current example I uploaded in my first post, the values of the elements of the c_exp subscript are defined in the future consumption variable as numbers: 10,12,14,16,18,20,22,24,26,28,30,32,34,36,38,40,42,44,46,48,50,50,50,50,50,50,50,50,50,50,48,46,44,42,40,39,36,34,32,30,28,26,24,22,20,18,16,14,12,10 (there are 50 elements for the subscript c_exp).

What we would like to do is to replace these numbers with expressions to make these values endogenous. So for instance instead of the numbers we would have, say, delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1),delay(consumption,1) or any other equation (of course in our full model the endogenous values for the subscript elements will differ).

[Louison]

I think you need VECTOR LOOKUP then.

ok thank you Tom, I will look into that when I can access DSS again.

[Administrator]

You can only separate constants by a comma, not equations. But you can add equations (and have one for each subscript element if you really wish).

http://www.vensim.com/documentation/ind ... ?22065.htm

[tomfid]

Is the mapping one to one, i.e. at each present point in time, is there one corresponding future expectation of consumption? In that case, you could use something much simpler, like

expected consumption = SMOOTH( consumption, horizon )

[Louison]

Admin: I hadn't thought about multiple equations to do it but looking again at the user guide, it seems it could be indeed a way to go. So using multiple equation we could define each element of a same subscript by an equation ? That's how I understand the example in the user guide, am I correct ?

Tom: at each present point in time, we would need the whole set of expected consumptions to be computed.

Many thanks for your help and time, I will post here feedbacks as soon as we have tried your suggestions.

[tomfid]

I would hope that the expectations are all structurally the same, so that you could have one subscripted equation, and supply a vector of parameters. as in:
expected C[e] = SMOOTH( consumption, horizon[e] )
horizon [e] = 1,2,3,4,...