Ventana software support forum

This thread is for issues pertaining to the current 64 bit beta DSS release (currently Windows only). The 64 bit release permits running huge models that address more than 2GB of memory, and 64 bit dlls/shared libraries can be called directly from 64 bit processes. First, some notes: The existing .v...

You're right - I should have said payoff = INTEG( - squared error, 0) There's another nuance that's missing from my pseudocode - but hopefully makes sense - which is that contributions to the payoff are only computed where the the data exists (unlike policy payoff contributions, which are computed e...

... or try "File>New Model" to create a blank workspace.

Absolutely. That's how I learn too - a combination of seeing far by standing on the shoulders of giants, and good judgment = accumulated experience from bad judgment.

I'm traveling, but hope to say a bit more about this soon.

A more elegant solution (helpful if you need speed) is to use mapping. I think there's an example in the help system, but here's the idea: subrange : (x1-x100) previous : (x1-x99) -> next next : (x2-x100) -> previous var a[subrange] = .... sum of var a[x1] = var a[x1] sum of var a[next] = sum of var...

Here's the sequence of computation:
http://www.vensim.com/documentation/ind ... quence.htm

Correct. We'll probably change the API when the 64 bit version is finalized (beta release imminent). The limiting factor is that the .vdf format uses floats, so we need to provide a double .vdfx and matching function calls.

Tom

Even if there's no integration, there's a step at time=initial time, which sets up stock initial values (if there were any). That may produce the warning, because the initialization step produces a different random # from the dynamic step. I'm looking into the RNG in general, so there may be a way t...

For a situation in which supply and demand are relatively constant, up to fixed willingness to pay/sell, you probably want a rectangular distribution (ptype=1) with a narrow pwidth (i.e. pwidth is smaller than the variation in prices across suppliers). Another possibility would be to use a linear pr...

Great. Here's the intuition (for all watching this thread): First, a physical (conserved) stock always has to have negative feedback that forces the outflow to zero when the stock is zero, as in the reality check "no inventory -> no sales." If you write, outflow = MIN( stock, desired outfl...

I assume you have: outflow = MIN( water stock {gal} , indicated outflow {gal/yr} ) which is dimensionally inconsistent. What's really going on is that there's some characteristic time that it would take to drain the stock, given a pump or gravity or whatever operating as fast as possible: outflow = ...

It may work with subranges (I don't recall), but I think that's not quite the issue. You have several instances of mixed subscript uses - for example: vector reorder(Desired expansion x hour[HOUR,EXPN2],price order inc[HOUR,TECH]) I think it's checking the size of the TECH family (=PDC) against EXPN...

A function you've used may be expecting a complete range rather than a subrange, and the addition of the new subscript converts TECH to a subrange. Hard to tell without an example though.

Are you using a VECTOR REORDER with PDC on the LHS and Tech on the RHS of the equation?

Hard to tell without seeing the equations in question.

You could try playing around with named ranges. There's an example of usage in the help system for the GET XLS functions. I don't remember off the top of my head whether it's possible to use that to limit the data range, but it might work. The downside is that named ranges are invisible in Excel. Th...

Ahh ... i'd forgotten, but there's a bug in the labeling of the fields in the equation editor, which may have misled you. You have: with lookup.JPG but the WITH LOOKUP on the RHS is not needed, because that's already taken care of by the choice of equation type with the dropdown menus. You need: wit...

Have you tried checking the initialization step results by using Diff integration?

You probably have an extra parentheses in one of the fields in the equation editor. We can't tell without a copy of the model, or perhaps a screenshot of the equation in the editor, so if you can post one of those we can help.

It appears that the problem is the dimensions of your lookup: population growth rate lookup --> Year If you change this to 1/year or fraction/year, it should work. Note that the units on a lookup represent the output, rather than the input. It's often good practice to normalize the inputs and output...

I'm not sure this is what you want though, because the alloc will happen across TECH but not HEXP. If you want to allocate over TECH and HEXP at the same time, you'll have to create a single dimension containing all the [TECH,HEXP] combinations.

The ALLOC functions work across the last subscript, so if you're allocating across producers, you need to reorder Expected supply[HOUR,TECH,HEXP] as Expected supply[HOUR,HEXP,TECH]. See attached.

This is actually a deliberate design decision in PLE, in order to ensure that the diagram corresponds exactly to the equations, and that each variable has a single, primary home view. Pro/DSS provide more flexibility, including text-only models.

There does appear to be a storage or display issue with the GET XLS variable in gaming mode. However, if you add an auxiliary, b=a, it takes on the correct values, so I think the underlying sim is OK. I'll investigate further.

I think what you're seeing is a consequence of the settings in the Scaling tab of the Control Panel. however, I'm not sure you'll be able to make a choice that's satisfactory at all possible times. Is this a Venapp or a regular model?

I'll take a look at your example.