What we now want to do is to connect these two models. We will do this one step at a time. First we will use the orders generated by the growth in customers to drive orders. Then we will use the production to drive total shipments.
We first delete reference orders. Next, we add total orders and its causes (new orders and replacement orders) from View 1 using the Model Variable tool. Switching to View 1, we cut total orders from the view (highlight with Pointer and Edit>Cut, or Ctrl-X). Back to View 2, we connect total orders to orders received and change the equation for orders received to:
orders received = total orders
Now simulate the model. Instead of using a test input for orders (reference orders), we are now using the output from the first model as the input. (If you are working with the models that come with Vensim please note that the model prod3.mdl will not generate the results that follow. To get these results you need to set total shipments = backlog/normal delivery delay in prod3.mdl or work from prod2.mdl as described above.) We get the results:
Backlog grows for two and a half years then begins to fall off as new orders slow. Production peaks in the third year (lagging a peak in orders in the second year) and there is excess capacity thereafter.
On the consumption side we still have the assumption that total shipments are equal to total backlog divided by normal delivery delay. This means that the diffusion process is unaffected by production capacity limitations. To finish coupling the two submodels, we change this to:
total shipments = production
and modify the diagram appropriately. When we simulate this new structure (be sure to use a new name for your run) we get different behavior:
The peaks in production and orders occur much later and are less extreme. The peak backlog is less than 3 million whereas it was almost 5 million in the previous run. This occurs because the diffusion process is now limited by the rate of capacity expansion.
