Help: how to work create numerous different arrivals without creating numerous sources

[anninhapgs]

Hi everyone,

I am working on a project that has more than 2,000 different arrival distributions. The distributions are all exponential, but they all change over time.

To illustrate: suppose I am modeling 30 companies around the world, each company with a group of suppliers (or clients), each supplier (or client) with a group of parts (or orders) that will be arriving in my company to be processed. Each one of these parts coming from each supplier to each company (I will call it a combination) has its own distribution that changes over time. [this is an analogy to my problem]

If I had only 1 company with a few suppliers and a few parts,  the natural solution would be to use Rate Tables. This would be good because I want to investigate the impact of different Rate Scale Factors (increase or decrease of arrivals) on my key performance indicators.

However, as I mentioned, I have more than 2,000 different distributions. So, my main question is: does anyone have a suggestion about how to create these arrivals in an efficient manner?

I am assuming creating one source and one table for each possible combination would not be efficient, as it may make the model super heavy. 

I have thought about two different solutions (which I am not sure if it is good and, even worse, I do not know how to implement yet). So, recommendations to any materials, SimBits or tutorials that might provide more information about my ideas are really appreciated.

#1) One idea was using an object inside an object in Simio. This may reduce at least a few combinations?! However, I do not know how to model objects inside objects in Simio.   

#2) Letting Simio auto-create all my servers and auto-fill the information using tables. This means that I would create one source for each combination. Although this may not be computationally efficient, at least I would not have to drag and drop each one of them. But again, I am new to this auto-create object tool and I do not know how to implement it. 

So, I really appreciate:

- any ideas on how to solve my problem in an efficient way

- materials that may teach me how to solve my problem using my idea #1 or #2.

Another issue is that RateTables (as far as I know) cannot be imported and binded. So, how can I input all my 2,000 columns of arrival rate data in Simio? Otherwise, my solution #2 is really bad.

I have also discussed the idea of grouping the arrivals together and using TableRowReferencing and discrete distributions ("mix") to reduce the number of possible combinations. However, the team I am working with does not think this is a good idea, since each distribution change over time in a different way. So, combining them may not be very accurate.

Thanks!