Art of modeling – choosing the abstraction level
As you become more experienced analytics professional in your career, you extend your expertise beyond the science of modeling into the “art” of modeling as well. There are many aspects to developing the skill of “The art” of modeling but a key area is to determine the abstraction level and thereby the appropriate modeling method by reviewing the problem that needs to be solved.
What is abstraction level and why is it important ?
Abstraction level is the granularity level of your model. It essentially means what level of entities do you want to model in your simulation model. Consider the example of a manufacturing network in the illustration below. I have shown only three levels but you can create more as well. The top level of granularity is more abstract (high level) and as you go down, the granularity (details) increase.
Remember that choosing the right abstraction level is critical to the success of the model of the modeling project. Once you have decided what entities need to be included in the model and the abstraction level, the choice of modeling method and the actual “coding” of the model is quite straightforward.
So you may ask the question -why can’t I use the same, say the most detailed level for all my models ? There can be many reasons but for those who are just getting started- the most important of them is that you can avoid unnecessary complexity in your model by choosing the right abstraction level. Going back to our Manufacturing network modeling example above, if you are creating a model for Manufacturing network capacity planning, you don’t need to model each and every process line and/or job shops on the manufacturing floor.
A key point to remember that abstraction levels are not cast in stone. It is normal (and even desirable) to periodically reconsider the abstraction level in the model development process. The rule of thumb from my experience is that you should start with high abstraction level and add details as they are needed.
Suggested cheat sheet of abstraction levels
In you career as an analytics professional in Supply Chain domain, you will end up developing models of various abstraction levels. As mentioned earlier, the rules of abstraction are not cast in stone and can (and should) vary based on the problem at hand. However, generally, certain types of simulation models will always align with a specific abstraction level.
The cheat sheet below should be able to guide you in your model development process to evaluate whether you are modeling at the right abstraction level.
Only experience makes you an “artist” in analytics domain
So while Science can be taught in schools and coursework, the art aspect of modeling is something that you acquire only through experience. And a personal suggestion is to not be afraid to experiment with the abstraction level.