SEARCH searchBY TOPIC
right_division Green SCM Distribution
Bookmark us
sitemap
SCDigest Logo

SCDigest Expert Insight: Supply Chain by Design

About the Author

Dr. Michael Watson, one of the industry’s foremost experts on supply chain network design and advanced analytics, is a columnist and subject matter expert (SME) for Supply Chain Digest.

Dr. Watson, of Northwestern University, was the lead author of the just released book Supply Chain Network Design, co-authored with Sara Lewis, Peter Cacioppi, and Jay Jayaraman, all of IBM. (See Supply Chain Network Design – the Book.)

Prior to his current role at Northwestern, Watson was a key manager in IBM's network optimization group. In addition to his roles at IBM and now at Northwestern, Watson is director of The Optimization and Analytics Group.

By Dr. Michael Watson

September 25, 2013



Systems Thinking and the “Limits” of Optimization

When You Model Your Supply Chain, the Goal Isn't to Capture Every Detail


Dr. Watson Says:

start
...you want to build models that are simple enough to capture the key trade-offs.
close
What Do You Say?



Click Here to Send Us Your Comments
feedback
Click Here to See Reader Feedback



Supply Chain Digest recently ran two articles on systems thinking for the supply chain.  These articles drew a lot of attention and Dan Gilmore wrote up some of the interesting reader responses

The reader response that caught my eye was this one: 


 

Trevor Miles of supply chain software vendor Kinaxis sent in a great letter on the topic, in part saying that a downside to some systems thinking is that it created a belief that a full a supply chain system can actually be accurately modeled in software and optimization engines.

"There are so many nuances and competing objectives in a decision that change frequently that I do not believe we can write an objective function for an optimization that captures all the trade-offs that need to be performed," Miles said. "How does one translate customer service objectives, inventory levels, revenue, and margin into a common UOM against which we can run an optimization engine? And all that assuming we know all the variables with sufficient certainty that the optimum is meaningful. Well, we don’t."


Some network design tools have the ability to model and solve multiple objectives.    And, this is certainly a powerful feature that you should be using. 

However, I don’t think Miles was looking for a new feature.   I think he is touching on a more important and deeper modeling issue.  And, that issue is that your model was never meant to be a one-to-one relationship to your supply chain.

Previous Columns by Dr. Watson

The Three Use Cases for Data Scientists

Learn Python, PuLP, Jupyter Notebooks, and Network Design

EOQ Model and the Hidden Costs of Fixed Costs

CSCMP Edge - Nike Quote: "It is All an Art Project Until you Get it on Someone's Feet"

Supply Chain by Design: Why Business Leaders should think of AI as an Umbrella Term

More

I’ve seen many cases where people want the model to be as complex as the real supply and capture every variable.  Then, presumably, when you run the model it gives you more insight than you would get otherwise.  Or, some think, without this complexity, the model is worse that useless—it is better to do nothing.

This is the wrong way to look at models.  First, it will be technically impossible to create such a model.  And, the tendency is to blame the current state of optimization technology.  Second, the real limit isn’t the technology—it is our ability to understand the model and use the model to make decisions. 

When you build a model, you want to build models that are simple enough to capture the key trade-offs and useful for making the decisions you need to make.  Earlier this year, we wrote up a good perspective on model building from Michael Schrage.

As a further case for simplicity, Miles points out the uncertainty in our inputs and variables.  We don’t know what demand will be next year and we don’t know what the transportation costs will be.  Instead of an overly complicated model that assumes these values, we are better off with a simple model that allows us to explore the key trade-offs and see what happens to our supply chain under and wide range of possible outcomes.


Final Thoughts

It is easy to say that we should build simple models that answer the questions we want.  Actually, this is quite difficult.  But, if you spend time thinking about the key trade-offs in your model or spend time building smaller models to capture different trade-offs, you will find you learn a lot more about your supply chain than if you had built a single complicated model.


Recent Feedback

 

No Feedback on this article yet

 

 
.