|
|
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
April 30, 2013
|
|
Don't Let the Term "Optimization" Become a Buzzword
Why You Should Care What Optimization Means, How it Fits in with Analytics, and Is it Real Time?
|
|
Dr. Watson Says: |
 |
| We want to give optimization a more precise definition so you can see the possibilities for using this technology to help your business and to help you evaluate solutions. |
|
What Do You Say?
|
|
|
|
During several Videocasts on SCDigest, several people have asked us how optimization is different from analytics.
I can understand where the confusion comes from.
First, analytics is a term that is used loosely. In a previous SCDigest column, I suggested that it is better to think about analytics as a field composed of descriptive analytics (for reporting and understanding the business), predictive analytics (for forecasting and projecting what might happen), and prescriptive analytics (suggesting what should happen, using optimization technology).
Second, optimization is often loosely used to mean that you are improving something. If we use optimization in this context, it can mean almost anything. We want to give optimization a more precise definition so you can see the possibilities for using this technology to help your business and to help you evaluate solutions.
Optimization technology allows you to model your business with the language of mathematics and then determine the best set of decisions. In the supply chain, it is most often associated with linear or integer programming. Optimization is a rich field with proven methods for coming up the best possible solution. To contrast this, other techniques for finding a solution (known as heuristics) may seem logical, but may not be giving you good answers.
I’ve heard supply chain managers claim that they do “optimization” because they locate their facilities with a network design tool. If you limit optimization to just this, you are missing out on a lot of opportunities.
In a supply chain, optimization technology can help you come up with better scheduling of your production lines (especially when you have issues with sequence dependent set-ups), help you schedule your workforce, and help you determine which items in the warehouse should be closest to the dock doors. By applying optimization technology to these areas, you can save significant money.
Optimization can also help with truck routing. This area is interesting because routing turns out to be a very difficult mathematical problem. In fact, for most large scale routing problems, it is impossible to find the proven optimal solution. Instead, most routing solution use optimization to help come up with answers, but need to supplement with heuristics. With good solutions, you can find 15%-30% reduction in transportation costs and while hitting delivery windows. But, since even good solution providers have to use heuristics, you can see differences in solutions for the same data set.
Optimization can also be embedded in your real-time systems. That is, as orders come in, you can rely on underlying optimization engines to assign the order to the right location and change your picking directions. Or, if a machine breaks down, optimization can help you rebalance the plant.
Final Thoughts
Although optimization problems can be hard to solve, the payoff can be very nice. If you start to look for areas to apply optimization technology in your supply chain, you may find a lot of potential.
|
|
|
|
Recent Feedback |
I found this very informative and helpful. As a Supply Chain management student, optimization is thrown around all too frequently. It is nice to get a clear definition. Optimization has the potential to greatly improve operations and logistics. Great insight!
Taylor Monk
Student
UT Austin, McCombs School of Business
May, 01 2013
|
I can't see in this article how optimization generates benefits to supply chain, but I still don't get why optimization will become a buzzword? Do you mean that managers introduce analytical tools into supply chain but still rely on heuristics to make a decision?
Another question is, how to divert from heuristics to optimization?
Response from Dr. Watson:
Thanks for the comment. Here are some answers to your questions:
Optimization benefits the supply chain by helping make better decisions. Optimization technology allows you to make better decisions by sorting through countless possible solutions to pick the best one. For example, if you are trying to pick the best 5 warehouse locations out of 200 potential locations there are more than 2 billion combinations. You need optimization technology to sort through these combinations.
Optimization can become a buzzword if people use it to just mean "do something better." The danger is that they need miss the ability to apply true optimization technology to a problem.
With optimization and heuristics, it is important to know the difference, to know how different problems in your supply chain are being solved, and to know when one approach might be better than the other. This will help you make better decisions about potential solutions.
Wenhuan(Mike) Wu
MBA student
RPI
May, 02 2013
|
It was good to see you differentiate between heuristics and optimization because many supply chain managers don't know that the two aren't the same thing. In introducing packaging optimization to potential clients, we regularly discuss optimization versus heuristics.
People are always amazed to learn that a 12 count case composed of rectangular primary units has 324 possible combinations of case configuration and pallet patterns. Most users do not have tools to pinpoint the optimal solution thus leaving money on the table.
The heuristics versus optimization consideration is even more complex in outbound e-commerce operations where significant companies will ship over one million unique combinations of weight & cube annually. The obvious question is "how many shipping cases are needed to optimize shipping costs?" Numerous executives mistakenly believe that an internal engineer will figure this out on a spread sheet. One exasperated engineer told his superior that it would take him hundreds of years to generate a manual solution when we were able to identify the optimal solution from 800 million possibilities in less than 90 days with the aid of software.
By the way, my MBA students at University of Massachusetts and Niagara University regularly quote Supply Chain Digest in their supply chain research papers. Keep up the good work.
Jack Ampuja
President
Supply Chain Optimizers
May, 09 2013
|
Dr. Watson,
I think I completely get where you're are coming from, it seems as though you may be calling this an extension of DSS (Decision Support Systems)?
I really do not know how to now differentiate DSS System from your analogy of "Optimization Technology". Are you simply calling DSS a different name? Please advise.
Maybe your perspective may allay the controversy in the industry definition of DSS.
As a SC practitioner, there is always that challenge of transferring common literal meaning to technical platforms or fields.
In my humble opinion, I think we should be able to still use both, literal/technical connotation, but as experts, be cognizant of application, or our intended audience. More so when "OPTIMIZATION" can be perceived as relative.
By the way, that was a brilliant article with a unique perspective.
Akin O
SC Practitioner
LW
Oct, 02 2013
|
|
|
| |
|
|
|
