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
May 7, 2013
Four Steps for Thinking About An Optimization Problem
Optimization Can Help you Get Value from your Supply Chain Data - Here’s How to Think About an Optimization Project
Dr. Watson Says:
These four steps will help lead you to better decisions, fewer surprises, and a more successful implementation.
As the amount of supply chain data continues to increase, there is a greater need to use that data to make better decisions. Optimization can help here—whether it is for better planning systems or to even react to new events.
In the previous post, we talked about a definition of optimization and how it can be used in many areas within your supply chain.
In this post, we want to cover the four steps for thinking about an optimization problem. Thinking about these four steps helps you make better decisions when running an optimization project. It will help lead you to better decisions, fewer surprises, and a more successful implementation.
Step 1: Determine the Decisions you Want to Make
The result of an optimization run is a decision you can take action on. For example, you want to determine where to locate a warehouse, you want to know the sequence to run jobs on your lines, or you want to know who should be scheduled to work each shift in the warehouse.
In this step, you want to determine what action the optimization engine should recommend.
Step 2: Quantify the Objective
In this step, you are quantifying how you will know if one solution is better than the other. That is, the first step gives you the action you should take, this step quantifies how one potential solution compares to another. It could be cost, it could be service, or any number of things. The main point of this step is that you do not want to be ambiguous. You want a formula to say how one solution compares to another.
If we are evaluating two sequences for jobs running down the line, we might want to pick the one with the lowest cost. Or, we might want to pick the one with the least amount of set-up time.
Step 3: Determine the Constraints or Limits in the System
In this step, we are defining what makes legitimate solutions. For example, we may want to define a constraint that says we need to meet all customer demand. So, this is not a criteria for comparing solutions, but a requirement for any solution.
Step 4: Understand the Data Available
With an optimization project, you need to keep an eye on the data requirements. Either you need to gather data that you do not yet have or you need to revise earlier steps if some data will not be available.
Final Thoughts
You definitively need to iterate through the steps—decisions at one step can impact another. And, it will be tempting to do multiple steps simultaneously. But, without this framework, it can be very difficult to keep track of the different aspects of the optimization project.
