Replies to This Discussion

Permalink Reply by Bruce C Hartman on June 8, 2010 at 9:42am

I'm not familiar with that package, but they all use optimization techniques. Mixed integer programming is the chief mathematical tool. Look up references to vehicle routing under various constraints, the traveling salesman problem, and packing problems to understand the depth. All of the problems are NP-hard, which means you can't just buy a bigger computer to solve a bigger problem. And the real world problems require that the algorithm produce something feasible-- that is, the planners could go with the current solution-- at any phase in the time consuming search for an optimal solution. Many times optimality cannot be reached in the time allowed for the computing. So naturally there is a lot of OR being done to continually improve the solution techniques, many of which involve heuristics-- methods for guessing a good plan for getting closer to optimal.

Permalink Reply by Yi-Chun Tsai on June 8, 2010 at 9:50am

Thanks Bruce. I understand the the MIP problems are hard to solve already in a deterministic optimization problem. I am also looking for a package that can deal with the stochastic nature of my supply chain in either customer demands or the manufaturing processes. Hence, I am looking at simulation optimization approach. I know there are simulation packages that come with optimization components such as ProModel with SimRunner and Simul8 with OptQuest. What do you think?

Permalink Reply by Andrew J. Blair on June 8, 2010 at 10:48am

I recommend taking an Operations Research course from Dr. Timothy Anderson at Portland State University. He is an excellent teacher. He can walk you through DEA (data envelopment analysis), integer programming, and linear programming. All of these techniques are related. The usefulness of these techniques for supply chain management is demonstrated in MS Excel. Anyway, I found Dr. Anderson's course on the subject to be superb.