I am currently enjoying the coursera module on discrete optimization with Minizinc [1]. Optimization modeling always seems to me that it has great potential, but I find it hard to find case studies of real applications, nor do I know anybody that uses it professionally. The case studies I have found are all military based, and I find that a little far from anything I would use it for.
Could all you operations researcher types share your experience? I would also love to hear from anyone who uses optimization modeling or OR in their professional capacity, and about projects.
Please note when I say "real applications," I don't mean a large scale simulation of a standard problem by an academic using made-up data and models. I know that you can use Minizinc to solve the traveling salesman problem, but I want to hear about an actual sales organization -- one that sells stuff (Amazon devs/ analysts, please chime in!).
For "real application" examples: in social services, I could see optimizing case work schedules, application of expensive interventions for best pay off, etc ; then I would want to hear about the actual roll out of this system in county X. In circuits, I could see using optimizing design problems, e.g. getting close to a required time constant using a dozen resistor and capacitor options; then how it was used to improve a robot that is actually produced. In dev ops, I could imagine a model for best machine allocation. Etc.
My two real application of optimizatioin: I used quadratic programming to fit population translation matrices to demographic data [2], and we once used AMPL with NEOS to find estimates of population by household type that would fit a time series with a few years of missing data.
Could all you operations researcher types share your experience? I would also love to hear from anyone who uses optimization modeling or OR in their professional capacity, and about projects.
Please note when I say "real applications," I don't mean a large scale simulation of a standard problem by an academic using made-up data and models. I know that you can use Minizinc to solve the traveling salesman problem, but I want to hear about an actual sales organization -- one that sells stuff (Amazon devs/ analysts, please chime in!).
For "real application" examples: in social services, I could see optimizing case work schedules, application of expensive interventions for best pay off, etc ; then I would want to hear about the actual roll out of this system in county X. In circuits, I could see using optimizing design problems, e.g. getting close to a required time constant using a dozen resistor and capacitor options; then how it was used to improve a robot that is actually produced. In dev ops, I could imagine a model for best machine allocation. Etc.
My two real application of optimizatioin: I used quadratic programming to fit population translation matrices to demographic data [2], and we once used AMPL with NEOS to find estimates of population by household type that would fit a time series with a few years of missing data.
[1] https://www.coursera.org/learn/basic-modeling/
[2] https://escholarship.org/uc/item/47t278gq