Comments (2)
  • Add a Comment
  • Edit
  • More Actions v
  • Quarantine this Entry

1 DenisAvrilionis commented Permalink

Excellent posting!

 
A few years ago there was a series of publications in the domain of software engineering dealing with classification of the various models used to build software. Maybe some of the ideas that emerged from these discussions could apply.
 
To further elicit the relationship between analytics and optimization a first rather trivial but extremely important aspect to reinstate is that both techniques aim at building models of the reality. As a few of my friends from the software engineering research community would say, some of these models are “contemplative” while some others are “actionable” (call them also operational/productive/…). As each model exists to serve a specific well-defined goal, there is no point in comparing analytics with optimization. These are all techniques to help us create a, as accurate as possible, view of the real world.
 
It is clear that the peak of maturity of the market is currently shifting from “BI” to “Analytics”, with “business optimization” still being a niche domain. However, the power and the value of optimization models are already proven beyond any doubt. At least this is what we see everyday in the domain of price optimization and revenue maximization where we are active.
 
I am looking forward to reading more post in that subject!

2 JeanFrancoisPuget commented Permalink

Thank you for the comment. The nice thing about optimization is that mathematical optimization models serve two purpose: they describe a reality (what needs to be improved) and they are actionable thanks to mathematical optimization solvers.