Hi everyone,

I want to share my perspective and experience on optimization software that I learned since college. Please correct me if my opinion is wrong and any comment is welcomed.

Excel

Advantage: easy to demonstrate the result visually. Also everyone has Excel installed in their laptop. Disadvantage: not very efficient in terms of modeling and solver performance when solving a large-scale (or even mid-scale) problem. Also Excel solver is based on branch-and-bound not branch-and-cut, the performance can be very poor when solving a Mixed-integer programming problem.

Good for business major students or introductory level OR class

AMPL/GAMS

Advantage: Algebraic language enables us to write the model simply like writing the math formulation directly. AMPL also provide script (I am sure GAMS has too but I've never used script in GAMS) for us to code some iterative algorithm such as Bender's decomposition, Lagrangian Relaxation, Column Generation. It's also possible to run batch data when testing the performance of algorithm or heuristic. Easy to do maintenance.

Disadvantage: Once AMPL pass the model and data to the solver such as CPLEX, users have no control on the solver. Aka no lower-level control on the solver. It is impossible to write branch-and-cut via AMPL since it requires to code separation problem at every node (interrupt the solving process) to determine the strongest cut.

However, AMPL enables users to specify "user cut" and it will report how many user cuts are added when terminated. The information is the total user cuts added, so we can't know how many cuts are added at what node. So far I have no ideas how AMPL select these user cuts to add.

C++/Java/Python (basic language) together with CPLEX, Gurobi

Advantage: allows you to have lower-level control on the solver and you have great flexibility on coding/testing any algorithm. This skill set is a must for every OR PhD student in my opinion.

Disadvantage: need to be familiar with those basic language first.

Recently, I am forced to use Lingo since the professor (Non-OR field) only knows Lingo. When I try to promote AMPL/GAMS to him, he claims that many companies are using Lingo, which I am very suspicious about it. Does anyone know the real situation about these optimization software market share?

asked 03 Nov '17, 00:57

leonard239's gravatar image

leonard239
212
accept rate: 0%


link

answered 03 Nov '17, 07:46

AlexFleischer's gravatar image

AlexFleischer
863
accept rate: 10%

Hello! Thanks for sharing the article :-)

(03 Nov '17, 08:53) leonard239

In my experience, I would estimate that the optimization software market is

  • 70% humans still doing it manually
  • 20% in-house algorithms (often First Fit variants)
  • 10% constraints solvers, mostly Cplex, Gurobi and (of course) OptaPlanner.

At this point in time, it's more about terraforming the market than competing for a share of the market.

This is an opinion, not hard data. I'd love to see any market data to dispute or affirm this. But most vendors don't publicly share download numbers or website visits.

link

answered 10 Nov '17, 04:59

Geoffrey%20De%20Smet's gravatar image

Geoffrey De ... ♦
3.6k32764
accept rate: 6%

Thanks for the comment. This is my first time heard OptaPlanner though.

(10 Nov '17, 14:00) leonard239
Your answer
toggle preview

Follow this question

By Email:

Once you sign in you will be able to subscribe for any updates here

By RSS:

Answers

Answers and Comments

Markdown Basics

  • *italic* or _italic_
  • **bold** or __bold__
  • link:[text](http://url.com/ "Title")
  • image?![alt text](/path/img.jpg "Title")
  • numbered list: 1. Foo 2. Bar
  • to add a line break simply add two spaces to where you would like the new line to be.
  • basic HTML tags are also supported

Tags:

×191
×39
×29
×22

Asked: 03 Nov '17, 00:57

Seen: 307 times

Last updated: 10 Nov '17, 14:00

OR-Exchange! Your site for questions, answers, and announcements about operations research.