Page 1 of 1

What is the benefit in paying for LINDO in stochastic optimization?

Posted: Thu May 03, 2018 2:38 pm
by Saku45
The LINDO solver only allows a maximum of 6 samples of a continuous distribution.

This seems a bit strange to be. If I have a continuous distribution, then I could just manually approximate it with a discrete distribution on my own, and then I could write up a stochastic programming model in GAMS using this discrete distribution. Now I don't need to limit myself to 6 samples.

So why does LINDO restrict me to 6 samples? Why do I need to pay for their license to get more than 6 samples, when I can just manually do the discrete approximation
on my own and build a discrete model with greater than 6 realizations of the stochastic variable?

Re: What is the benefit in paying for LINDO in stochastic optimization?

Posted: Thu May 03, 2018 4:26 pm
by bussieck
With a Lindo license you get more than "just" the sample library. It comes with a couple of unique features, e.g. variance reduction, SP Benders, and a full set of commercially supported algorithms (LP, MIP, NLP, Global, ...)

-Michael