5 ms·

The Life and Times of the Father of Linear Programming (2005)

31 points by the_d00d 12y ago

dminor 12y ago
In my first job out of college we used linear programming to solve combinatorial auctions - two things I had no idea existed until I interviewed with the company. It was sort of the nexus between OR and econometrics. Very interesting to learn there was one man who helped give birth to both fields.
- eru 12y ago
  You should read about von Neumann and Morgenstern, too.
  - the_d00d 12y ago
    There is a great story mathematicians tell about Von Neumann when he first met Dantzig. From what I remember, Dantzig goes to the chalkboard to explain the finer points of his simplex method and Von Newmann quickly grows bored and interrupts him by saying "oh that". Von Newmann then procedes to point out some deep insights that would only have been obvious to Von Newmann himself. Dantzig later recalls being left speechless by the brilliance of Von Newmann.
philip1209 12y ago
This is great. I had a college course that basically taught only the simplex algorithm - we did all all of the iterations by hand using tableaus and matrices. We are using operations research optimization techniques at my side project, StaffJoy. The greatest innovation in usability of OR has been the JuMP project - there is now a fairly universal way to express optimization problems that is lower-level than Excel and higher-level than C. http://juliaopt.org http://juliaopt.org
doug1001 12y ago
Leonid Kantorovich doesn't deserve even a mention in this article (or in the HN comments)? pretty sure that LK is the one who first developed and applied the technique--and did so with more conviction than most have in their own work: in Russia, during the Siege of Leningrad, no less--he used this new technique to calculate the optimum distance between vehicles carrying food & supplies along the frozen Ladoga river.
- hudibras 12y ago
  Uh, you need to read the whole article.
qwerta 12y ago
There is no mention of Antonin Svoboda. He developed linear computer which would target moving airplanes in pre-nazi Czechoslovakia https://en.wikipedia.org/wiki/Anton%C3%ADn_Svoboda https://en.wikipedia.org/wiki/Anton%C3%ADn_Svoboda His book from 1948 is probably first book about applied programming: https://archive.org/details/ComputingMechanismsLinkages https://archive.org/details/ComputingMechanismsLinkages
- the_d00d 12y ago
  The wikipedia article on Linear Programming doesn't mention him either: https://en.wikipedia.org/wiki/Linear_Programming https://en.wikipedia.org/wiki/Linear_Programming A keyword search inside the link to the publication you shared returns no results for linear programming or simplex. I don't see how your links support your claim that he should have been mentioned in an article devoted to linear programming; which, btw, is different from linear targeting. To be fair, I understand your respect for the man and I don't think anyone would downplay Svoboda's contributions to computer design and mathematical/scientific computing.

31 points by the_d00d 12y ago

In my first job out of college we used linear programming to solve combinatorial auctions - two things I had no idea existed until I interviewed with the company. It was sort of the nexus between OR and econometrics. Very interesting to learn there was one man who helped give birth to both fields.

eru 12y ago
You should read about von Neumann and Morgenstern, too.
- the_d00d 12y ago
  There is a great story mathematicians tell about Von Neumann when he first met Dantzig. From what I remember, Dantzig goes to the chalkboard to explain the finer points of his simplex method and Von Newmann quickly grows bored and interrupts him by saying "oh that". Von Newmann then procedes to point out some deep insights that would only have been obvious to Von Newmann himself. Dantzig later recalls being left speechless by the brilliance of Von Newmann.

philip1209 12y ago

This is great. I had a college course that basically taught only the simplex algorithm - we did all all of the iterations by hand using tableaus and matrices. We are using operations research optimization techniques at my side project, StaffJoy. The greatest innovation in usability of OR has been the JuMP project - there is now a fairly universal way to express optimization problems that is lower-level than Excel and higher-level than C. http://juliaopt.org http://juliaopt.org

doug1001 12y ago

Leonid Kantorovich doesn't deserve even a mention in this article (or in the HN comments)? pretty sure that LK is the one who first developed and applied the technique--and did so with more conviction than most have in their own work: in Russia, during the Siege of Leningrad, no less--he used this new technique to calculate the optimum distance between vehicles carrying food & supplies along the frozen Ladoga river.

hudibras 12y ago
Uh, you need to read the whole article.

qwerta 12y ago

There is no mention of Antonin Svoboda. He developed linear computer which would target moving airplanes in pre-nazi Czechoslovakia https://en.wikipedia.org/wiki/Anton%C3%ADn_Svoboda https://en.wikipedia.org/wiki/Anton%C3%ADn_Svoboda His book from 1948 is probably first book about applied programming: https://archive.org/details/ComputingMechanismsLinkages https://archive.org/details/ComputingMechanismsLinkages

the_d00d 12y ago
The wikipedia article on Linear Programming doesn't mention him either: https://en.wikipedia.org/wiki/Linear_Programming https://en.wikipedia.org/wiki/Linear_Programming A keyword search inside the link to the publication you shared returns no results for linear programming or simplex. I don't see how your links support your claim that he should have been mentioned in an article devoted to linear programming; which, btw, is different from linear targeting. To be fair, I understand your respect for the man and I don't think anyone would downplay Svoboda's contributions to computer design and mathematical/scientific computing.