0% found this document useful (0 votes)

122 views

Linear Programming

Linear programming was developed in the late 1940s by George Dantzig while working for the US Air Force to help optimize logistics and supply scheduling. Dantzig devised the simplex method for solving linear programming problems, which uses computers to quickly perform calculations. Testing on thousands of military scenarios showed that nearly all planning problems could be represented with linear programming. Further refinements to the simplex method by Dantzig and others demonstrated it could solve problems with surprisingly few steps. Linear programming now has many applications across business, science, and government for optimization and resource allocation.

Uploaded by

beebee

Available Formats

Download as DOC, PDF, TXT or read online on Scribd

0% found this document useful (0 votes)

122 views

Linear Programming

Uploaded by

beebee

Available Formats

Download as DOC, PDF, TXT or read online on Scribd

You are on page 1/ 2

Linear Programming: A Brief History

Linear Programming is one of the youngest areas of mathematics. In 1939, the Russian
mathematician Leonid Vitalyevich Kantorovich had done some work with optimization
problems. However, the modern concept of Linear Programming seems to have been formulated
in 1947 when George Bernard Dantzig and his colleagues documented a noticeable pattern in a
series of military problems.
Dantzig was working for the US Department of the Air Force on a program called Project
SCOOP (Scientific Computation of Optimum Programs), after World War II. The military
needed to organize and expedite supplies to troops. “Programming” was a “military term that, at
that time, referred to plans or schedules for training, logistical supply or deployment of men.
Dantzig mechanized the planning process by introducing ‘programming in a linear structure’,
where ‘programming’ has the military meaning explained above.” (2)
Dantzig devised what is commonly known as the “simplex method” for solving linear
programming problems. This method was created, in part, to employ the use of computers
(which were in their beginning stages) to do numerous calculations quickly and accurately.
Dantzig (and colleagues) spent the better part of a year assessing thousands of situations taken
from their wartime experience and deciding whether this model could be used to formulate
realistic scheduling problems. They reviewed each of the “ground rules,” for planning and
scheduling, individually and showed that nearly every one could be converted into linear
programming format (with the exception of those involving discreteness and non-convexity.)
It was actually in the summer of 1947 that Dantzig set out to create an algorithm for
solving real planning problems (under pressure from the Air Force). The first thing he observed
was that these problems could be represented graphically, specifically as a polyhedral set. He
noticed that the solutions improved as he moved along the edges of the convex body from one
vertex to the next, eventually reaching the “optimal” solution. At that time, he wasn’t using an
objective function and soon felt that this method was extremely inefficient, so rejected the idea in
search of something better.
In June of this same year, Dantzig got in touch with T.J. Koopmans of the Cowles
Foundation in Chicago. At first Koopmans seemed unresponsive to the presentation but
suddenly became very interested in this linear programming model. “As if he suddenly saw its
significance to economic theory.”(5) In fact, Koopmans was eventually awarded the Nobel Prize
in 1975, the culmination of heading a group of economists who developed the theory of
distribution of resources and its correlation to linear programming.
Leonid Hurwitz, a student Koopmans, helped Dantzig with what they called “climbing up
the beanpole.” It was the predecessor to the simplex algorithm. It was enhanced by eliminating
the convexity constraint and assuming the variables summed to unity. However, Dantzig still
felt that though this model was more efficient, it was probably very impractical.
Dantzig consulted with Johnny von Neumann about solution procedures the following
fall. And surprisingly (having searched in vain for any literature on the subject) received a
lecture on the theory of linear programs. It turned out that von Neumann (and Oskar
Morgenstern) had just completed a text on game theory and suspected that the theory behind
Dantzig’s model was an analogue to the one they had developed for games.
Meanwhile, Dantzig’s group at the Pentagon was testing his simplex method and finding
that it was working very well. According to Dantzig, this was unforeseen. He hadn’t trusted his
intuition in higher dimensions. He thought “that the procedure would require too many steps
wandering from one adjacent vertex to the next.” When in fact, most of the time, his technique
solved problems with m equations in 2m or 3m steps…”truly amazing,” thought Dantzig.

Dantzig wrote a book (over a nine year period) on his work entitled, Linear Programming
and Extensions, published in 1963. He also received many awards for his work.