A system for constructing linear programming models
Shmuel M. Katz, Lawrence J. Risman, Michael Rodeh · IBM Systems Journal · 1980
So far several problems from agriculture have been successfully formulated in LPMODEL. Among these have been the question of how much of each kind of fish to raise in some fishponds, analyses of whether to dry up fishponds or uproot orchards so that the land could be used for raising cotton, and linear programming for entire farms. Both farm managers with and without previous linear pgroamming experience have used LPMODEL. The users have adapted to the notation very rapidly (after only one or two introductory explanations), and the abstract models have helped the farmers perfect the statement of the real constrains on the farms. In these experiments, the typical size of the final matrix has been 10 to 50 variables and less than 100 contraints. Presently, using a workspace of 128K, the APL procedures that solve the linear programming cannot handle much larger matrices, but this restriction will not be present in future versions of LPMODEL.