# The programs in this directory are "student" versions that # restrict problem sizes (to 300 variables and 300 general # constraints) and are intended for noncommercial use. # For serious commercial use, you should (eventually) license # commercial versions, even if your applications never exceed # the limits of the student versions. For simplicity, these # are 32-bit Power PC binaries, which generally also run on # Intel Macs (albeit more slowly than native Intel binaries). # Should feedback suggest it warranted, Intel student binaries # could be made available. file README file ampl.gz for student ampl processor file gjh.gz for "solver" that computes the objective gradient (g), the constraint , Jacobian (J), and the Hessian of the Lagrangian (H) at the , current point (primal and dual variable values). The solver , message shows commands to read these as params and delete , the temporary file that conveys them. When multiple objectives , are present, the first is used unless the objectives have suffix , objweight ==> g is the gradient of the weighted sum of objectives. , Source = /netlib/ampl/solvers/examples/gjh.c. file lpsolve.gz for LP/MIP solver based on lp_solve 4.0; see ampl/solvers/lpsolve/README file minos.gz for student MINOS solver file snopt.gz for student SNOPT solver