Procedures

Routine to calculate gradient information by finite difference.

Routine to determine conjugate direction vector or direction of steepest descent for unconstrained function minimization.

Routine to solve one-dimensional search in unconstrained minimization using 2-point quadratic interpolation, 3-point cubic interpolation and 4-point cubic interpolation.

Routine to find first xbar>=eps corresponding to a minimum of a one-dimensional real function by polynomial interpolation.

Routine to solve direction finding problem in modified method of feasible directions.

Routine to solve one-dimensional search problem for constrained function minimization.

Routine to find first xbar>=epscorresponding to a real zero of a one-dimensional function by polynomial interpolation.

Routine to solve special linear problem for imposing s-transpose times s<=1 bounds in the modified method of feasible directions.

Routine to solve constrained or unconstrained function minimization.