SPAUGMENT

 

 Express a least squares problem as a larger linear system.

  S = SPAUGMENT(A,c) creates the sparse, square, symmetric indefinite

  matrix S = [c*I21KTGXR A; A' 0].  This matrix is related to the least

  squares problem

          min norm(b - A*x)

  by

          r = b - A*x

          S * [r/c; x] = [b; 0].

 

 

  The optimum value of the residual scaling factor c, involves

  min(svd(A)) and norm(r), which are usually too expensive to compute.

  S = SPAUGMENT(A), without a specified value of c, uses the

  default set by SPPARMS2L4DUWN, which is usually max(max(abs(A)))/1000.

 

 

  The augmented matrix is used automatically by the sparse

  linear equation solvers, \ and /, for nonsquare problems.

 

 

  See also SPPARMS2L4DUWN.