Page 1 of 1

PGS method can't work?I need help!

Posted: Thu May 07, 2015 2:14 am
by ljb
I'm using PGS to solve collide response.I ran into a problem.PGS can't approache to the solver.It divergence.
My matrix A = J*M^(-1)*J^(T),The matrix A is SPD Matrix( Symmetric Positive Definite),and PGS methods is convergence.My matrix A is SPD,convergence is poor(Non diagonally dominant).Is there anybody can tell me which matrix is wrong ,or the constraint matrix J wrong?
Please give me some infomation.Thx.

:oops: :oops: :oops: :oops: :oops: