/LAGMUL

Block Format Keyword

/LAGMUL - Lagrange Multiplier Option

Description

Describes the Lagrange multiplier option.

Format

(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)
/LAGMUL
Lagmod	Lagopt	Tol		Alpha		Alpha_s

Field	Contents
Lagmod	Conjugates gradient pre-conditioning algorithm Default = 1 (Integer) = 1: Cholesky pre-conditioning = 2: Polynomial first degree pre-conditioning
Lagopt	Lagrange multiplier matrix scaling option Default = 0 (Integer) = 0: No scaling = 1: diagonal scaling = 2: L2 norm matrix
Tol	Convergence criteria Default = 1.E-11 (Real)
Alpha	Iterative shift parameter Default = 0.001 (Real)
Alpha_s	Initial shift value Default = 0.0 (Real)

1.	The Tol value defines a solver precision tolerance for the kinematical conditions treated by Lagrange multipliers:

/BCS/LAGMUL; /GJOINT; /IMPVEL/LAGMUL; /INTER/LAGMUL; /MPC; /RWALL/LAGMUL; and /RBODY/LAGMUL

2.	Alpha and Alpha_s are only used with Cholesky pre-conditioning, and may be used to optimize the matrix factorization speed and quality. These are parameters added to matrix diagonal to avoid failure of incomplete factorization algorithm.

3.	Kinematical conditions in RADIOSS solver may be optionally treated by a Lagrange multiplier method.

These conditions are incompatible with standard kinematical conditions, and a warning is issued in Starter if an attempt to merge different solution methods for the same nodes is made.

Otherwise, all Lagrange multiplier conditions are compatible; except in cases of physical incompatibility.