Difference between revisions of "Algorithms"
Jump to navigation
Jump to search
| Line 4: | Line 4: | ||
===preconditioned gradient solver=== | ===preconditioned gradient solver=== | ||
The preconditioned gradient solver is {{{lmt}}} default solver. It does not require the explicit construction of any mixed model equation, and it therefore less resource demanding than the direct solver. That is, many models which cannot be solved using the direct solver can still be solved using the pre-conditioned gradient solver. The solver has converged to a stable solution if | The preconditioned gradient solver is {{{lmt}}} default solver. It does not require the explicit construction of any mixed model equation, and it therefore less resource demanding than the direct solver. That is, many models which cannot be solved using the direct solver can still be solved using the pre-conditioned gradient solver. The solver has converged to a stable solution if | ||
$$log(\frac{|Ax-b|}{|b|})<t$$ | $$log\left(\frac{|Ax-b|}{|b|}\right)<t$$ | ||
Revision as of 12:55, 4 January 2021
Algorithms
Solving Linear Mixed model Equations
lmt supports two types of solver for solving MME's: a direct solver and a pre-conditioned gradient solver
preconditioned gradient solver
The preconditioned gradient solver is {{{lmt}}} default solver. It does not require the explicit construction of any mixed model equation, and it therefore less resource demanding than the direct solver. That is, many models which cannot be solved using the direct solver can still be solved using the pre-conditioned gradient solver. The solver has converged to a stable solution if $$log\left(\frac{|Ax-b|}{|b|}\right)<t$$