Package jline.api.cache

Class Cache_t_lrum_map

java.lang.Object

jline.api.cache.Cache_t_lrum_map

public final class Cache_t_lrum_map
extends Object

Method Summary

Modifier and Type	Method	Description
static Matrix	cache_t_lrum_map(Matrix[] D0c, Matrix[] D1c, Matrix m)	Characteristic times for one MAP per item, the MATLAB signature t = cache_t_lrum_map(D0c, D1c, m): D0c[k], D1c[k] are the (d x d) hidden-transition and arrival matrices of item k.
static Matrix	cache_t_lrum_map(MatrixCell[] D0, MatrixCell[] D1, Matrix m)	Characteristic times T_1..T_h equating the expected occupancy of each list to its capacity.
static Matrix	lrummapTime(double[] x, MatrixCell[] D0Matrix, MatrixCell[] D1Matrix, Matrix m, int n, int h)	Capacity residuals m_l - E[occupancy of list l] for times x.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Method Details

cache_t_lrum_map

public static Matrix cache_t_lrum_map(Matrix[] D0c,
 Matrix[] D1c,
 Matrix m)

Characteristic times for one MAP per item, the MATLAB signature t = cache_t_lrum_map(D0c, D1c, m): D0c[k], D1c[k] are the (d x d) hidden-transition and arrival matrices of item k.

cache_t_lrum_map

public static Matrix cache_t_lrum_map(MatrixCell[] D0,
 MatrixCell[] D1,
 Matrix m)

Characteristic times T_1..T_h equating the expected occupancy of each list to its capacity. As MATLAB, the capacity system is ROOT-SOLVED on y = log(T) from y = 0 (fsolve(@capres, zeros(1,h)), MaxIter 1e4, MaxFunEvals 1e5), here by Levenberg-Marquardt with a finite-difference Jacobian; the residual is driven to zero rather than its norm minimised to a loose tolerance. The per-list cells generalise the MATLAB inputs (one MAP per item): D0[k].get(l) is the D0 of item k in list l.

lrummapTime

public static Matrix lrummapTime(double[] x,
 MatrixCell[] D0Matrix,
 MatrixCell[] D1Matrix,
 Matrix m,
 int n,
 int h)

Capacity residuals m_l - E[occupancy of list l] for times x.