Class Pfqn_rgfmc
THINK TIMES ARE NOT IN EITHER PAPER. Both write the generating function as the RATIONAL prod_i (1 - rho_i z)^-m_i, every node a load-independent single server. An infinite server multiplies it by the ENTIRE exp(sum_r Z_r z_r), which breaks the step Thm 1 rests on: G_n(z') = -sum_i r_i holds only because the residues of n(z)/d(z) sum to zero when deg d >= deg n + 2 (Bertozzi and McKenna, SIAM Review 35(2):239-268, 1993, fact (IV), p. 246), and an exponential numerator does not decay at infinity. The delay is therefore carried by their own repair, eqs. (3.19)-(3.21): only the first k_r+1 Taylor coefficients of exp(Z_r z_r) can reach the coefficient of z_r^k_r, so replacing the exponential by that polynomial is EXACT and leaves a rational integrand. The price is that the eliminated class's population re-enters the term count, which is the population-insensitivity Harrison-Lee Sec. 4 advertises; the class kept for the base case pays nothing.
The elimination is exact in exact arithmetic but is an ALTERNATING sum over residues, so near-coincident loads over an eliminated class destroy significance. The worst cancellation ratio is tracked and the routine REFUSES past maxcancel rather than returning a confidently wrong lG.
-
Nested Class Summary
Nested ClassesModifier and TypeClassDescriptionstatic final classNormalizing constant and its logarithm. -
Field Summary
FieldsModifier and TypeFieldDescriptionstatic final doubleDefault nats of cancellation tolerated before refusing.static final intDefault cap on residue terms carried between eliminations.static final doubleDefault relative tolerance for calling two affine forms proportional. -
Method Summary
Modifier and TypeMethodDescriptionstatic Pfqn_rgfmc.Resultpfqn_rgfmc(Matrix L, double[] N, double[] Z) Exact multiclass normalizing constant by recursion on generating functions.static Pfqn_rgfmc.Resultpfqn_rgfmc(Matrix L, double[] N, double[] Z, double tol, int maxterms, double maxcancel)
-
Field Details
-
DEF_TOL
public static final double DEF_TOLDefault relative tolerance for calling two affine forms proportional.- See Also:
-
DEF_MAXTERMS
public static final int DEF_MAXTERMSDefault cap on residue terms carried between eliminations.- See Also:
-
DEF_MAXCANCEL
public static final double DEF_MAXCANCELDefault nats of cancellation tolerated before refusing.- See Also:
-
-
Method Details
-
pfqn_rgfmc
Exact multiclass normalizing constant by recursion on generating functions.- Parameters:
L- service demand matrix (M x R)N- population vector (R), nonnegative integersZ- think time vector (R)- Returns:
- the normalizing constant and its logarithm
-
pfqn_rgfmc
public static Pfqn_rgfmc.Result pfqn_rgfmc(Matrix L, double[] N, double[] Z, double tol, int maxterms, double maxcancel) - Parameters:
L- service demand matrix (M x R)N- population vector (R), nonnegative integersZ- think time vector (R)tol- relative tolerance for calling two affine forms proportionalmaxterms- cap on residue terms carried between eliminationsmaxcancel- nats of cancellation tolerated before refusing- Returns:
- the normalizing constant and its logarithm
-