4 % @brief Iterative PB(k) proportional bounds (Eager-Sevcik)
for single-
class
5 % closed networks with delay.
9function [Xlo,Xhi] = pfqn_pbk(L,N,Z,k)
12 % @brief PB(k):
the Eager-Sevcik proportional (performance) bound at
13 % iteration
count k, i.e.
the level-k member of
the Performance Bound
14 % Hierarchy computed at populations N, N-1, ..., N-k (Casale-Muntz-
15 % Serazzi 2008 cite this as
the iterative extension used in Tables 5,7).
16 % Reduces to
the noniterative asymptotic bound at k=0 and tightens to
17 % exact as k->N. Backed by
the validated PBH recursion (pfqn_pbh).
18 % @fn pfqn_pbk(L, N, Z, k)
19 % @param L Service demand vector (M x 1).
20 % @param N Population (scalar).
21 % @param Z Think time (scalar, default 0).
22 % @param k Iteration
count >= 0 (default 1).
23 % @return Xlo Lower throughput bound.
24 % @return Xhi Upper throughput bound.
27if nargin < 3 || isempty(Z), Z = 0; end
28if nargin < 4 || isempty(k), k = 1; end
29[Xlo,Xhi] = pfqn_pbh(L,N,Z,k);