Non-Product-Form Networks

Approximations for non-BCMP queueing networks.

The npfqn module provides approximation methods for non-product-form queueing networks that do not satisfy BCMP conditions.

Key Functions

• npfqn_nonexp_approx - handler for non-exponential service and arrival processes
• npfqn_traffic_merge_cs - Given a cell array of n MMAPs with arrivals of R classes, produces a new
• npfqn_traffic_merge - Given a cell array of n MMAPs with arrivals of R classes, produces a new
• npfqn_traffic_split_cs - Given a MMAP, produces a new array after split and class switching

npfqn_nonexp_approx

Description: handler for non-exponential service and arrival processes

Syntax:

[ST, gamma, nservers, rho, scva, scvs, eta] = npfqn_nonexp_approx(method, sn, ST, V, SCV, T, U, gamma, nservers)

Parameters:

Name	Description
method	Solution method or algorithm name
sn	Stochastic network structure
ST	Service time matrix
V	Visit ratios
SCV	Squared coefficient of variation
T	Service time or transition matrix
U	Utilization values
gamma	Item popularity probabilities
nservers	Number of servers

Returns:

Name	Description
ST	Service time values
gamma	Item popularity or rate parameters
nservers	Number of servers
rho	Traffic intensity or utilization
scva	SCV of interarrival times
scvs	SCV of service times
eta	Adjustment factors

npfqn_traffic_merge_cs

Description: Given a cell array of n MMAPs with arrivals of R classes, produces a new

Syntax:

npfqn_traffic_merge_cs(...)

npfqn_traffic_merge

Description: Given a cell array of n MMAPs with arrivals of R classes, produces a new

Syntax:

npfqn_traffic_merge(...)

npfqn_traffic_split_cs

Description: Given a MMAP, produces a new array after split and class switching

Syntax:

npfqn_traffic_split_cs(...)