Next: Calculation of Stationary Probabilities
Up: Introduction
Previous: Outline
Assumptions, Notations, and Objectives
We regard call center as a queuing system including the features for call rate prediction and service optimization.
We assume that
- incoming calls are united into one stream with the call rate
, a call rate is the average number of calls in a time unit
- there are
servers with the same service rate
, the service rate
is the average number of calls which can be served in a time unit,
- each server can serve any call, one at a time
- calls are Poisson with the rate
- service times are exponential with the rate
, in such a case
 |
|
|
(61) |
 |
|
|
(62) |
where
is an average time between the calls
and
is an average service time.
- arriving calls enter the first available server, if all the servers are busy
then the call takes a free waiting place
- there are
waiting places , that means that a call waits if there are not more than
other calls waiting, otherwise the call disappears,
the waiting places are common for all the calls
- the system is stationary, meaning that the parameters
and
are constant.
is the server running cost ( $ per time unit)
is the customer time cost ( $ per time unit)
is the lost customer cost ( $ per lost call)
- the optimal number of servers
minimizes the total cost 
per time unit,
including the server running cost, the customer time cost, and the lost customer cost at fixed parameters
and
- results are presented as the waiting time
distribution functions
,
where
is the probability that the waiting time
will be less than
at fixed number of servers
and waiting places
.
- a family of functions
is defined for different
parameters
and presented in a clear format, assuming that the number of servers
is defined by
minimizing the total cost
.
Next: Calculation of Stationary Probabilities
Up: Introduction
Previous: Outline
mockus
2008-06-21