The Solution
Feasible solution found at step: 0
Variable Value
AR 25.00000
STM 6.000000
STH 0.1000000
FB 0.5000000E-01
NS 5.475485
Solution: EZQUEUE
Because we cannot have fractional servers, we need at least six servers to meet our requirement that no more than 5% of customers find all servers busy.
Suppose that you install a sufficient number of incoming lines so customers finding all servers busy can wait. Further, you will still use six servers. You want to find the following:
1. | the fraction of customers finding all servers busy, |
2. | the average waiting time for customers who wait, |
3. | the average overall waiting time, and |
4. | the average number waiting. |
The following variation on the previous model computes these four statistics:
! Arrival rate of customers/ hour;
AR = 25;
! Service time per customer in minutes;
STM = 6;
! Service time per customer in hours;
STH = STM/ 60;
! The number of servers;
NS = 6;
! The PEL function finds number of servers
needed, NS;
FB = @PEB( AR * STH, NS);
! The conditional wait time for those who wait;
WAITC = 1 / ( NS / STH - AR);
! The unconditional wait time;
WAITU = FB * WAITC;
! The average number waiting;
NWAIT = AR * WAITU;
Note how we now use the @PEB function, rather than @PEL, to account for the presence of a queue to hold callers finding all lines busy. The solution to the modified model is:
Variable Value
AR 25.00000
STM 6.000000
STH 0.1000000
NS 6.000000
FB 0.4744481E-01
WAITC 0.2857143E-01
WAITU 0.1355566E-02
NWAIT 0.3388915E-01
Remember the unit of time is an hour, so the expected waiting time for those who wait is .2857 * 60 = 1.7 minutes.