You are on page 1of 5

Part I)

In this part the blocking probability was calculated and compared with that of erlangb formula. The results are shown in the figure below:
x=0:50; semilogy(x,pb1,'b',x,pb2,'g',x,(abs(pb2-pb1)),'r') title('Blocking probability versus offered load') ylabel('blocking probability');xlabel('offered load') legend('simulation','erlangb fromula','difference','location','best')
Blocking probability versus offered load

10

10

-10

10 blocking probability

-20

10

-30

simulation erlangb fromula difference

10

-40

10

-50

10

-60

10

-70

10

15

20

25 30 offered load

35

40

45

50

The following results can be inferred from the above figure:

first of all we see that for offered loads less than 25, blocking probability

decreases extremely and blocking becomes highly unlikely, which is the reason that in the simulation code the blocking probaility is 0(no one was blocked) and as a result is not shown in the logarithmic plot.

Another point to be considered is the closeness of the results of the

simulation and that of the erlangb formula. In those part of the garph which the green line is shown the blue line is exactly in the same coordinate and as a result is covered by the green line, since the green

line has been drawn after the blue line, but in those parts that both lines the maximum difference between the results:
max(abs(pb1-pb2)) ans = 9.9720e-003

are shown there is difference between the results. this claim is verified by

Part II)

The resulting figure is shown below:


x=0:42; semilogy(x,pb,'b',x,pd,'g',x,pt,'r') title('Qulaity of service parameters versus offered load') ylabel('probability');xlabel('offered load') legend('pb','pd','pt','location','best')

10

-1

Qulaity of service parameters versus offered load

probability

10

-2

pb pd pt 10
-3

10

15

20 25 offered load

30

35

40

45

In this part we expext that blocking and dropping probabilities act the same way since no specific effort has been made to differ one from another which is shown in the above figure as well. Part III)

the resulting figure is shown below:


x=0:10; semilogy(x,pb3,'b',x,pd3,'g',x,pt3,'r') title('Quality of service parameters versus dedicated channels') ylabel('probability');xlabel('dedicated channels') legend('pb','pd','pt','location','best')

10

Quality of service parameters versus dedicated channels

10

-1

probability

10

-2

10

-3

10

-4

pb pd pt 0 1 2 3 4 5 6 dedicated channels 7 8 9 10

10

-5

In this part we expect that blocking and dropping probabilities increase and decrease, respectively as the number of dedicated requests increase. the number of dedicated channels and becomes zero for N>4 which is In the above figure we see that dropping probabilty falls with increase in not shown in the logarithmic plot. The behavior of blocking probability

blocking probability, Pt is mainly determind by Pb. Part IV)

combination of Pb and Pd and since dropping probability is far less than

is as it's supposed to be, which is increasing. and the behavior of Pt is a

The results are shown below:


x=0:4; semilogy(x,pb4,'b',x,pd4,'g',x,pt4,'r') title('probability versus queue length') ylabel('probability');xlabel('queue length')

legend('pb','pd','pt','location','best')
probability versus queue length

10

-1

10

-2

prob ability

pb pd pt 10
-3

10

-4

10

-5

0.5

1.5

2 2.5 queue length

3.5

From the figure we see that considering the queue yields to reduction of again like the previous part Pt is mainly affected by the Pb due to major queue length mean waiting time decreases.

the dropping probability, and increase in the blocking probability, and difference in their magnitudes. also we expect that with increase in the

You might also like