International Journal of Advances in Science and Technology,

Vol. 4, No.1, 2012

Profit Analysis of Two Non - Identical Parallel System with Failure
under Preferable Inspection Policy and Discrete Distribution
Jasdev Bhatti
1
, Ashok Chitkara
2
, Nitin Bhardwaj
3

1
Research Scholar, Chitkara University, Himachal Pradesh,India.
[email protected]

2
Chancellor, Chitkara University, Himachal Pradesh, India.
[email protected]

3
Asst. Professor, Royal Institute of Technology & Management, Rohtak, Haryana, India.

[email protected]
Abstract
This paper has analyzed probabilistically a system consisting of two non-identical operative parallel
units by using regenerative point technique. A single repair facility is available for repairing both
kinds of failed units. Before the repair of failed automatic unit, it has to go through inspection policy.
But the manual one is free from such policy. The distribution of failure, inspection and repair time
are taken as discrete distribution. Various important measures of system effectiveness such as mean
time to system failure, steady state availability, busy period of supervisor and repairman are obtained.
The profit function, availability and MTSF are also studied graphically.
Keywords: Geometric distribution, Regenerating point technique, MTSF, Availability, Busy period
and Profit function.
1. Introduction
Due to the new researches and advancement in the feild of science, new technologies have been
introduced for the fulfillment of modern needs. These advanced technologies give rise to many
complexities in industrial sector. Many researchers had analyzed reliability models in which the life
time and repair time distribution were taken as exponential distribution. But in some situations when the
observed data are small the discrete failure time distribution are appropriate one as compaired to
continous distribution.
In the field of reliability using discrete distribution Bhardwaj (2008), Gupta (2007) had studied two
identical unit parallel systems with Geometric failure and repair time distributions. In this paper, the
device named Autoclave has been studied.
Autoclave is a device which uses high temperature for destruction of micro organisms. This device is
mainly used for sterilization of media, glasswares, solutions, discarded culture and contaminated
materials in labs and industries mainly pharma industries. For sterilization process moist heat is used and
this moist heat is in the form of Saturated Steam under Pressure.Autoclave is operated approximately 15
lb/in
2
at 121
0
C. It also destroys most of the vegetative and spore forming organisms.
In this paper the inspection and repair time are taken as geometric distribution. Initially both
automatic and manual unit are operative. On the failure of an automatic unit, an inspection is being
performed before being repaired by the repairman. But in case of manual one, direct repairman has been
Special Issue Page 20 of 99 ISSN 2229 5216


International Journal of Advances in Science and Technology,
Vol. 4, No.1, 2012



called for repairing the failed unit.Preference will be given to the manual failure as the repair time taken
by the manual one is less as compaired to automatic.
The model is analysed stochastically and the expressions for the various reliability measures of
system effectiveness such as mean time to system failure, steady state availability, and busy period for
both inspector and repairman were obtained.Graphs were also been drawn to analyzed the behaviour of
MTSF , availability and profit function with respect to repair and failure rate.
2. Model Description
The following assumptions are associated with the model:
- A system consists of two non-identical parallel units. Initially both automatic and manual unit
are in operative condition.
- Upon the failure of an automatic unit, inspection policy is being introduced for inspecting the
failed unit. But the manual one is free from such policy.
- The system is assumed to be in the failed state when both units were together in failed
conditions.
- A single repairman is available to repair both types of failed unit but the preference will be
given to the manual unit on the automatic one.
- A repaired unit works as good as new.
2.1 Nomenclature
Table 1: Nomenclature
O : Unit is in operative mode
A
o
: Automatic unit is in operative mode
M
o
: Manual unit is in operative mode.
M
r
/ M
W
: Manual unit is in failure mode and under repair / waiting for repair.
A
i
: Automatic unit is in failure mode and under inspection.
A
r
/ A
w
: Automatic unit is in failure mode and under repair /waiting for repair.
p
1 /
q
1
: Probability that automatic unit goes to failed state or not.
p
2 /
q
2
: Probability of the failed automatic unit to be inspected satisfactory or not.
p
3 /
q
3
: Probability that manual unit goes to failed state or not.
r
: Unit is under repair.
q
ij
(t) /

Q
ij
(t) : p.d.f and c.d.f of first passage time from regenerative state i to regenerative
state j.
P
ij
(t) : Steady state transition probability from state S
i
to S
j
.
i

: Mean sojourn time in state S
i
.


Special Issue Page 21 of 99 ISSN 2229 5216


International Journal of Advances in Science and Technology,
Vol. 4, No.1, 2012



Up States
S
0
(A
O
, M
O
), S
1
(A
O
, M
r
), S
2
(A
i
, M
O
), S
4
(A
r
, M
O
)
Down State
S
3
(A
i
, M
r
), S
5
(A
rw
, M
r
)


Figure 1: Transition Diagram
3. Transition Probabilities and Sojourn Times
Q
01
(t) =
( )
3 1
) 1 (
3 1 3 1
1
] 1 [
q q
q q p q
t

+
Q
02
(t) =
( )
3 1
) 1 (
3 1 3 1
1
] 1 [
q q
q q q p
t

+

Q
03
(t) =
( )
3 1
) 1 (
3 1 3 1
1
] 1 [
q q
q q p p
t

+

Q
10
(t) =
( )
1
) 1 (
1 1
1
] 1 [
sq
sq rq
t

+

Q
12
(t) =
( )
1
) 1 (
1 1
1
] 1 [
sq
sq rp
t

+
Q
13
(t) =
( )
1
) 1 (
1 1
1
] 1 [
sq
sq sp
t

+


Q
23
(t) =
( )
3 2
) 1 (
3 2 3 2
1
] 1 [
q q
q q p q
t

+
Q
24
(t) =
( )
3 2
) 1 (
3 2 3 2
1
] 1 [
q q
q q q p
t

+

Q
25
(t) =
( )
3 2
) 1 (
3 2 3 2
1
] 1 [
q q
q q p p
t

+

Q
32
(t) =
( )
2
) 1 (
2 2
1
] 1 [
sq
sq rq
t

+

Q
34
(t) =
( )
2
) 1 (
2 2
1
] 1 [
sq
sq rp
t

+

Q
35
(t) =
( )
2
) 1 (
2 2
1
] 1 [
sq
sq sp
t

+

Q
40
(t) =
( )
3
) 1 (
3 3
1
] 1 [
sq
sq rq
t

+

Q
41
(t) =
( )
3
) 1 (
3 3
1
] 1 [
sq
sq rp
t

+

Special Issue Page 22 of 99 ISSN 2229 5216


International Journal of Advances in Science and Technology,
Vol. 4, No.1, 2012



Q
45
(t) =
( )
3
) 1 (
3 3
1
] 1 [
sq
sq sp
t

+

Q
54
(t) =
s
s r
t

+
1
] 1 [
) 1 (
(1-16)
The steady state transition probabilities from state S
i
to S
j
can be obtained from
P
ij
=
t
lim Q
ij

It can be verified that
P
01
+ P
02
+ P
03
= 1, P
10
+ P
12
+ P
13
= 1, P
23
+ P
24
+ P
25
= 1,
P
32
+ P
34
+ P
35
= 1, P
40
+ P
41
+ P
45
= 1, P
54
= 1. (17-22)
3.1 Mean Sojourn Times
Let T
i
be the sojourn time in state S
i
(i = 0, 1, 2, 3, 4, 5), then mean sojourn time in state S
i
is given
by

=
> = =
0
) ( ) (
t
i i i
t T P T E

so that

0
=
,
1
1
3 1
q q

1
=
1
1
1
sq

2
=
,
1
1
3 2
q q

3
=
,
1
1
2
sq

4
=

,
1
1
3
sq

5
=
,
1
1
s
(23-28)
Mean sojourn time (m
ij
) of the system in state S
i
when the system is to transit into S
j
is given by
m
ij
=

=0 t
ij
) t ( q t

m
01
+ m
02
+ m
03
= q
1
q
3

0,
m
10
+ m
12
+ m
13
= sq
1

1,
m
23
+ m
24
+ m
25
= q
2
q
3

2
,
m
32
+ m
34
+ m
35
= sq
2

3,
m
40
+ m
41
+ m
45
= sq
3

4
, m
54
= s
5
. (29-34)

4. Reliability and Mean Time to System Failure
Let R
i
(t) be the probability that system works satisfactorily for atleast t epochs cycles when it is
initially started from operative regenerative state S
i
(i = 0, 1, 2, 4).
R
0
(t) = Z
0
(t) + q
01
(t 1) R
1
(t1) + q
02
(t 1) R
2
(t1).
R
1
(t) = Z
1
(t) + q
10
(t 1) R
0
(t1) + q
12
(t1) R
2
(t1).
R
2
(t) = Z
2
(t) + q
24
(t1) R
4
(t1).
R
4
(t) = Z
4
(t) + q
40
(t1) R
0
(t1) + q
41
(t1) R
1
(t1). (35-38)

Taking geometric transformation on both sides, we get

) (
) (
) (
1
1
0
h D
h N
h R =

The mean time to system failure is
Special Issue Page 23 of 99 ISSN 2229 5216


International Journal of Advances in Science and Technology,
Vol. 4, No.1, 2012




i
=
1
1
1
1
1 h D
N
1
) h ( D
) h ( N
lim =

where
N
1
= (
0
-1) (1- P
12
P
24
P
41
) + (
1
+ P
10
) (P
01
+ P
02
P
24
P
41
) + (P
02
+P
01
P
12
) (
2
+
4
P
24
+ P
24
P
40
).
D
1
= (1- P
12
P
24
P
41
) - P
10
(P
01
+ P
02
P
24
P
41
) - P
24
P
40
(P
02
+P
01
P
12
). (39-40)
5. Availability Analysis
Let A
i
(t) be the probability that the system is up at epoch t when it is initially started from
regenerative state S
i
by simple probabilistic argument the following recurrence relations are obtained.
A
0
(t) = Z
0
(t) + q
01
(t1) A
1
(t1) + q
02
(t1) A
2
(t1) + q
03
(t1) A
3
(t1).
A
1
(t) = Z
1
(t) + q
10
(t1) A
0
(t1) + q
12
(t1) A
2
(t1) + q
13
(t1) A
3
(t1).
A
2
(t) = Z
2
(t) + q
23
(t1) A
3
(t1) + q
24
(t1) A
4
(t1)+ q
25
(t1) A
5
(t1).

A
3
(t) = q
32
(t1) A
2
(t1) + q
34
(t1) A
4
(t1) + q
35
(t1) A
5
(t1).
A
4
(t) = Z
4
(t) + q
40
(t1) A
0
(t1) + q
41
(t1) A
1
(t1)+ q
45
(t1) A
5
(t1).
A
5
(t) = q
54
(t1) A
4
(t1). (41-46)
By taking geometric transformation and solving the equation

) (
) (
) (
2
2
0
h D
h N
h A =

and
i
Z (h) =
i

The steady state availability of the system is given by
A
0
=
t
lim A
0
(t)
Hence, by applying L Hospital Rule, we get
A
0
= -
) 1 (
) 1 (
2
2
D
N
'

where
N
2
(1) = (1 P
23
P
32
)[
0
( P
40
+ P
41
P
10
) +
1
( P
41
+ P
40
P
01
)

+
4
( 1 P
10
P
01
)] +
2
[(P
02
+ P
03
P
32
)
(P
40
+ P
41
P
10
)

+ (P
12
+ P
13
P
32
) (P
41
+ P
40
P
01
) ]. (47)
2
D' (1) = {(1 P
23
P
32
) [
0
(P
40
+ P
41
P
10
) +
1
(P
41
+ P
40
P
01
)

+
4
(1 P
10
P
01
)] +
2
[(P
02
+ P
03
P
32
)
(P
40
+ P
41
P
10
)

+ (P
12
+ P
13
P
32
) (P
41
+ P
40
P
01
)] +
3
[(P
03
+ P
02
P
23
) (P
40
+ P
41
P
10
)
+ (P
13
+ P
12
P
23
)

(P
41
+ P
40
P
01
)] +
5
{ P
45
(1 P
23
P
32
) (1 P
10
P
01
) +

(P
40
+ P
41
P
10
)[ P
02
(P
25
+
P
23
P
35
) + P
03
(P
35
+ P
32
P
25
)] + (P
41
+ P
40
P
01
) [

P
12
(P
25
+ P
23
P
35
) + P
13
(P
35
+ P
32
P
25
)]}} (48)

Special Issue Page 24 of 99 ISSN 2229 5216


International Journal of Advances in Science and Technology,
Vol. 4, No.1, 2012



6. Busy Period Analysis
6.1 Busy Period of Inspector
Let B
i
(t) be the probability of the inspector who inspect the failed automatic unit before being
repaired by repairman.Using simple probabilistic arguments, as in case of reliability and availability
analysis the following recurrence relations can be easily developed.
B
0
(t) = q
01
(t1) B
1
(t1) + q
02
(t1) B
2
(t1) + q
03
(t1) B
3
(t1).
B
1
(t) = q
10
(t1) B
0
(t1) + q
12
(t1) B
2
(t1) + q
13
(t1) B
3
(t1).
B
2
(t) = Z
2
(t) + q
23
(t1) B
3
(t1) + q
24
(t1) B
4
(t1)+ q
25
(t1) B
5
(t1).
B
3
(t) = Z
3
(t) + q
32
(t1) B
2
(t1) + q
34
(t1) B
4
(t1) + q
35
(t1) B
5
(t1).
B
4
(t) = q
40
(t1) B
0
(t1) + q
41
(t1) B
1
(t1)+ q
45
(t1) B
5
(t1).
B
5
(t) = q
54
(t1) B
4
(t1). (49-54)
By taking geometric transformation and solving the equation

) (
) (
) (
2
3
0
h D
h N
h B =

The probability that the inspection facility is busy in inspecting the failed unit is given by
B
0
=
t
lim B
0
(t)
Hence, by applying L Hospital Rule, we get
B
0
= -
) 1 ( D
) 1 ( N
2
3
'

where
N
3
(1) =
2
[(P
02
+ P
03
P
32
)( P
40
+ P
41
P
10
)

+ ( P
12
+ P
13
P
32
) ( P
41
+ P
40
P
01
) ]
+
3
[(P
03
+ P
02
P
23
) ( P
40
+ P
41
P
10
) +( P
13
+ P
12
P
23
)

( P
41
+ P
40
P
01
)]. (55)
and
) 1 ( D
2
' is the same as in availability analysis.
6.2 Busy Period of Repairman
Let ) (
'
t B
i
be the probability that the repairman is busy in repair of failed unit when the system
initially starts from regenerative state S
i
. Using simple probabilistic arguments, the following recurrence
relations can be easily developed.
) (
'
0
t B = q
01
(t1)
) 1 (
'
1
t B
+ q
02
(t1) ) 1 (
'
2
t B + q
03
(t1) ) 1 (
'
3
t B
) (
'
1
t B = Z
1
(t) + q
10
(t1) ) 1 (
'
0
t B + q
12
(t1) ) 1 (
'
2
t B + q
13
(t1) ) 1 (
'
3
t B .
) (
'
2
t B
= q
23
(t1)
) 1 (
'
3
t B
+ q
24
(t1) ) 1 (
'
4
t B + q
25
(t1)
) 1 (
'
5
t B
.
) (
'
3
t B = Z
3
(t) + q
32
(t1) ) 1 (
'
2
t B + q
34
(t1) ) 1 (
'
4
t B + q
35
(t1) ) 1 (
'
5
t B .
) (
'
4
t B = Z
4
(t) + q
40
(t1)
) 1 (
'
0
t B
+ q
41
(t1) ) 1 (
'
1
t B + q
45
(t1) ) 1 (
'
5
t B .
) (
'
5
t B = Z
5
(t) + q
54
(t1) ) 1 (
'
4
t B . (56-61)
Special Issue Page 25 of 99 ISSN 2229 5216


International Journal of Advances in Science and Technology,
Vol. 4, No.1, 2012



By taking geometric transformation and solving the equation

) (
) (
) (
2
4 '
0
h D
h N
h B =

The probability that the repairman is busy in repairing the failure of failed unit is given by

'
0
B
=
t
lim ) (
'
0
t B

Hence, by applying L Hospital Rule, we get

'
0
B
= -
) 1 ( D
) 1 ( N
2
3
'

where
N
4
(1) = (1 P
23
P
32
)[
1
( P
41
+ P
40
P
01
)

+
4
( 1 P
10
P
01
)] +
3
[(P
03
+ P
02
P
23
) ( P
40
+ P
41
P
10
)
+ ( P
13
+ P
12
P
23
)

( P
41
+ P
40
P
01
)] +
5
{P
45
(1 P
23
P
32
) ( 1 P
10
P
01
) +

( P
40
+ P
41
P
10
)[ P
02
(P
25
+
P
23
P
35
) + P
03
(P
35
+ P
32
P
25
)] + ( P
41
+ P
40
P
01
) [

P
12
(P
25
+ P
23
P
35
) + P
13
(P
35
+ P
32
P
25
)]} (62)
and
) 1 ( D
2
'
is the same as in availability analysis.
7. Profit Function Analysis
The expected total profit in steady-state is
P = C
0
A
0
C
1
B
0
C
2
'
0
B (63)

where
C
0
: per unit up time revenue by the system.
C
1
& C
2:
per unit down time expenditure on the system during inspection and repair.
8. Graphical Representation
The behaviour of the MTSF, availability and the profit function w.r.t failure rate and repair rate have
been studied through graphs by fixing the values of certain parameters C
0
, C
1
and C
2
as
C
0
= 2000, C
1
= 400 and C
2
= 600.
On the basis of the numerical values taken as:
r = 0.15 and s = 0.85
The values of various measures of system effectiveness are obtained as:

Mean time to system failure (MTSF) = 10.22329.
Availability (A
0
) = 0.620626.
Busy period of Inspector (B
0
) = 0.080572.
Busy period of repairman (
'
0
B ) = 0.829718.
Profit = 711.19331
Figure: 2 show the behavior of MTSF w.r.t failure rate (p
1
) for different values of repair rate (r).It
appears from graph that MTSF decreases with increase in failure rate.
Figure: 3 show the behavior of profit function w.r.t failure rate (p
1
) for different values of repair rate
(r).It appears from graph that profit decreases with increase in failure rate.
Figure: 4 show the behavior of availability w.r.t failure rate (p
1
) for different values of repair rate (r).It
appears from graph that availability decreases with increase in failure rate.
Special Issue Page 26 of 99 ISSN 2229 5216


International Journal of Advances in Science and Technology,
Vol. 4, No.1, 2012



Figure: 5 show the behavior of MTSF w.r.t repair rate (r) for different values of failure rate (p
1
). It
appears from graph that MTSF increases with increase in repair rate.
Figure: 6 show the behavior of profit function w.r.t repair rate (r) for different values of failure rate (p
1
).
It appears from graph that profit increases with increase in repair rate.
Figure: 7 show the behavior of availability function w.r.t repair rate (r) for different values of failure
rate (p
1
). It appears from graph that availability increases with increase in repair rate.

9. Conclusion

This paper provided the numerical results for MTSF, availability and busy period of repairman and
inspector which concluded that the preventive maintenance of units increases both the availability and
profit of the system. It also provides information which will be useful for other researchers and
companies following such systems to prefer the equipments which satisfied the conditions as discussed.



Figure 2. MTSF vs Failure Rate

Figure 3. Profit vs Failure Rate

Special Issue Page 27 of 99 ISSN 2229 5216


International Journal of Advances in Science and Technology,
Vol. 4, No.1, 2012




Figure 4. Availability vs Failure Rate

Figure 5. MTSF vs Repair Rate

Figure 6. Profit vs Repair Rtae
Special Issue Page 28 of 99 ISSN 2229 5216


International Journal of Advances in Science and Technology,
Vol. 4, No.1, 2012





Figure 7. Availability vs Repair Rate
10. References
[1] Bhardwaj N, Kumar A, Kumar S (2008) Stochastic analysis of a single unit redundant system with two
kinds of failure and repairs, Reflections des. ERA-JMS, Vol. 3 Issue 2 (2008), 115-134.
[2] Kumar A, Bhardwaj N (2009) Analysis of two unit redundant system with imperfect switching and
connection time, International transactions in mathematical sciences and Computer,July-Dec. 2009,Vol. 2, No.
2, pp. 195-202.
[3] Gupta R, Varshney G (2007) A two identical unit parallel syatem with Geometric failure and repair time
distributions, J. of comb. Info. & System Sciences, Vol. 32, No.1-4, pp 127-136 (2007)
[4] Rander, M.C. Kumar S., Kumar A. (1994) Cost analysis of two dissimilar cold standby system with
preventive maintenance and replacement of standbyMicraelectron. Reliability. 34(7), 171-174.
[5] Haggag. M.Y. (2009), Cost analysis of two-dissimilar unit cold standby system with three states and
preventive maintenance using linear first order differential equations J.Math & Stat., 5(4):395-400, 2009, ISSN
1549-3644.
[6] Padgett , Spurrier (1985), On discrete failure models, IEEE Transactions on Reliability, Vol. R-34, No. 3,
1985.
[7] Said K.M.EL, Salah M, Sherbeny EL (2005) Profit analysis fof a two unit cold standby system with
preventive maintenance and random change in Units, 1(1):71-77, 2005, ISSN 1549-3644 (2005)
[8] Taneja G, Tyagi V K, Bhardwaj P (2004) Profit analysis of a single unit programmable logic controller
(PLC), Pure Appl. Math. Sci., vol. LX 1-2, September 2004.
.

Author Profile


Mr. Jasdev Bhatti is persuing his Ph.D degree from Chitkara University, Baddi,
Himachal Pradesh, India.He had completed his M.Sc and M.Phill in mathematics
from Himachal University, Shimla, Himachal Pradesh, India.



Your photo
Comes
Here
Special Issue Page 29 of 99 ISSN 2229 5216