Assignment 4

(Assignment 4)
Scenario B. GAMMA costs $80k, desired rate 5.5 parts/hr
(NOT REQUIRED)
This case is more interesting. First, from the isolated production rates, we screen out any combinations involving the BETA driller, since these wont achieve a rate faster than 5 parts/hour.
For the remaining combinations, we use the continuous material two-machine line analysis model
in cell1 to compute the average production rate P and average buffer level n
for different values
of the buffer size. Fig. 1 shows the results; we plot the production rate of the line, as a function of
the average buffer level.
6.1
AD,AR
AD,BR
G
Production rate (parts/hour)
6.0
5.9
5.8
5.7
5.6
5.5
5.4
50
100
Average buffer level
150
200
Figure 1: Production rate versus average buffer levels. (Note: AD=alpha drill, AR=alpha reamer,
BR=beta reamer, G=gamma machine.)
Assuming that the machine is operating 24 hours, 365 days a year, the inventory costs for one
year can be found by
IC = 80 n
.
Therefore, the investment costs for each configuration can be determined by finding the minimum
average buffer level that leads to the desired production rate, substituting in the previous equation,
and then adding up the capital costs of the machines. The results are summarized in the table
below.
Configuration
ALPHA Driller + ALPHA Reamer
ALPHA Driller + BETA Reamer
GAMMA
Average buffer level

0
4.5
0
Inventory costs ($)

0
360
0
Total costs ($)

75k
70.36k
80k
Therefore, the optimal configuration consists of ALPHA Driller and BETA Reamer.
To get the lead time, which we consider here as the average time that a part spends from
entering the drilling process to leaving the reaming process, we consider the different stages of the
process. Thus, we have that
Drilling: Wd = 1/6.286 = 0.159 hr
Buffer queue wait time: Wb = 4.5/5.5 = 0.818 hr
Reaming: Wr = 1/5.66 = 0.177 hr
Thus, W = Wd + Wb + Wr = 1.154 hr. (Note that this expression is approximate, since the average
buffer level used in Littles law comes from a continuous material model.)
Problem 2
For this problem, we will use the deterministic processing time long line model. This decision is
motivated by the fact that all 4 machines have the same deterministic processing time (100 parts/day). Note that the parameters that are given to us, pi and ri , are probabilities, not probability
rates. Thus, we can use them directly in our calculations.
(a)
(i) The graph is shown in Fig. 2. To explain its behavior, consider what happens upstream of
buffer 3. When buffer 3 is small, there is a high chance of blockage in this buffer, which in turns
implies that the average levels of buffers 1 and 2 will be high, as parts cannot get through the
system very easily. As you increase N3 and the probability of buffer 3 blockage decreases, the
first three machines become more and more decoupled from machine 4, the lines production rate
increases, and the average levels in buffers 1 and 2 decrease as material flows more easily through
the line. The average level in buffer 3 increases as we make N3 larger, but only up to a point.
This is due to the fact that, when N3 is large, buffer 3 will not tend to be full anymore, because
machine 4 is faster than the upstream three-machine-line. Therefore, increasing N3 any further
has negligible effects on the system. We can also see that the level in buffer 1 is greater than the
level in buffer 2, which follows from the fact that buffer 1 has a smaller upstream impedance and
a larger downstream impedance than buffer 2.
(ii) The resulting plot is shown in Fig. 3. We now have the reverse situation as in the previous
analysis. For small values of N1 , parts cannot flow downstream very easily, and thus the levels in
buffers 2 and 3 are small. As we increase N1 , the production rate increases and the levels in buffers
2 and 3 increase. The average level in buffer 1 will keep rising as we make N1 larger, because
machine 1 is faster than the three-machine-line downstream of buffer 1. Therefore, buffer 1 will
always tend to be full, regardless of its size. Finally, we notice that the average level in buffer 2 is
greater than in buffer 3, which can be explained in the same way as in part (ii).
35
n1
n2
n3
buffer levels
30
25
20
15
10
5
0
0
50
100
150
200
250
300
N3
Figure 2: Change in average buffer levels as N3 is increased.

300
n1
n2
n3
buffer levels
250
200
150
100
50
0
0
50
100
150
200
250
300
N1
Figure 3: Change in average buffer levels as N1 is increased.

(b)
(i) The equation is reasonable, as a smaller mean time to repair (i.e., faster repair rate), will lead
to larger costs. The first term represents the daily revenue; the second term corresponds to daily
costs of storage space; the third term corresponds to the daily cost of holding inventory; the fourth
term corresponds to the daily costs due to repair.
(ii) Let M T T R1 range from 0.01 day to 0.20 day, so that r1 will range from 1.0 to 0.05. Using
cell1, we get the results summarized in Table 2.
We can find the most profitable value of M T T R1 from the plot in Fig. 4.
(iii)
The result does make sense, since it shows the tradeoff between the revenue and the different
cost components. When we decrease M T T R1, the revenue increases due to a larger production
rate, but the repair costs will increase as well. Additionally, a faster machine 1 means that there
4
Table 2: Table with calculations for the profit as M T T R1 is varied.
profit
MTTR1
r1
Production rate
inventory n
1
inventory n
2
inventory n
3
profit ($)
0.01
1.00
82.87
28.67
17.78
12.22
72105.14
0.02
0.50
82.87
28.04
17.78
12.22
77111.45
0.05
0.20
82.85
25.48
17.63
12.18
80118.16
0.08
0.13
82.61
22.81
16.96
11.92
80667.01
0.10
0.10
82.27
21.25
16.28
11.60
80601.64
0.12
0.08
81.80
19.89
15.52
11.21
80317.18
0.15
0.07
80.88
18.17
14.36
10.55
79597.73
0.18
0.06
79.77
16.77
13.26
9.87
78637.41
0.20
0.05
78.97
15.98
12.60
9.44
77908.70
81000
80000
79000
78000
77000
76000
75000
74000
73000
72000
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
0.2
MTTR1
Figure 4: Profit versus M T T R1 (in days).

will be more parts in the system on average, so our holding costs will also increase. On the other
hand, if we decrease M T T R1, our repair costs go down, but we also loose revenue because of a
smaller production rate. The curve in Fig. 4 reflects these tradeoffs and shows that the optimal
value of M T T R1 is 0.085 days.
Problem 3
We start by scheduling production for the heel flange. First fill the row of projected on-hand
inventory, which starts with 25 units at week 0. We note that at week 2 there will be a shortage
of 15 units, so this creates a requirement of that many units by the beginning of week 2. Since we
have exhausted all our initial on-hand inventory, the requirements for the remaining weeks come
from the gross requirements row.
Now, from the net requirements row we can determine the planned order receipts, so that we
guarantee that the demand for each week is covered, while satisfying the lot size constraint (35
units per lot). We finally adjust the timing of the orders, by releasing orders one week before they
are due, because of the lead time (1 week). The resulting MRP table is shown in Fig. 5.
Figure 6 shows the graph of cumulative net requirements, planned order receipts, and planned
releases. Notice that the cumulative planned receipts curve always stays above the net requirements,
and that the order releases curve has a lead of 1 week.
We now schedule the production for Type A. Since we need those parts before we can begin
to produce each heel flange, the planned order release dates of the heel flanges become the gross
Figure 5: MRP table for heel flange production.

requirements for Type A. The table is filled in the same way (see Fig. 7), and the production curves
are shown in Fig. 8. Notice that, although the lead time is 1 week, the scheduled receipt for week 4
is actually released on week 2, since production of Type A parts can only occur in even-numbered
weeks.
Finally, for Type B we need to produce two parts for each heel flange required. Figures 9 and 10
show the MRP table and production plot. We assume that we will produce in a lot-for-lot policy,
i.e., leaving no inventory at the end of any period; therefore, the cumulative net requirements and
cumulative receipts curve coincide in Fig. 10. Notice also that now the receipts lag the releases
curve by three weeks, consistent with the lead time for this part type.
Figure 6: Heel flange production.
Figure 7: MRP table for Type A production.
Figure 8: Type A production.
Figure 9: MRP table for Type B production.
Figure 10: Type B production.

Assignment 4

Uploaded by

Copyright:

Available Formats

Assignment 4

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Assignment 4

Uploaded by

Copyright:

Available Formats

(Assignment 4)

Scenario B. GAMMA costs $80k, desired rate 5.5 parts/hr

Production rate (parts/hour)

Average buffer level

Inventory costs ($)

Total costs ($)

Figure 2: Change in average buffer levels as N3 is increased.

Figure 3: Change in average buffer levels as N1 is increased.

Table 2: Table with calculations for the profit as M T T R1 is varied.

Figure 4: Profit versus M T T R1 (in days).

Figure 5: MRP table for heel flange production.

Figure 6: Heel flange production.

Figure 7: MRP table for Type A production.

Figure 8: Type A production.

Figure 9: MRP table for Type B production.

Figure 10: Type B production.

You might also like