22-Mec-B4 Integrated Manufacturing Systems · December 2017
Question 7 of 7: Optimum ratio of men to semiautomatic machines
Nivaar worked solution (AI-drafted; not reviewed by a licensed engineer)
Notes on this paper
Paper format. 16-Mec-B4 Integrated Manufacturing Systems,
December 2017 — a three-hour open-book examination; any non-communicating
calculator is permitted. The cover page states “Any five (5) questions constitute a complete
paper. Only the first five (5) questions as they appear in your answer book will be marked”
and “All questions are of equal value”, so each of the seven printed
questions carries 20 marks against a 100-mark paper. Note 1 invites the candidate to submit a clear
statement of any assumptions made where a question is open to interpretation; this paper needs that
licence twice, and both places are flagged below. All seven questions are worked here, because this
set is a study resource rather than a timed sitting.
Reference texts. E. S. Buffa and R. K. Sarin, Modern Production /
Operations Management, 8th ed. (requirements-schedule lot sizing, economic order interval,
part-period balancing, plant location, machine coupling and the man-machine chart);
R. B. Chase, F. R. Jacobs and N. J. Aquilano, Operations and Supply Chain Management,
16th ed. (aggregate planning strategies, categories of forecasting technique, weighted factor
rating for facility location); M. P. Groover, Automation, Production Systems, and
Computer-Integrated Manufacturing, 5th ed. (CAD geometric transformations, computer-aided
process planning, routing sheets, group technology); D. C. Montgomery, Introduction to
Statistical Quality Control, 8th ed. (Shewhart constants, process capability indices);
A. J. Duncan, Quality Control and Industrial Statistics, 5th ed. (natural tolerance versus
specification); C. E. Ebeling, An Introduction to Reliability and Maintainability
Engineering, 3rd ed. (when preventive maintenance pays).
Question 7: Optimum ratio of men to semiautomatic machines (20 marks)
Operator wage rate $4.90 per hour; machine burden rate $18.00 per
hour; a large number of identical machines, so any integer assignment is available.
Find. The number of machines one operator should tend, and the resulting lowest
cost per unit produced.
Check: the time study does not say which elements stop the machine. The
reading taken here is the physically natural one — the machine is idle only while it is being
unloaded and re-loaded ($l = 0.6 + 3.1 = 3.7$ min), and the operator inspects and packs the piece
he has just removed while the machine is already running the next one, so inspecting, packing and
walking ($w = 2.4+1.9+0.4 = 4.7$ min) are independent work. The alternative reading, that all four
handling elements are charged to the machine ($l = 8.0$, $w = 0.4$), is worked at the end of the
answer: it gives the same optimum of five machines per operator, at
$15.60 per unit instead of $14.24. State whichever reading is
assumed, as Note 1 on the cover page invites.
Approach. Split the operator's work into the part that stops the machine and
the part that does not, compute the theoretical number of machines that exactly fills the
operator's time, then cost the two neighbouring integers — because the machines and the
operator are paid at very different rates, the cheapest integer is not always the nearest one.
Classify the time-study elements. Three quantities govern a machine-coupling
problem: the servicing time during which the machine is stopped, the operator's independent work
plus walking, and the machine's own running time.
$$l = 3.1 + 0.6 = 3.7\ \text{min},\qquad
w = 2.4 + 1.9 + 0.4 = 4.7\ \text{min},\qquad
t = 41.3\ \text{min}$$
The operator therefore spends $l + w = 8.4$ min on each machine he visits, while each machine is
occupied for $l + t = 45.0$ min per piece.
Compute the theoretical number of machines. One operator can just keep $n$
machines busy when the time he spends going round them all equals the time a machine needs to
complete a piece, $n(l+w) = l+t$:
$$n' = \frac{l+t}{l+w} = \frac{3.7+41.3}{3.7+4.7} = \frac{45.0}{8.4}
= \boxed{5.36\ \text{machines per operator}}$$
Since $n$ must be an integer, the answer is 5 or 6, and the two cases behave quite differently.
Below $n'$ the system is machine-limited — every machine runs a full
45.0 min cycle and the operator has idle time. Above $n'$ it is
operator-limited — the operator works continuously and every machine waits
for him.
Write the cost per unit. Over one cycle the plant pays for one operator and
$n$ machines, and produces $n$ units:
$$\text{cost per unit}=\frac{T_c}{60}\cdot\frac{K_{\text{op}}+n\,K_{\text{mc}}}{n}
\qquad\text{where}\quad
T_c=\begin{cases} l+t & n \le n' \\[2pt] n(l+w) & n > n' \end{cases}$$
with $K_{\text{op}} = $$4.90 per hour and $K_{\text{mc}} = $$18.00 per
hour. Both terms matter: adding a machine spreads the operator's wage over more units but adds a
full machine burden, and once past $n'$ it also stretches the cycle.
Cost the machine-limited case, n = 5. Here
$5(8.4) = 42.0 \le 45.0$, so the cycle is the machine cycle of 45.0 min = 0.75 h and the operator
carries 3.0 min of idle time per cycle:
$$\text{cost per cycle}=0.75\,(4.90 + 5 \times 18.00)=0.75(94.90)=71.175$$
$$\text{cost per unit}=\frac{71.175}{5}=\boxed{14.24\ \text{per unit}}$$
in dollars, at an output of $5(60)/45.0 = 6.67$ units per hour.
Cost the operator-limited case, n = 6. Now
$6(8.4) = 50.4 > 45.0$, so the operator sets the pace and the cycle stretches to 50.4 min = 0.84 h,
leaving each machine idle for 5.4 min waiting to be served:
$$\text{cost per cycle}=0.84\,(4.90+6 \times 18.00)=0.84(112.90)=94.836
\qquad\Rightarrow\qquad
\frac{94.836}{6}=15.81\ \text{per unit}$$
in dollars — 11 per cent dearer than five machines, because the sixth machine's burden is
paid in full while the cycle it must share has grown by 12 per cent.
Check the other neighbour and conclude. Four machines give the same 45.0 min
cycle but spread the operator over fewer units,
$0.75(4.90+4 \times 18.00)/4 = 76.90(0.75)/4 = 14.42$ dollars per unit, so the cost curve has its
minimum at five:
$$\boxed{n = 5\ \text{machines per operator, at }14.24\ \text{dollars per unit}}$$
The result is the general one for machine coupling with a cheap operator and expensive machines:
round the theoretical ratio down, because idle machine time costs
$18.00 an hour and idle operator time only $4.90.
Test the alternative reading of the time study. If instead every handling
element stops the machine, $l = 3.1+0.6+2.4+1.9 = 8.0$ min and $w = 0.4$ min, so
$n' = (8.0+41.3)/8.4 = 5.87$. Costing the neighbours, $n = 5$ gives a 49.3 min cycle and
$49.3(94.90)/(60 \times 5) = 15.60$ dollars per unit while $n = 6$ gives
$50.4(112.90)/(60 \times 6) = 15.81$. The optimum is again five machines, so the
ambiguity in the time study changes the cost figure by about 9 per cent but not the answer to the
question actually asked.
Man-machine chart for the optimum assignment of five machines to one operator. The operator serves each machine for 3.7 min, then inspects, packs and walks for 4.7 min, cycling through all five in 42.0 min and idling for the remaining 3.0 min of the 45.0 min machine cycle.
Question 7 — cost per unit against the number of machines per operator
Machines per operator, n
Governing cycle, min
Limited by
Cost per cycle
Cost per unit
4
45.0
machine
$57.68
$14.42
5 — optimum
45.0
machine
$71.18
$14.24
6
50.4
operator
$94.84
$15.81
Theoretical ratio n′ = 5.36; operator idle time at n = 5 is 3.0 min per 45.0 min cycle; output 6.67 units per hour.