You are on page 1of 6

Homework 3

1. Construction
a.
Yi = 1 if we build plant,
Yi = 0 if we don’t

Objective: min 2,000,000y1 + 1,500,000y2 + 1,200,000y3 + 2,400000y4+ 3,000,000y5

Constraints
6000y1 + 8000 y2 + 4000y3 + 7000y4 + 7000y5 >= 16,000 units
X1 <= 6000
X2<= 8000
X3<= 4000
X4<= 7000
X5<= 7000
Yi is binary
Xi is positive integer

b.

c. Yi = 0 if plant not built


yi = 1 if plant is built

Objective:
Min 500x1 + 650x2 + 450x3+600x4+ 300x5+ 2,000,000y1 + 1,500,000y2 + 1,200,000y3 +
2,400000y4+ 3,000,000y5

Constraints
X1 + x2 + x3 + x4 + x5 >= 16000
Y1+y2+y3+y4+y5 >= 0
X1 <= 6000
X2 <= 8000
X3 <= 4000
X4 < = 7000
X5 < = 7000
Yi is binary
Xi is non negative and integer
d. $12,600,000

The solutions from part b and d are fairly different but they are similar in structure.
Part B
- part A just considers what plants to build so that we can meet demand.
← - by constructing plants 1,2,3 we are able to meet demand. The total cost is 4,700,000.
Part D
- considers the costs of construction and production to meet demand. The total cost is 12,600,000
- Has the same constraints as part A plus some additional ones

e.
y1 = 0 if plant not built
y1 = 1 if plant is built
Xij = # of monitors shipped from plant i where i = 1,2,3,4,5 to destination j where j=1,2,3,4

Objective:
Min 270x11 + 280x12 + 300x13 + 350x14 + 170x21 + 370x22 + 250x23 + 600x24 + 700x31 + 630x32 + 550x33 +
600x34 + 300x41 + 150x42 + 100x43+ 500x44 + 280x51 + 500x52 + 550x53 + 400x54

Constraints
X11 + x21 + x31 + x41 + x51 = 4000
X12 + x22 + x32 + x42 + x52 = 3000
X13 + x23 + x33 + x43 + x53 = 2000
X14 + x24 + x34 + x44 + x54 = 7000
X11 + x12 + x13 + x14 + x15 <= 6000
X21 + x22 + x23 + x24 + x25 < = 8000
X31 + x32 + x33 + x34 + x35 < = 4000
X41 + x43 + x43 + x44 + x45 < = 7000
X51 + x52 + x53 + x54 + x55 < = 7000
X11 + x12 + x13 + x14 + x15 ……….x54 >=16000

f.
This solution is different compared to part b

2. Customer Service
Q2

a. Probability that you will get assigned to a knowledgeable employee = 0.75


Probability that you will get assigned to a non-knowledgeable employee = 0.25
This is a Bernoulli distribution.

b.
Rate of Knowledgeable Employee: λ = 6
Mean: 1/ λ = 1/6 = 0.167
Variance: 1/ λ2 = 1/36 = 0.0278
Rate of Not Knowledgeable Employee: λ = 2
Mean: 1/ λ = 1/2 = 0.5
Variance: 1/ λ2 = 1/4 = 0.25

c. Probability that call will last more than an hour : P(X >1) = 1 – P(X<1) which is probability of call lasting at least an
hour.
P(X<1) =EXPON.DIST(1,6,FALSE) = 0.01487
So, P(X >1) = 1 - 0.01487 = 0.98512
d. Expected time on the phone: E(X) = 0.75 M(Knowledgeable) + 0.25 M(Not Knowledgeable)
= 0.75 X 0.167 + 0.25 X 0.5
= 0.12525 + 0.125
= 0.25025 hours
i.e. 15 minutes

e. Probability that you will get a knowledgeable employee 10 times out of your first 10 calls:
=BINOM.DIST(10,10,0.75, FALSE)
= 0.05631

f. Probability that you will get a knowledgeable employee exactly 8 times out of your first 10 calls:
=BINOM.DIST(8,10,0.75, FALSE)
= 0.28156

g.
Probability of getting a knowledgeable employee : P(Knowledgeable) = 0.75
For first 10 calls, N =10
Mean, E(X) = np = 10 X 0.75 = 7.5
Variance, Var(X) = np(1-p) = 10 X 0.75 X 0.25 = 1.875

h. Probability that you will get a knowledgeable employee, for the first time, the 3rd time you call: (1-p)2 x p
= 0.25 X 0.25 X 0.75
= 0.046875

i.The number of calls you need to place to get a knowledgeable employee are in geometric progression where p
= 0.75.
 Mean: 1/ p = 1/0.75 = 1.333
 Variance: 1/ p2 = 1/0.5625 = 1.778

j. To reduce the average customer waiting time :-

Option A:

All the agents are knowledgeable.


E(X) = 1 X M(Knowledgeable)
= 1 X 0.167
= 0.167

Option B:

Increase λ and µ by 20%. So the new values are,


λ = 1.2 X 6 = 7.2
Mean: 1/ λ = 1/7.2 = 0.1389

µ = 1.2 X 2 = 2.4
Mean: 1/µ= 1/ 2.4 = 0.4167
 Expected time on the phone after 20% increase:
E(X) = 0.75 M(Knowledgeable) + 0.25 M(Not Knowledgeable)
= 0.75 X 0.1389 + 0.25 X 0.4167
= 0.104175 + 0.104175
= 0.20835

Waiting Time in option A is less than option B, so choose option A.

3.

a.

b.

The model follows the general shape of the actual data but the peaks and valleys are shallow and
delayed. It isn’t very helpful. The MAPE is .0426
c.

The 3 month moving average follows much more closely and has helpful data. The MAPE is .1306

d. The differences between the two models for January 2017 differ because the first model considers
12 months of data so it smooths out fluctuations. The second model only uses 3 months and
unfortunately those 3 months were low

You might also like