Case 1: Toy Manufacturing Company

a. Linear Mathematical Model:

Decision Variables:
SC: Number of units of small car to be produced and sold
MC: Number of units of midsize car to be produced and sold
FC: Number of units of fullsize car to be produced and sold
MM: Number of units of midsize minivan to be produced and sold
LM: Number of units of large minivan to be produced and sold

Objective function:
Maximize Total Profit (in Rs.): 18SC + 20MC + 30FC + 50MM + 55LM

The objective here is to maximize profit. For small cars, since number of units to be produced
and sold is SC and profit per unit is 18, 18SC represents profit from sale of all units of small
cars produced and sold. Similarly, 20MC and 30FC represent profit from sale of all units of
midsize and fullsize cars, respectively, and 50MM and 55LM represent profit from sale of all
units of midsize minivan and large minivan, respectively. Total profit is summation of profit
from all vehicle types.

Constraints:

(i) 0.20SC + 0.25MC + 0.40FC + 0.60MM + 0.70LM <= 6,000

Plastic consumption per unit of small car produced is 0.20 kg; hence, plastic consumption in
production of all units of small car is 0.20SC kgs. Similarly, plastic consumption in production
of midsize car and fullsize car is 0.25MC kgs and 0.40FC kgs, respectively, and that in
production of midsize minivan is 0.60MM kgs and large minivan is 0.70LM kgs. Since a total of
6,000 kgs of plastic is available in the coming month, the total plastic consumption needs to
be less than or equal to 6,000 kgs.

(ii) 18SC + 15MC + 24FC + 30MM + 40LM <= 36,000

Labour consumption per unit of small car produced is 18 minutes; hence, labour consumption
in production of all units of small car is 18SC minutes. Similarly, labour consumption in
production of midsize car and fullsize car is 15MC minutes and 24FC minutes, respectively,
and that in production of midsize minivan is 30MM minutes and large minivan is 40LM
minutes. Since a total of 600 hours (36,000 minutes) of labour is available in the coming
month, the total labour consumption needs to be less than or equal to 36,000 minutes.

(iii) MM <= 1,000

Sales of midsize minivan cannot exceed 1,000 units

(iv) LM <= 1,000

Sales of large minivan cannot exceed 1,000 units

(v) SC >= 0, MC >= 0, FC >=0, MM >=0, LM >=0

Number of cars produced and sold cannot be negative

b. Optimal product mix (please refer to the attached spreadsheet for calculations)

Produce Units
Midsize Minivan 1,000
Large Minivan 150

Maximized profit: Rs. 58,250

c. Change in optimal product mix based on change in profit for fullsize car

To assess sensitivity of the optimal product mix to profit per unit of fullsize car, one would
need to understand the certain key definitions.

Allowable increase/decrease in coefficient of a decision variable: In the sensitivity table, with

respect to coefficient of a decision variable, allowable increase represents the value by which
the objective function coefficient can be increased without change in the optimal solution.
Similarly, allowable decrease represents the value by which the objective function coefficient
can be decreased without change in the optimal solution.

In this example, the objective function coefficient (profit in Rs.) for fullsize car is 30. Per the
sensitivity report, the allowable increase in this coefficient is 3. Thus, for the optimal product
mix to change, the profit from fullsize car would need to increase from Rs. 30 per unit to
beyond Rs. 33 per unit. This implies that, if the profit from fullsize car increases from Rs. 30 to
Rs. 33, the optimal product mix would remain unchanged; however, if it increases (beyond Rs.
33) to Rs. 34 or Rs. 35, the optimal product mix would change. This can be confirmed by
plugging these values in the mathematical model. The results of this exercise are captured in
the table below:
Optimal Midsize Fullsize Midsize Large
Profit per unit Small car
profit car car Minivan Minivan
Fullsize car (Rs.) (units)
(Rs.) (units) (units) (units) (units)
30 58,250 0 0 0 1,000 150
31 58,250 0 0 0 1,000 150
32 58,250 0 0 0 1,000 150
33 58,250 0 0 0 1,000 150
34 58,500 0 0 250 1,000 0
35 58,750 0 0 250 1,000 0
On the contrary, the allowable decrease is 10^30. Thus, the optimal product mix would not
change even if the coefficient (profit) declines to a very large extent. This can be confirmed by
plugging these values in the mathematical model. The results of this exercise are captured in
the table below:
Optimal Midsize Fullsize Midsize Large
Profit per unit Small car
profit car car Minivan Minivan
Fullsize car (Rs.) (units)
(Rs.) (units) (units) (units) (units)
30 58,250 0 0 0 1,000 150
29 58,250 0 0 0 1,000 150
28 58,250 0 0 0 1,000 150
20 58,250 0 0 0 1,000 150
10 58,250 0 0 0 1,000 150
0 58,250 0 0 0 1,000 150
d. Purchase of additional amount of plastic

To assess the attractiveness of purchasing additional amount of plastic, it is critical to

understand the concept of shadow price.

Shadow price of a resource: In the sensitivity table, shadow price represents the expected
increase in profit from a unit increase in availability of a resource. Only constrained resources
can have a shadow price. If a resource is not fully exhausted, its shadow price is zero and thus
it is not attractive to purchase additional units of this resource.

In this example, after arriving at the optimal solution, 705 kgs of plastic is consumed and 5,295
kgs is left unutilized. Considering the availability of surplus, the shadow price of plastic is zero.
Purchasing additional amount of plastic would not contribute to incremental profit but would
cost Rs. 5 per kg or Rs. 10 per kg, leading to suboptimal solution. Hence, additional units should
not be purchased.

e. Purchase of additional amount of labour

All the 600 hours (36,000 minutes) of labour were fully utilized to arrive at the optimal
solution. As a result, labour commands a shadow price. Per the sensitivity table, shadow price
of labour stands at Rs. 1.375 per minute (or Rs. 82.5 per hour). This indicates that every
additional hour of labour would contribute Rs. 82.5 to profitability.

Since purchasing labour at Rs. 75 and Rs. 80 would lead to net gain of Rs. 7.5 and Rs. 2.5,
respectively, additional units of labour should be purchased at these prices. On the contrary,
purchasing labour at Rs. 90 and Rs. 100 would lead to net loss of Rs. 7.5 and Rs. 17.5; hence,
purchase of additional units at these prices should be avoided.

Allowable increase in a resource: In the sensitivity table, with respect to a resource, allowable
increase represents the extent to which increase in availability of a constrained resource
would contribute to a constant increase in profit. Beyond this threshold, the shadow price of
the constrained resourced will change. Since, per the sensitivity table, the allowable increase
for labour is 34,000 (minutes), it would be beneficial to purchase up to 34,000 minutes (566.7
hours) of labour.
Case 2: Portfolio Manager

a. Linear Mathematical Model:

Decision Variables:
A: Amount to be invested in bond A ($)
B: Amount to be invested in bond B ($)
C: Amount to be invested in bond C ($)
D: Amount to be invested in bond D ($)
E: Amount to be invested in bond E ($)

Objective function:
Maximize after-tax earnings (in $): 4.3%A + 2.7%B + 2.5%C + 2.2%D + 4.5%E

The objective here is to maximize after-tax earnings for the next year. For bond A, since
amount to be invested is A and after-tax earnings are 4.3%, 4.3%A represents after-tax
earnings from investment in bond A. Similarly, 2.7%B, 2.5%C, 2.2%D, and 4.5%E represent
after-tax earnings from investment in bonds B, C, D, and E respectively. Total after-tax
earnings are summation of after-tax earnings from all bond types.

Constraints:

(i) B + C + D >= 4,000,000

Investment in Government and Agency bonds is required to be $4 million at minimum

(ii) A + B + C + D + E <= 10,000,000

Total investment all bonds cannot exceed $10 million

(iii) (2A + 2B + C + D + 5E) / (A + B + C + D + E) <= 1.4

Weighted-average bond quality cannot exceed 1.4

The above inequality can be simplified as follows (maintaining linearity is required by

the solver to function):
(2A – 1.4A) + (2B – 1.4B) + (C – 1.4C) + (D – 1.4D) + (5E – 1.4E) <= 0
0.6A + 0.6B – 0.4C – 0.4D + 3.6E <= 0

(iv) (9A + 15B + 4C + 3D + 2E) / (A + B + C + D + E) <= 5

Weighted-average years to maturity cannot exceed 5

The above inequality can be simplified as follows:

(9A – 5A) + (15B – 5B) + (4C – 5C) + (3D – 5D) + (2E – 5E) <= 0
4A + 10B – C – 2D – 3E <= 0

(v) SC >= 0, MC >= 0, FC >=0, MM >=0, LM >=0

Amounts invested cannot be negative

b. Optimal investment mix (please refer to the attached spreadsheet for calculations)

Invest in Amount ($)

Bond A 2,181,818
Bond C 7,363,636
Bond E 454,545
Total 10,000,000

Maximized after-tax earnings: $298,364

Case 3: Distribution problem

a. Linear Mathematical Model:

Decision Variables:
The number of tons to be distributed from factory to depot/customer and from depot to
customer are identified as follows:

Supplied to Liverpool Brighton Newcastle Birmingham London Exeter Total

Factory Factory Depot (N) Depot (B) Depot (L) Depot (E)
(LI) (BR)
Newcastle LIN -
Birmingham LIB BRB
London LIL BRL
Exeter LIE BRE
C1 LI1 BR1 - B1 - - 50,000
C2 - - N2 B2 L2 - 10,000
C3 LI3 - N3 B3 L3 E3 40,000
C4 LI4 - N4 B4 - E4 35,000
C5 - - - B5 L5 E5 60,000
C6 LI6 - N6 - L6 E6 20,000
Total 150,000 200,000 70,000 50,000 100,000 40,000

In the above table, LIN and LI1 represent number of tons to be distributed from the Liverpool
Factory to the Newcastle Depot and to customer C1, respectively. Similarly, BRB and BR1
represent number of tons to be distributed from Brighton Factory to Birmingham Depot and
to customer C1, respectively. Lastly, N3, B3, L3, and E3 represent number of tons distributed
from the 4 depots to customer 3.

Objective function:
Minimize distribution cost:
0.5LIN + 0.5LIB + 1.0LIL + 0.2LIE + 1.0LI1 + 1.5LI3 + 2.0LI4 + 1.0LI6 + 0.3BRB + 0.5BRL + 0.2BRE
+ 2.0BR1 + 1.5N2 + 0.5N3 + 1.5N4 + 1.0N6 + 1.0B1 + 0.5B2 + 0.5B3 + 1.0B4 + 0.5B5 + 1.5L2 +
2.0L3 + 0.5L5 + 1.5L6 + 0.2E3 + 1.5E4 + 0.5E5 + 1.5E6

The objective here is to minimise the overall distribution cost. 0.5LIN represents cost of
distributing 1 ton from Liverpool Factory to Newcastle Depot. Similarly, 1.0B1 represents cost
of distributing 1 ton from Birmingham Depot to customer C1. All terms in the objective
function together represent the total cost of distributing the products from factory to
depot/customers and from depot to customers.

Constraints:

(i) LIN + LIB + LIL + LIE + LI1 + LI3 + LI4 + LI6 <= 150,000

The monthly production capacity of the Liverpool Factory is 150,000 tons

(ii) BRB + BRL + BRE + BR1 <= 200,000

The monthly production capacity of the Brighton Factory is 200,000 tons

(iii) N2 + N3 + N4 + N6 <= 70,000

The monthly throughput capacity of the Newcastle Depot is 70,000 tons

(iv) B1 + B2 + B3 + B4 + B5 <= 50,000

The monthly throughput capacity of the Birmingham Depot is 50,000 tons

(v) L2 + L3 + L5 + L6 <= 100,000

The monthly throughput capacity of the London Depot is 100,000 tons

(vi) E3 + E4 + E5 + E6 <= 40,000

The monthly throughput capacity of the Exeter Depot is 40,000 tons

(vii) LIN – N2 – N3 – N4 – N6 = 0

At Newcastle Depot, demand must equal supply

(viii) LIB + BRB – B1 – B2 – B3 – B4 – B5 = 0

At Birmingham Depot, demand must equal supply

(ix) LIL + BRL – L2 – L3 – L5 – B6 = 0

At London Depot, demand must equal supply

(x) LIE + BRE – E3 – E4 – E5 – E6 = 0

At Exeter Depot, demand must equal supply

(xi) LI1 + BR1 + B1 = 50,000

The monthly requirement of C1 is 50,000 tons

(xii) N2 + B2 + L2 = 10,000

The monthly requirement of C2 is 10,000 tons

(xiii) LI3 + N3 + B3 + L3 + E3 = 40,000

The monthly requirement of C3 is 40,000 tons

(xiv) LI4 + N4 + B4 + E4 = 35,000

The monthly requirement of C4 is 35,000 tons

(xv) B5 + L5 + E5 = 60,000

The monthly requirement of C5 is 60,000 tons

(xvi) LI6 + N6 + L6 + E6 = 20,000

The monthly requirement of C6 is 20,000 tons

(xvii) All decision variables >= 0

Number of tons produced cannot be negative

b. Optimal distribution mix (please refer to the attached spreadsheet for calculations)

Distribute from Distribute to Tons

Liverpool factory Exeter depot 40,000
Liverpool factory Customer 1 50,000
Liverpool factory Customer 6 20,000
Liverpool factory total 110,000
Brighton factory Birmingham depot 50,000
Brighton factory London depot 55,000
Brighton factory total 105,000
Birmingham depot Customer 2 10,000
Birmingham depot Customer 4 35,000
Birmingham depot Customer 5 5,000
Birmingham depot total 50,000
London depot Customer 5 55,000
Exeter depot Customer 3 40,000

Minimized distribution cost: GBP198,500

Case 4: Paper recycling plant

a. Linear Mathematical Model:

Acronyms used:
To economize usage of words, we have used certain acronyms in this write up. They are as
stated below:
 G1, G2, G3 – Grade 1, Grade 2, Grade 3 paper
 BB, TP, NP, BP – Box Board, Tissue Paper, Newsprint, and Book Paper
 DI, AD – De-inking, Asphalt Dispersion processing methods

Decision Variables:
The number of tons to be produced are identified as follows:
Process De-inking Asphalt Dispersion
Input BB TP NP BP BB TP NP BP
Grade 1 - - G1DINP G1DIBP - - G1ADNP G1ADBP
Grade 2 G2DIBB G2DITP - G2DIBP G2ADBB G2ADTP - G2ADBP
Grade 3 G3DIBB G3DITP G3DINP - G3ADBB G3ADTP G3ADNP -
In the above table, G1DINP represent number of tons of G1 paper to be produced using the
DI process and NP, while G1ADBP represents number of tons of G1 paper to be produced
using AD process and BP. Similarly, G2DITP represents number of tons of G2 paper to be
produced using DI process and TP and G2ADBB represents number of tons of G2 paper to be
produced using AD process and BB. Lastly, G3DINP represent number of tons of G3 paper to
be produced using DI process and NP and G3ADTP represent number of tons of G3 paper to
be produced using AD process and TP.

Objective function:
To state the objective function (minimization of total cost of meeting demand), it is imperative
to understand the total cost per ton of output. The table below explains this:

Total Cost
De-inking Asphalt dispersion
Particulars BB TP NP BP BB TP NP BP
Cost per ton ($) A 5 6 8 10 5 6 8 10
Pulp content B 15% 20% 30% 40% 15% 20% 30% 40%
Process efficiency factor C 90% 90% 90% 90% 80% 80% 80% 80%
Input required to produce 1 ton of output D=(1/B)/A 7.41 5.56 3.70 2.78 8.33 6.25 4.17 3.13
Cost of inputs per ton of output ($) E=A*D 37.04 33.33 29.63 27.78 41.67 37.50 33.33 31.25
Processing cost per ton of input ($) F 20.00 20.00 20.00 20.00 15.00 15.00 15.00 15.00
Processing cost per ton of output ($) G=D*F 148.15 111.11 74.07 55.56 125.00 93.75 62.50 46.88
Total cost per ton of output ($) H=E+G 185.19 144.44 103.70 83.33 166.67 131.25 95.83 78.13

Minimize total cost:

185.19G1DINP + 83.33G1DIBP + 95.83G1ADNP + 78.13G1ADBP + 185.19G2DIBB +
144.44G2DITP + 83.33G2DIBP + 166.67G2ADBB + 131.25G2ADTP + 78.13G2ADBP +
185.19G3DIBB + 144.44G3DITP + 103.70G3DINP + 166.67G3ADBB + 131.25G3ADTP +
95.83G3ADNP

The objective here is to minimise the total cost. 185.19G1DINP represents total cost of
producing 1 ton of G1 using DI process and NP. Similarly, 95.83G3ADNP represents total cost
of producing 1 ton of G3 using AD process and NP. All terms in the objective function together
represent the total cost of meeting the demand.
Constraints:

(i) 3.70G1DINP + 2.78G1DIBP + 7.41G2DIBB + 5.56G2DITP + 2.78G2DIBP + 7.41G3DIBB +

5.56G3DITP + 3.70G3DINP <= 3,000

Since input for the DI process using NP is 3.70 times output, 3.70G1DINP represents input used
in production of G1 paper using DI process and NP. The summation of all terms represents total
input to the DI process from all sources (BB, TP, NP, BP) and for production of all grades (G1,
G2, G3). The DI process can at most process 3,000 tons of input.

(ii) 4.17G1ADNP + 3.13G1ADBP + 8.33G2ADBB + 6.25G2ADTP + 3.13G2ADBP +

8.33G3ADBB + 6.25G3ADTP + 4.17G3ADNP <= 3,000

Since input for the AD process using NP is 4.17 times output, 4.17G1DINP represents input used
in production of G1 paper using AD process and NP. The summation of all terms represents
total input to the AD process from all sources (BB, TP, NP, BP) and for production of all grades
(G1, G2, G3). The AD process can at most process 3,000 tons of input.

(iii) G1DINP + G1DIBP + G1ADNP + G1DIBP = 500

The terms on the LHS represent total production of G1 paper. Total production must equal
demand (500 tons).

(iv) G2DIBB + G2DITP + G2DIBP + G2ADBB + G2ADTP + G2ADBP = 500

The terms on the LHS represent total production of G2 paper. Total production must equal
demand (500 tons).

(v) G3DIBB + G3DITP + G3DINP + G3ADBB + G3ADTP + G3ADNP = 600

The terms on the LHS represent total production of G3 paper. Total production must equal
demand (600 tons).

(xviii) All decision variables >= 0

Number of tons produced cannot be negative

b. Optimal production mix (please refer to the attached spreadsheet for calculations)

Produce Tons
G1 paper using DI process and BP 340
G1 paper using AD process and BP 160
G2 paper using DI process and BP 500
G3 paper using AD process and NP 600

Minimized total cost: $140,000