IE 301 Assignment 1

Introduction to mathematical modelling LP Graphical solving

1 Problem 1

Decision variables

𝑥 = number of cakes, 𝑦 = number of bread loaves.

Resources and limits

• Flour availability: 20 cups. One cake uses 3 cups; one loaf uses 8 cups. Therefore

3𝑥 + 8𝑦 ≤ 20.

• Oven time availability: 3 hours = 180 minutes. One cake takes 45 minutes; one loaf takes 30 minutes.
Because baking cannot be done simultaneously across products, the total time is

45 30 180
45𝑥 + 30𝑦 ≤ 180 ⟺ 𝑥+ 𝑦≤ ⟺ 3𝑥 + 2𝑦 ≤ 12.
15 15 15
• Nonnegativity and integrality (for Problem 1’s “possible solutions”): 𝑥, 𝑦 ∈ ℤ≥0 .

Revenue. Each cake yields $10, each loaf yields $6; revenue is 𝑅 = 10𝑥 + 6𝑦.

Systematic enumeration of all integer-feasible (𝑥, 𝑦)

From flour: 3𝑥 + 8𝑦 ≤ 20 ⇒ 8𝑦 ≤ 20 − 3𝑥 ⇒ 𝑦 ≤ 20−3𝑥
8
.
From time: 3𝑥 + 2𝑦 ≤ 12 ⇒ 2𝑦 ≤ 12 − 3𝑥 ⇒ 𝑦 ≤ 12−3𝑥
2
.
We can bound 𝑦: from 8𝑦 ≤ 20 ⇒ 𝑦 ≤ ⌊20/8⌋ = 2. Consider 𝑦 = 0, 1, 2 and for each compute the largest integer
𝑥 that satisfies both constraints.

𝑦 From flour: From time: Feasible 𝑥

3𝑥 ≤ 20 − 8𝑦 3𝑥 ≤ 12 − 2𝑦
0 3𝑥 ≤ 20 ⇒ 𝑥 ≤ 3𝑥 ≤ 12 ⇒ 𝑥 ≤ 4 𝑥 = 0, 1, 2, 3, 4
⌊6.6⌋ = 6
1 3𝑥 ≤ 12 ⇒ 𝑥 ≤ 4 3𝑥 ≤ 10 ⇒ 𝑥 ≤ 3 𝑥 = 0, 1, 2, 3
2 3𝑥 ≤ 4 ⇒ 𝑥 ≤ 1 3𝑥 ≤ 8 ⇒ 𝑥 ≤ 2 𝑥 = 0, 1

Thus the complete list of integer-feasible pairs is:

(0, 0), (1, 0), (2, 0), (3, 0), (4, 0), (0, 1), (1, 1), (2, 1), (3, 1), (0, 2), (1, 2).

1
IE 301 Assignment 1 2

Revenue evaluation.

(0, 0) → 0, (1, 0) → 10, (2, 0) → 20, (3, 0) → 30, (4, 0) → 40,

(0, 1) → 6, (1, 1) → 16, (2, 1) → 26, (3, 1) → 36,
(0, 2) → 12, (1, 2) → 22.

The maximum revenue among integer-feasible solutions is $40 at (𝑥, 𝑦) = (4, 0).

Conclusion (Problem 1)

Bake 4 cakes and 0 loaves; revenue $40.

2 Problem 2
2.1 Linear programming model (full statement)
max 10𝑥 + 6𝑦
s.t. 3𝑥 + 8𝑦 ≤ 20 (flour),
3𝑥 + 2𝑦 ≤ 12 (oven time),
𝑥 ≥ 0, 𝑦 ≥ 0.

2.2 Graphical analysis with complete corner computations

Solve the intersection:

3𝑥 + 8𝑦 = 20 4
{ ⇒ (3𝑥+8𝑦) − (3𝑥+2𝑦) = 8 ⇒ 6𝑦 = 8 ⇒ 𝑦 = .
3𝑥 + 2𝑦 = 12 3

Substitute 𝑦 = 43 into 3𝑥 + 2𝑦 = 12:

4 8 28 28
3𝑥 + 2 ⋅ = 12 ⇒ 3𝑥 + = 12 ⇒ 3𝑥 = ⇒𝑥= .
3 3 3 9
Other extreme points: (0, 0), (4, 0), and (0, 2.5). Objective values:

(4, 0) → 40, (0, 2.5) → 15, ( 28

9 3
, 4 ) → 352
9
≈ 39.11.

So 𝑥 = 4, 𝑦 = 0 is optimal.
IE 301 Assignment 1 3

Figure 1: Problem 2: Feasible region and optimal point

3 Problem 3

Decision variables and model

𝑥 = number of chairs, 𝑦 = number of tables,

max 400𝑥 + 100𝑦
s.t. 8𝑥 + 10𝑦 ≤ 80 (labor hours),
2𝑥 + 6𝑦 ≤ 36 (wood in pounds),
𝑥 ≤ 6 (demand limit on chairs),
𝑥 ≥ 0, 𝑦 ≥ 0.

3.1 Corner points and objective values

• Labor ∩ wood:

8𝑥 + 10𝑦 = 80
{ ⇒ 10𝑥 + 30𝑦 = 180 (second × 5) ⇒ 2𝑥 + 20𝑦 = 100 ⇒ 𝑥 + 10𝑦 = 50.
2𝑥 + 6𝑦 = 36

Substitute into 2𝑥 + 6𝑦 = 36:

2(50 − 10𝑦) + 6𝑦 = 36 ⇒ 100 − 20𝑦 + 6𝑦 = 36 ⇒ −14𝑦 = −64 ⇒ 𝑦 = 32

7
, 𝑥 = 30
7
.

• Labor with 𝑥 = 6: 8(6) + 10𝑦 = 80 ⇒ 𝑦 = 3.2 ⇒ (6, 3.2).

• Wood with 𝑥 = 6: 2(6) + 6𝑦 = 36 ⇒ 𝑦 = 4 (but labor = 88 > 80 ⇒ infeasible).
• Feasible axes points include (6, 0) and (0, 6).
Objective values:

(0, 6) → 600, (6, 0) → 2400, (6, 3.2) → 2720, ( 30

7 7
, 32 ) → 15200
7
≈ 2171.43.
IE 301 Assignment 1 4

Optimal solution and slack

Optimal (𝑥, 𝑦) = (6, 3.2) with profit $2720. Labor used = 80 ⇒ labor slack = 0. Wood used = 31.2 ⇒ wood
slack = 36 − 31.2 = 4.8 pounds.

3.2 Effect of changing the table profit from $100 to $500

Let 𝑍 = 400𝑥 + 500𝑦. On 8𝑥 + 10𝑦 = 80 ⇒ 𝑦 = 8 − 0.8𝑥:

𝑍 = 400𝑥 + 500(8 − 0.8𝑥) = 400𝑥 + 4000 − 400𝑥 = 4000.

Hence every feasible point along the labor boundary is optimal (multiple optima).

Figure 2: Problem 3: Feasible region and optimal point

Data & decision variables

Let
𝑃 = number of permanent operators per day, 𝑇 = number of temporary operators per day.

4 Problem 4

1. LP formulation

min 𝑍 = 64𝑃 + 42𝑇

s.t. 16𝑃 + 12𝑇 ≥ 450 (throughput)
𝑃 + 𝑇 ≤ 40 (workstations)
0.5𝑃 + 1.4𝑇 ≤ 25 (defective-claims cap)
𝑃, 𝑇 ≥ 0 (nonnegativity).
IE 301 Assignment 1 5

2. Graphical solution (continuous LP)

Corner (extreme) points of the feasible region (all units are operators):
• Throughput & error cap: (𝑃, 𝑇 ) ≈ (20.122, 10.671)
• Throughput & axes: (28.125, 0) (the (0, 37.5) point violates the error cap)
• Seats & error cap: (34.444, 5.556)
• Seats & axis: (40, 0)
Evaluate the objective 𝑍 = 64𝑃 + 42𝑇 at the feasible corners:

(20.122, 10.671) ∶ 𝑍 ≈ 1,735.98

(28.125, 0) ∶ 𝑍 = 1,800.00
(34.444, 5.556) ∶ 𝑍 ≈ 2,437.78
(40, 0) ∶ 𝑍 = 2,560.00

Optimal continuous solution:

𝑃 ⋆ ≈ 20.12, 𝑇 ⋆ ≈ 10.67, 𝑍min ≈ $1,735.98.

(If integrality were required, one near-feasible integer choice is (𝑃, 𝑇 ) = (21, 10) with 𝑍 = $1,764.)

Figure 3: Baseline feasible region and optimal solution (𝑍 = 64𝑃 + 42𝑇 ).

3. Pay-change sensitivity
(a) Permanent pay $64 → $54; temporary stays $42. Minimize 𝑍 = 54𝑃 + 42𝑇 on the same feasible region.
The optimum moves to
(𝑃, 𝑇 ) = (28.125, 0), 𝑍min = $1,518.75.
Interpretation: all-permanent (just enough to hit 450 claims) becomes best because permanents are relatively cheaper.
IE 301 Assignment 1 6

Figure 4: Optimal solution when permanent pay is $54 (𝑍 = 54𝑃 + 42𝑇 ).

(b) Temporary pay $42 → $36; permanent stays $64. Minimize 𝑍 = 64𝑃 + 36𝑇 . The current mix remains
optimal:
(𝑃 ⋆ , 𝑇 ⋆ ) ≈ (20.12, 10.67), 𝑍min ≈ $1,671.95,
i.e., same staffing mix, lower cost.

Figure 5: Optimal solution when temporary pay is $36 (𝑍 = 64𝑃 + 36𝑇 ).

IE 301 Assignment 1 7

4. No cap on defective claims

Drop 0.5𝑃 + 1.4𝑇 ≤ 25. With only throughput and seats, cost per processed claim is

perm = 64
16
= 4, 42
temp = 12 = 3.5,

so all-temporary is cheapest. Meet 16𝑃 + 12𝑇 ≥ 450 with 𝑇 = 37.5 and 𝑃 = 0 (uses 37.5 ≤ 40 seats):

(𝑃, 𝑇 ) = (0, 37.5), 𝑍min = $1,575.

(Nearest integers: 𝑇 = 38 gives 456 claims, cost $1,596.)

Figure 6: Feasible region and optimum with no error cap (𝑍 = 64𝑃 + 42𝑇 ).

5. Raise minimum throughput to at least 650

Keep seats ≤ 40 and errors ≤ 25, but require 16𝑃 + 12𝑇 ≥ 650. Even at the intersection of the seat and error
boundaries (𝑃, 𝑇 ) = (34.444, 5.556) we get

16𝑃 + 12𝑇 ≈ 617.78 < 650,

so the system cannot reach 650 claims/day under those two limits. Hence:

Infeasible with seats ≤ 40 and errors ≤ 25.

IE 301 Assignment 1 8

Figure 7: Throughput ≥ 650 with seats ≤ 40 and errors ≤ 25: infeasible.

5 Problem 5

Data

Each pound of coffee produces 16 cups. There are 10 pounds each of Colombian, Kenyan, and Indonesian
coffees. Brewing capacity is 30 gallons per day = 30 × 8 = 240 cups. Sales ratio: 𝑃 = 1.5 𝐶.

Pomona: (0.20, 0.35, 0.45), Coastal: (0.60, 0.10, 0.30)

0.20𝑃 + 0.60𝐶
Colombian: ≤ 10 ⟺ 𝑃 + 3𝐶 ≤ 800,
16
0.35𝑃 + 0.10𝐶
Kenyan: ≤ 10 ⟺ 7𝑃 + 2𝐶 ≤ 3200,
16
0.45𝑃 + 0.30𝐶
Indonesian: ≤ 10 ⟺ 9𝑃 + 6𝐶 ≤ 3200,
16
Brew: 𝑃 + 𝐶 ≤ 240, 𝑃 = 1.5𝐶, 𝑃, 𝐶 ≥ 0,
Objective: max 𝑃 + 𝐶.

5.1 Reduction with the ratio 𝑃 = 1.5𝐶 and solution

Col: 1.5𝐶 + 3𝐶 ≤ 800 ⇒ 𝐶 ≤ 1600/9 ≈ 177.78,
Ken: 10.5𝐶 + 2𝐶 ≤ 3200 ⇒ 𝐶 ≤ 256,
Ind: 13.5𝐶 + 6𝐶 ≤ 3200 ⇒ 𝐶 ≤ 3200/19.5 ≈ 164.10,
Brew: 2.5𝐶 ≤ 240 ⇒ 𝐶 ⋆ = 96, 𝑃 ⋆ = 144, 𝑃 ⋆ + 𝐶 ⋆ = 240 .
IE 301 Assignment 1 9

5.2 Resource usage at the optimum

0.20(144) + 0.60(96) 86.4
Col: = = 5.4 < 10,
16 16
0.35(144) + 0.10(96) 60
Ken: = = 3.75 < 10,
16 16
0.45(144) + 0.30(96) 93.6
Ind: = = 5.85 < 10.
16 16
Beans are slack; brew capacity binds.

5.3 Managerial questions

One more pound of any bean does not increase sales (bean constraints are nonbinding). If brewing increases to 40
gallons = 320 cups:
2.5𝐶 ≤ 320 ⇒ 𝐶 = 128, 𝑃 = 192, total = 320,
and beans remain slack.

Figure 8: Problem 5: Feasible region, ratio line, and optimum

6 Question 6

Definitions and master formula

Fixed monthly cost 𝐹 = $25,000, variable cost 𝑣 per pound, price 𝑝 per pound.
Total revenue: 𝑇 𝑅(𝑄) = 𝑝𝑄, total cost: 𝑇 𝐶(𝑄) = 𝐹 + 𝑣𝑄.
Break-even: 𝑇 𝑅(𝑄) = 𝑇 𝐶(𝑄) ⇒ (𝑝 − 𝑣)𝑄 = 𝐹 ⇒ 𝑄BE = 𝐹 /(𝑝 − 𝑣) .

6.1 Parts (a)–(e)

(a) With 𝑝 = 0.40, 𝑣 = 0.15 ⇒ 𝑐 = 0.25: 𝑄BE = 25,000/0.25 = 100,000 pounds.
(b) The lines 𝑇 𝑅 and 𝑇 𝐶 intersect at 𝑄 = 100,000 and $40,000.
IE 301 Assignment 1 10

Figure 9: Question 6b: Break-even plot at 𝑝 = $0.40, 𝑣 = $0.15

(c) If 𝑝 = 0.60 and 𝑣 = 0.15 ⇒ 𝑐 = 0.45: 𝑄BE = 25,000/0.45 ≈ 55,555.56.

(d) If 𝑣 = 0.22 (and 𝑝 = 0.60 ⇒ 𝑐 = 0.38): 𝑄BE = 25,000/0.38 ≈ 65,789.47.
(e) Add $14,000 per year to 𝐹 ⇒ $1,166.67 per month, so 𝐹new = 26,166.67. With 𝑐 = 0.38: 𝑄BE = 26,166.67/0.38 ≈
68,860.71.
IE 301 Assignment 1 11

(a) Part (c): Price increase to $0.60.

(b) Part (d): Variable cost increase to $0.22.

(c) Part (e): Added advertising (higher fixed cost).

Figure 10: Question 6c–e: Comparative break-even plots