Group Assignment 1

Submitted by:

Group 1
Ashishkumar Prajapati - 101502564
Magdiel Rebelo - 101533146
Needesh Selvakumar - 101494357
Pitiyavarun Sivanesasuntharam - 101540762
Avinash Suryawanshi - 101436278

Business Decision making through Advance Analytics

Submitted to:
Prof. Ryan Brownlee

June 19th, 2024

George Brown College

1
Table of Contents
Case 1: Enactus...................................................................................2
Constraints.....................................................................................3
CASE 2: TESLA.....................................................................................4
Decision Variables:......................................................................................................... 4
Standard LP Formulation:............................................................................................... 4
Decision Variables:......................................................................................................... 6
Standard LP Formulation:............................................................................................... 6
Objective Function:..................................................................................................... 6

CASE: 3.............................................................................................9
2. Executable Solution Implementation...........................................10
3.Sensitivity Analysis and Scenario Planning...................................13

Case study

2
 Case 1: Enactus
Enactus is a GBC student run business that has staffing requirements for different
numbers of students on different days of the week.

Day of the Week Minimum Number of Students Required

Monday 17
Tuesday 13
Wednesday 15
Thursday 19
Friday 14
Saturday 16
Sunday 11

Develop an LP model that relates five-day shift schedules to daily numbers of students
available and implement a solution to find a schedule that uses the fewest number of
students and meets all daily workforce requirements.

Answer
We need to ensure that the total number of students available each day meets the
minimum requirements. Each student works a five-day shift, so the constraints for
each day of the week will be formulated based on the overlap of these shifts.

3
Constraints
Equations derived for all the seven days.

Excel solver used to find out the fewest number of students and meets all
daily workforce requirements (Find the answers below).

4
 CASE 2: TESLA
Part a) Cost Minimization Model
Tesla needs to determine the most cost-effective way to ship vehicles from three
manufacturing plants to four regions while meeting regional demands and staying within
plant capacities. The goal is to minimize the total shipping cost.
We used a linear programming (LP) model to find the optimal shipping plan that minimizes
costs. The model includes decision variables for the number of vehicles shipped from each
plant to each region, an objective function to minimize total shipping costs, and constraints
for plant capacities and regional demands. Assumptions include fixed shipping costs and
deterministic demand and capacity.

Decision Variables:

Let xi,j be the number of vehicles shipped from plant i to region j, where:

 i is the plant index (1, 2, 3)

 j is the region index (1, 2, 3, 4)

Standard LP Formulation:

Objective function:
Minimize Z=131x1,1+218x1,2+266x1,3+120x1,4+250x2,1+116x2,2+263x2,3+278x2,4+178x3,1+132x3,
2+122x3,3+180x3,4

Subject to:
Plant capacity constraints:
x1,1+x1,2+x1,3+x1,4≤450
x2,1+x2,2+x2,3+x2,4≤600
x3,1+x3,2+x3,3+x3,4≤500

Regional demand constraints:

x1,1+x2,1+x3,1≥450
x1,2+x2,2+x3,2≥200
x1,3+x2,3+x3,3≥300
x1,4+x2,4+x3,4≥300

Non-negativity constraints:
xi,j≥0 for all i and j

5
The results of the linear programming model indicate the optimal number of vehicles that
should be shipped from each plant to each region to minimize the total shipping cost while
meeting the regional demands and adhering to the plant capacities.
The solution minimizes the total shipping cost to $176,050, which is achieved by optimally
distributing the shipments from the three plants to the four regions. Each shipment route is
chosen to ensure that the cost is kept as low as possible.
 Plant 1 ships its maximum capacity of 450 vehicles;
 Plant 2 ships 300 vehicles, which is under its capacity of 600 vehicles, indicating
some underutilization;
 Plant 3 ships its maximum capacity of 500 vehicles;
All regional demands are fully met.
The model suggests shipping 150 vehicles from Plant 1 to Region 1 and the remaining 300
vehicles to Region 4. This indicates that shipping from Plant 1 to Region 4 is relatively cost-
effective despite the long distance.
Plant 2 ships 100 vehicles to Region 1 and 200 vehicles to Region 2, effectively balancing the
cost with its shipping capacity.
Plant 3 focuses on supplying Region 1 with 200 vehicles and Region 3 with 300 vehicles,
fully utilizing its capacity.
Recommendations:
The optimal solution suggests specific quantities to be shipped from each plant to each
region, minimizing total shipping costs to $176,050. Plant 1 is fully utilized, Plant 2 is
6
partially utilized, and Plant 3 is fully utilized. The solution ensures that all regional demands
are met efficiently.
Sensitivity and Constraint Impact Summary:
The sensitivity report shows that Plant 1's capacity is a binding constraint, meaning its full
utilization is critical for the cost minimization solution. The shadow prices indicate the cost
impact of increasing plant capacities: increasing Plant 1's capacity would further reduce costs.
However, increasing capacity for Plant 2 or Plant 3 would have minimal impact as they are
not fully utilized. The marginal values highlight the cost savings or additional costs
associated with each shipping route.

Part b) Profit Maximization Model

Tesla aims to maximize after-tax profits from shipping vehicles, considering varying
production costs, tax rates, and selling prices across different regions. The goal is to optimize
the production and shipping strategy to maximize profitability.
We enhanced the initial model by incorporating production costs, tax rates, and selling prices.
The new objective function maximizes after-tax profits instead of minimizing shipping costs.
The model still includes the same constraints on plant capacities and regional demands.
Assumptions include fixed production costs, tax rates, and selling prices.

Decision Variables:

Let xi, be the number of vehicles shipped from plant i to region j, where:

 i is the plant index (1, 2, 3)

 j is the region index (1, 2, 3, 4)

Standard LP Formulation:

Objective Function:

The profit before tax for shipping from plant i to region j is:
Profitbefore tax,i,j=Selling Pricej−Production Costi−Shipping Costi,j

The after-tax profit for shipping from plant i to region j is:

Profitafter tax,i,j=Profitbefore tax,i,j×(1−Tax Ratei)

Thus, the objective is to maximize: Maximize Z=∑3i =1 ∑4j=1 xi,j ×

((Selling Pricej−Production Costi−Shipping Costi,j) × (1−Tax Ratei))

7
The optimal solution provided by the enhanced model shows how to maximize after-tax
profits given the production costs, selling prices, and tax rates.
The optimal solution maximizes after-tax profits to approximately $3,407,310. Plant 1 ships
all its capacity to Region 1, Plant 2 ships to Regions 3 and 4, and Plant 3 ships to Region 2.
The model suggests prioritizing routes that yield higher after-tax profits, even if they involve
higher shipping costs.
 Plant 1: Fully utilized by shipping 450 vehicles to Region 1, taking advantage of
lower production costs and favorable tax rates;
 Plant 2: Utilized by shipping 300 vehicles each to Region 3 and Region 4, ensuring
high-profit routes are prioritized;
 Plant 3: Fully utilized by shipping 500 vehicles to Region 2, optimizing after-tax
profit due to lower tax rates;
In this enhanced model, the capacity of Plant 1 remains a binding constraint, highlighting its
critical role in profit maximization. The shadow prices reveal that increasing Plant 1's
capacity would significantly enhance profits, while changes in Plant 2 or Plant 3's capacities
have lesser effects. Sensitivity ranges show the stability of the solution within specific cost
and demand variations, indicating robustness against moderate changes in input parameters.

What If:
Selling prices in all regions increase by 5% due to market demand.
New Selling Prices:
 Region 1: $20,254.5
 Region 2: $21,546
 Region 3: $18,448.5
 Region 4: $19,236

8
 With the new selling prices, Tesla can achieve a total after-tax profit of $4,461,195. This
is an increase from the baseline scenario, demonstrating the impact of higher selling
prices on profitability.
Regional Pricing Impact:
 Region 1: The highest profit after tax for Plant 1 is in Region 1 ($4,041), making
it an attractive destination for shipments from Plant 1.
 Region 2: Plant 1 also sees significant profits from Region 2 ($4,885), indicating a
strong incentive to prioritize shipments to this region.
 Region 3 and 4: The profits from Plant 1 to Region 3 and Region 4 are lower but
still substantial ($2,683 and $3,336, respectively). Plant 2 and Plant 3 have lower
profits across all regions compared to Plant 1, suggesting a strategic advantage in
utilizing Plant 1 for these high-profit regions.
By analyzing these updated profit values and incorporating them into the model, Tesla can
strategically maximize after-tax profits in response to increased selling prices. The model
provides valuable insights into optimal shipping strategies and highlights the importance of
balancing plant capacities and regional demands to achieve the highest profitability.

CASE: 3
9
1.Executive
Summary of Solution Approach and
Recommendations
Presenting Problem:

Krishna Copiers needs to determine the optimal locations for three service
centers among 11 cities to minimize the total annual distance traveled by
service representatives. Each customer city must be assigned to one of
the three service centers.

Solution Approach:

Model Formulation: A mixed-integer linear programming (MILP) model was

formulated where binary variables represent the decision of whether to
place a service center in a city and the assignment of each customer city
to a service center.

Objective: Minimize the total annual distance traveled.

Constraints:

Each city must be assigned to exactly one service center.

Exactly three service centers must be established.

Cities can only be assigned to cities with service centers.

Assumptions:

Distances and trips are accurately represented.

Costs are linear with distance and number of trips.

Solution Tool: The PuLP library in Python was used to solve the MILP.

Recommendations:

Selected Service Centers: The optimal locations for the service centers are
New York, Chicago, and LA.

Customer Assignments: Each customer city is assigned to the nearest

service center based on the optimized solution.

Implementation: The model should be regularly reviewed and adjusted

based on changing travel patterns or new data.

10
2. Executable Solution Implementation
Below is the Python code to solve the problem using PuLP.

python

import pulp

# Define the problem

problem = pulp.LpProblem("Krishna_Copiers", pulp.LpMinimize)

# List of cities

cities = ["Boston", "Chicago", "Dallas", "Denver", "LA", "Miami", "New

York", "Phoenix", "Pittsburgh", "San Francisco", "Seattle"]

# Distance matrix (in miles)

distances = [

[0, 983, 1815, 1991, 3036, 1593, 213, 2664, 792, 2385, 2612],

[983, 0, 1205, 1050, 2112, 1390, 840, 1729, 457, 2212, 2052],

[1815, 1205, 0, 801, 1425, 1332, 1604, 1027, 1237, 1765, 2404],

[1991, 1050, 801, 0, 1174, 2041, 1780, 836, 1411, 1765, 1373],

[3036, 2112, 1425, 1174, 0, 2757, 2825, 398, 2456, 403, 1909],

[1593, 1390, 1332, 2041, 2757, 0, 1258, 2359, 1250, 3097, 3389],

[213, 840, 1604, 1780, 2825, 1258, 0, 2442, 386, 3036, 2900],

[2664, 1729, 1027, 836, 398, 2359, 2442, 0, 2073, 800, 1482],

[792, 457, 1237, 1411, 2456, 1250, 386, 2073, 0, 2653, 2517],

[2385, 2212, 1765, 1765, 403, 3097, 3036, 800, 2653, 0, 817],

[2612, 2052, 2404, 1373, 1909, 3389, 2900, 1482, 2517, 817, 0]

11
# Annual trips to each city

trips = [885, 760, 1124, 708, 1224, 1152, 1560, 1222, 856, 1443, 612]

# Decision variables

x = pulp.LpVariable.dicts("service_center", cities, cat='Binary')

y = pulp.LpVariable.dicts("assignment", [(i, j) for i in cities for j in cities],

cat='Binary')

# Objective function

problem += pulp.lpSum([distances[cities.index(i)][cities.index(j)] *
trips[cities.index(i)] * y[(i, j)] for i in cities for j in cities])

# Constraints

# Each city should have exactly one service center assigned to it

for i in cities:

problem += pulp.lpSum([y[(i, j)] for j in cities]) == 1

# Limit the number of service centers to 3

problem += pulp.lpSum([x[j] for j in cities]) == 3

# A city can only be assigned to a city that has a service center

for i in cities:

for j in cities:

problem += y[(i, j)] <= x[j]

# Solve the problem

problem.solve()

# Display the results

print("Status:", pulp.LpStatus[problem.status])
12
print("\nService centers will be located in:")

for city in cities:

if pulp.value(x[city]) == 1:

print(f" - {city}")

print("\nCity assignments to service centers:")

for i in cities:

for j in cities:

if pulp.value(y[(i, j)]) == 1:

print(f" - {i} is assigned to {j}")

print("\nTotal annual distance travelled:", pulp.value(problem.objective))

You can implement these scenarios by modifying the trips and

distances arrays in the code and rerunning the optimization. Here
is an example of how to implement Scenario 1:

python

Copy code

# Scenario 1: Increase demand in New York

trips_scenario_1 = trips.copy()

trips_scenario_1[cities.index("New York")] *= 1.5

# Re-define the problem with updated trips

problem_scenario_1 = pulp.LpProblem("Krishna_Copiers_Scenario_1",
pulp.LpMinimize)

# ... (rest of the code remains the same, but use trips_scenario_1 instead
of trips)

13
This structure ensures you cover all the requested elements in your
assignment, providing a comprehensive analysis and solution approach.

3.Sensitivity Analysis and Scenario Planning

Given that PuLP does not directly provide sensitivity analysis for integer
programming problems, you can perform scenario analysis by manually
adjusting parameters. Here are a few scenarios you can analyze:

Scenario 1: Increased Demand in New York

Increase the annual trips to New York by 50% and observe the impact on
service center locations and total distance traveled.

Scenario 2: Additional Service Center Constraint

Add a fourth service center and determine how this impacts the total
distance traveled and the location of service centers.

Scenario 3: Distance Reduction Between Key Cities

Reduce the distance between Chicago and LA by 20% to simulate

improved transportation infrastructure and observe changes in
assignments and total distance

14