Unit-1

(a) Introduction to Decision-Making Process

What is Decision-Making?

Decision-making is the process of selecting the best alternative from multiple available
options. It is used in business, finance, and daily life to achieve desired outcomes.

Steps in Decision-Making:

1. Identify the Problem – Recognize the issue that requires a decision.

2. Gather Information – Collect relevant data and information.
3. Identify Alternatives – List possible solutions.
4. Evaluate Alternatives – Analyze the benefits and drawbacks of each option.
5. Choose the Best Alternative – Select the most suitable option.
6. Implement the Decision – Put the decision into action.
7. Review the Decision – Assess the effectiveness of the decision.

Example:
A company wants to launch a new product but must decide whether to manufacture it in-
house or outsource it. It collects data on costs, risks, and benefits before making a decision.

(b) Single-Stage Decision-Making Problems

A single-stage decision-making problem occurs when a decision is made at a single point in
time without considering future consequences.

Types of Decision Environments:

1. Decision Making under Certainty – The outcome is known with certainty.

o Example: Choosing between two machines when one clearly has a lower cost
and higher efficiency.
2. Decision Making under Risk – Probabilities of outcomes are known.
o Example: Investing in a stock where past performance and market trends give
a probability of gain or loss.
3. Decision Making under Uncertainty – Probabilities of outcomes are unknown.
o Example: Launching a new product in a completely new market without past
data.

(c) Developing Pay-off & Regret Tables

1. Pay-off Table

A pay-off table shows the possible outcomes for different decisions based on different states
of nature. It helps in evaluating the best alternative.
 Example: A company is deciding between launching Product A or Product B. The
possible profits under different market conditions are:

Decision High Demand (₹) Low Demand (₹)

Product A 1,00,000 20,000
Product B 80,000 50,000

• If demand is high, Product A is better.

• If demand is low, Product B is better.

2. Regret Table (Opportunity Loss Table)

A regret table shows the difference between the best possible outcome and the actual
outcome for each decision.

Decision High Demand Regret Low Demand Regret

Product A 0 30,000
Product B 20,000 0

Regret values are calculated by subtracting actual profits from the highest possible profit in
that state.

(d) Decision Rules Using Criteria

Decision rules help in choosing the best option when dealing with uncertainty or risk.

1. Maximax Criterion (Optimistic Approach)

• Selects the decision with the highest possible outcome.

• Best for risk-takers.
• Example: Choosing Product A because it has the highest profit potential of
₹1,00,000.

2. Maximin Criterion (Pessimistic Approach)

• Selects the decision with the best of the worst outcomes.

• Best for risk-averse decision-makers.
• Example: Choosing Product B because its worst-case profit is ₹50,000 (higher than
Product A's ₹20,000).

3. Minimax Regret Criterion

• Selects the decision that minimizes the maximum regret.

• Example: Choosing Product B as it has a lower maximum regret (₹20,000 vs.
₹30,000).

4. Expected Monetary Value (EMV) – Probabilistic Approach

 • Uses probabilities to find the best option.

Example: If the probability of high demand is 0.7 and low demand is 0.3, then:

EMV(ProductA)=(1,00,000×0.7)+(20,000×0.3)=76,000EMV(Product A) = (1,00,000 \times

0.7) + (20,000 \times 0.3) = 76,000EMV(ProductA)=(1,00,000×0.7)+(20,000×0.3)=76,000
EMV(ProductB)=(80,000×0.7)+(50,000×0.3)=71,000EMV(Product B) = (80,000 \times 0.7)
+ (50,000 \times 0.3) = 71,000EMV(ProductB)=(80,000×0.7)+(50,000×0.3)=71,000

Since Product A (₹76,000) > Product B (₹71,000), Product A is the best choice.

(e) Multi-Stage Decision-Making Problems – Decision

Trees
A decision tree is a graphical representation of multiple decision stages, where each branch
represents a possible choice or event.

Components of a Decision Tree:

1. Decision Nodes (Squares) – Represent choices.

2. Chance Nodes (Circles) – Represent uncertain events with probabilities.
3. Branches – Represent different outcomes.
4. Pay-offs – Represent final outcomes (profits/losses).

Example: A company must decide whether to invest in a new factory (₹50 lakh) or do
nothing (₹0 loss/gain).

• If it invests, demand can be high (60% probability, ₹1 crore profit) or low (40%
probability, ₹20 lakh loss).

Decision Tree Representation:

[Decision: Invest or Not]
/ \
Invest (₹50L) No Investment (₹0)
/ \
High Demand Low Demand
(60% | ₹100L) (40% | -₹20L)

Expected Value Calculation:

EV=(0.6×100L)+(0.4×(−20L))=60L−8L=52LEV = (0.6 \times 100L) + (0.4 \times (-20L)) =

60L - 8L = 52LEV=(0.6×100L)+(0.4×(−20L))=60L−8L=52L

Since ₹52L is greater than ₹0, investing in the new factory is the better decision.

Unit-2
Theory of Games: Notes on Quantitative
Techniques
(a) Introduction to Game Theory and Concept of Strategy
What is Game Theory?

Game Theory is a mathematical approach to decision-making where two or more players

(decision-makers) compete or cooperate to achieve their goals. It helps in strategic decision-
making in business, economics, politics, and even war.

Key Concepts in Game Theory

1. Players – The decision-makers in the game. (Example: Two competing companies)

2. Strategies – The possible actions each player can take. (Example: High pricing or
Low pricing)
3. Payoff – The outcome (profit/loss) for each player based on their chosen strategy.
4. Rules – Defined conditions of the game, such as moves and payoffs.
5. Equilibrium – A situation where no player can improve their position by changing
their strategy unilaterally (Nash Equilibrium).

Example:
Two competing companies (A & B) must decide on a pricing strategy. If both set high prices,
they make moderate profits. If one lowers its price, it gains more market share but reduces its
profit margin.

(b) Classification of Game Models

Game models are classified based on the number of players, strategies, and payoffs.

Types of Game Models:

1. Based on Number of Players

o Two-Person Games – Involves two players. (Example: Chess, Company A
vs. Company B)
o Multi-Person Games – Involves more than two players. (Example: Poker,
Business market competition)
2. Based on Payoffs
o Zero-Sum Games – One player’s gain is equal to the other’s loss. (Example:
Betting, war strategies)
o Non-Zero-Sum Games – Players can cooperate, and both may gain.
(Example: Business partnerships)
3. Based on Strategy Choices
o Pure Strategy Games – Players use a fixed strategy always.
o Mixed Strategy Games – Players use a probability-based approach to make
choices.
4. Based on Information Availability
 o Perfect Information Games – Players know all past moves. (Example:
Chess)
o Imperfect Information Games – Players don’t know the full history.
(Example: Poker)

(c) Two-Person Zero-Sum Games – Formulation of Payoff

Table
A zero-sum game means that one player's gain is exactly equal to the other player's loss.

Payoff Table Representation

A payoff table shows the rewards or penalties for each combination of strategies chosen by
two players.

Example:
Consider two companies, A and B, competing in the market. Each can choose either
Aggressive Marketing (AM) or Normal Marketing (NM) strategy. The payoffs
(profits/losses) are given in the table below:

Company B AM NM
Company A (AM) (–5, +5) (+10, –10)
Company A (NM) (+15, –15) (0, 0)

• If Company A chooses AM and Company B chooses NM, Company A gains ₹10

and Company B loses ₹10.
• If both choose NM, no one gains or loses.

Since this is a zero-sum game, the sum of payoffs in each case is zero.

(d) Solution of Games – Algebraic Method – 2×2 and 3×3

Matrix
1. 2×2 Game Solution Using Algebraic Method

In a 2×2 matrix, each player has two strategies. The objective is to find the optimal strategy
using algebraic equations.

Example:
Consider the following payoff matrix for Player A (row player) and Player B (column
player). The values represent A’s payoffs (B’s payoffs will be negative of these values in a
zero-sum game).

B1 B2
A1 4 1
A2 3 2
Steps to Solve Algebraically:

1. Let Player A play A1 with probability p and A2 with probability (1 – p).

2. Let Player B play B1 with probability q and B2 with probability (1 – q).
3. Expected payoff for Player A:

E(A1)=4q+1(1−q)=4q+1−q=3q+1E(A1) = 4q + 1(1 - q) = 4q + 1 - q = 3q +
1E(A1)=4q+1(1−q)=4q+1−q=3q+1 E(A2)=3q+2(1−q)=3q+2−2q=q+2E(A2) = 3q +
2(1 - q) = 3q + 2 - 2q = q + 2E(A2)=3q+2(1−q)=3q+2−2q=q+2

4. Set both expected payoffs equal to find equilibrium:

3q+1=q+23q + 1 = q + 23q+1=q+2 2q=12q = 12q=1 q=0.5q = 0.5q=0.5

5. Similarly, solve for p using Player B’s equations.

Thus, A should play A1 with probability 0.5 and A2 with probability 0.5 for an optimal
strategy.

2. 3×3 Game Solution Using Algebraic Method

For a 3×3 game, similar steps are followed, but with three equations involving probabilities
p, q, r for Player A and x, y, z for Player B. The solution involves solving a system of three
linear equations.

(e) Dominance Rule – Row-wise and Column-wise

What is the Dominance Rule?

The dominance rule is used to simplify large payoff matrices by eliminating weak
strategies.

1. Row-wise Dominance

• A row is dominated if all its values are less than or equal to another row.
• The dominated row is removed.

Example:
If in a payoff matrix, row R2 has smaller values than row R1, then R2 is eliminated.

2. Column-wise Dominance

• A column is dominated if all its values are greater than or equal to another column.
• The dominated column is removed.

By applying these rules, the matrix becomes simpler, making calculations easier.

Limitations and Applications of Game Theory

Limitations:

1. Assumes Rationality – Assumes all players are logical and make rational decisions.
2. Limited to Two Players in Basic Models – Real-world situations involve multiple
players.
3. Difficult to Determine Probabilities – In uncertain conditions, probability values
may not be available.
4. Not Always Practical – In dynamic markets, real-time decision-making differs from
theoretical models.

Applications of Game Theory:

1. Business Strategy – Pricing, advertising, and competition analysis.

2. Economics – Understanding market behaviors and auctions.
3. Politics – Election campaigns and diplomatic negotiations.
4. Military Strategy – War planning and defense tactics.
5. Sports – Decision-making in competitive games like chess and football.
 Unit-3
📘 Linear Programming (LP)
(a) Introduction to Linear Programming

Definition:
Linear Programming is a mathematical technique used to allocate limited resources in an
optimal way. It helps in maximizing or minimizing a linear objective function, subject to a set
of linear constraints.

Key Points:

 “Linear” means the relationships are proportional.

 “Programming” refers to planning or optimization.

Example:
A company makes two products: A and B.
Profit on A = ₹3/unit, Profit on B = ₹5/unit.
Company wants to maximize profit under constraints like labor hours and materials.

(b) Formulation of Linear Programming Problem

Steps:

1. Define decision variables: What are we trying to decide?

2. Write the objective function: What are we maximizing or minimizing?
3. Identify the constraints: What limits the decisions?
4. Add non-negativity condition: Decision variables can’t be negative.

Example:
Let xxx = units of Product A, yyy = units of Product B
Maximize Profit: Z=3x+5yZ = 3x + 5yZ=3x+5y
Subject to:
2x+3y≤122x + 3y \leq 122x+3y≤12 (Labor hours constraint)
x+y≤5x + y \leq 5x+y≤5 (Material constraint)
x≥0,y≥0x \geq 0, y \geq 0x≥0,y≥0

(c) General Statement of LP Problems

There are two main types:

1. Maximization Case:
  Objective: Maximize profit, output, etc.
 Constraints: Resources available are limited.

Example: Maximize Z=3x+2yZ = 3x + 2yZ=3x+2y

Subject to:
x+y≤4x + y \leq 4x+y≤4
x+2y≤6x + 2y \leq 6x+2y≤6

2. Minimization Case:

 Objective: Minimize cost, time, wastage, etc.

Example: Minimize Z=4x+3yZ = 4x + 3yZ=4x+3y

Subject to:
3x+y≥123x + y \geq 123x+y≥12
x+2y≥8x + 2y \geq 8x+2y≥8

(d) Assumptions, Advantages, and Applications of LP

Assumptions:

1. Proportionality: Increase in input leads to proportional increase in output.

2. Additivity: Total effect = sum of individual effects.
3. Certainty: All coefficients are known and constant.
4. Non-Negativity: Negative production doesn’t exist.
5. Linearity: Both the objective and constraints are linear.

Advantages:

 Efficient resource utilization.

 Helps in decision-making.
 Flexible and applicable in many areas.

Applications:

 Manufacturing: To plan production.

 Transportation: Least-cost routing.
 Finance: Portfolio optimization.
 Marketing: Optimal advertising mix.
 HR: Shift scheduling.

(e) Solution of LP Problem – Graphical Method (First Quadrant Only)

Used when there are two variables only.

Steps:
 1. Convert constraints to equations.
2. Plot them on graph paper (x-axis and y-axis).
3. Identify feasible region (shaded area satisfying all constraints).
4. Find corner points (vertices) of the feasible region.
5. Evaluate the objective function at each corner point.
6. Choose the point giving maximum or minimum value.

Example:

Maximize Z=3x+5yZ = 3x + 5yZ=3x+5y

Subject to:
x+y≤4x + y \leq 4x+y≤4
x+2y≤6x + 2y \leq 6x+2y≤6
x≥0,y≥0x \geq 0, y \geq 0x≥0,y≥0

Graph:

 Draw lines for each constraint.

 Shade the feasible region in first quadrant.
 Calculate ZZZ at each corner point.

Point Z = 3x + 5y
(0,0) 0
(0,3) 15
(2,2) 16 ✅ (Max)
(4,0) 12

Optimal Solution: x=2,y=2,Z=₹16x = 2, y = 2, Z = ₹16x=2,y=2,Z=₹16

Unit-4

📘 Assignment Problem & Transportation Problem

Assignment Problem
(a) Introduction to Assignment Problem

Definition:
An assignment problem deals with assigning n tasks to n workers (or jobs to machines, persons to
projects, etc.) so that the total cost is minimized or profit is maximized.

Key Features:

 One-to-one assignment.
  Objective: Minimize cost or maximize efficiency.

Example:
Assign 3 workers (A, B, C) to 3 jobs (X, Y, Z) where cost/time is given in a matrix. Choose assignments
with the least total cost.

(b) Complete Enumeration Method (Theoretical)

This method involves:

1. Listing all possible combinations of assignments.

2. Calculating total cost for each combination.
3. Choosing the combination with the least cost (for minimization).

Limitations:

 Not practical for large problems (factorial increase).

 Only useful for understanding small-size problems.

(c) Hungarian Assignment Method (HAM)

A step-by-step algorithm to solve assignment problems efficiently.

Steps:

1. Create the Cost Matrix

Rows = workers, Columns = jobs
2. Row Reduction
Subtract the minimum value in each row from all elements of that row.
3. Column Reduction
Subtract the minimum value in each column from all elements of that column.
4. Cover all zeros using minimum number of horizontal and vertical lines
5. Test for Optimality
o If number of lines = number of rows/columns → Optimal solution found.
o Else, go to step 6.
6. Adjust the Matrix
o Find the smallest uncovered element.
o Subtract it from all uncovered elements.
o Add it to the intersection of covering lines.
o Repeat step 4.
7. Make Assignments
o Assign zero elements such that no two are in the same row/column.

Example (Simplified):

Job A Job B Job C

 W1 9 2 7
W2 6 4 3
W3 5 8 1

Use the steps above to reduce and assign.

(d) Unbalanced Assignment Problem

If number of workers ≠ number of jobs, the problem is unbalanced.

Solution:

 Add dummy row or column with zero cost to make it square.

 Then apply Hungarian Method.

Example:
3 workers, 4 jobs → Add a dummy worker (4th row) with 0 costs for each job.

(e) Constrained Assignment Problem

Sometimes assignments are not allowed (e.g., a worker can’t do a job).

Solution:

 Put a very high cost (like 999 or M) in those cells to restrict assignment.
 HAM will naturally avoid assigning such cells.

(f) Unique vs Multiple Optimal Solutions

 Unique Optimal Solution: Only one assignment combination gives the minimum total cost.
 Multiple Optimal Solutions: Two or more combinations give the same minimum cost.

Tip to identify:

 If you can make more than one independent assignment of zeros during final step in HAM,
then multiple optimal solutions exist.

(g) Maximization Case in Assignment Problems

By default, the Hungarian Assignment Method is used for minimization.

To handle maximization, we convert it to a minimization problem.

Steps:
 1. Identify the largest element in the cost/profit matrix.
2. Subtract every element in the matrix from this largest value.
3. Apply Hungarian Method on the new matrix.
4. The assignments will give the maximum total profit.

Example (Profit Matrix):

Job A Job B Job C

W1 10 15 20
W2 20 25 15
W3 15 10 30

Largest value = 30 → Subtract each from 30 to convert it to a minimization matrix.

📦 Transportation Problem
(h) Introduction to Transportation Problem

Definition:
A transportation problem deals with shipping goods from multiple sources to multiple
destinations at minimum cost, satisfying supply and demand constraints.

Structure:

 Sources (factories) with certain supply.

 Destinations (warehouses) with certain demand.
 Cost matrix showing cost per unit for each route.

Objective:
Minimize total transportation cost.

(i) Solution to Transportation Problem

To solve, we start with an initial feasible solution using any of the following methods:

1. North-West Corner Rule (NWCR)

Steps:

 Start from the top-left (north-west) cell.

 Allocate as much as possible (min of supply/demand).
 Adjust supply/demand and move right/down.
✅ Simple but not always cost-effective.

2. Least Cost Method (LCM)

Steps:

 Find the lowest cost cell in the matrix.

 Allocate as much as possible.
 Adjust supply/demand and cross out row/column once exhausted.
 Repeat with next least cost.

✅ More cost-efficient than NWCR.

3. Vogel’s Approximation Method (VAM)

Steps:

1. For each row & column, compute a penalty = difference between lowest & second
lowest cost.
2. Select the row/column with the highest penalty.
3. Allocate as much as possible in the lowest cost cell of that row/column.
4. Adjust supply/demand, cross out exhausted rows/columns.
5. Repeat until all allocations are made.

✅ Generally gives the best initial solution.

(k) Testing Optimality using MODI Method

MODI (Modified Distribution Method) checks whether the current solution is optimal.

Steps:

1. Calculate u and v values:

o Assign u1=0u_1 = 0u1=0
o Use ui+vj=Ciju_i + v_j = C_{ij}ui+vj=Cij for occupied cells to find other
values.
2. Compute opportunity cost for each unoccupied cell:
Δij=Cij−(ui+vj)\Delta_{ij} = C_{ij} - (u_i + v_j)Δij=Cij−(ui+vj)
3. Check for optimality:
o If all Δ ≥ 0 → optimal solution found.
o If any Δ < 0 → improve the solution.
Improving the Solution (Closed Loop Method)

If solution is not optimal:

1. Select the unoccupied cell with the most negative Δij\Delta_{ij}Δij.

2. Draw a closed loop:
o Start at the selected cell.
o Move alternately in horizontal and vertical lines through occupied cells only.
o Mark + and − signs alternately.
3. Adjust allocations:
o Subtract the smallest value from the − positions.
o Add it to the + positions.
4. Update the solution and re-calculate using MODI.

✅ Repeat until all Δ≥0\Delta \geq 0Δ≥0

Unit-5

📘 Queuing Theory

(a) Introduction to Queuing Theory & Applications

Definition:
Queuing Theory is the mathematical study of waiting lines or queues. It helps analyze
systems where customers arrive, wait for service, and then depart after being served.

Objective:
To balance waiting time and service efficiency by minimizing total cost (cost of waiting +
cost of service).

Real-life Applications:

 Bank counters (customer queues)

 Call centers (calls waiting for agents)
 Hospitals (patients waiting for treatment)
 Supermarkets (checkout lines)
 Computer networks (data packets waiting to be processed)

(b) General Structure of a Queuing System

A queuing system has 5 key components (often called the “Kendall Notation”):
 1. Input/Arrival Process:
o Describes how customers arrive.
o Usually follows a Poisson distribution.
2. Queue Discipline:
o Rules for serving customers.
o Most common: FCFS (First Come, First Serve).
3. Service Mechanism:
o Number of servers and service time distribution.
o Often Exponential Distribution.
4. Capacity of System:
o Maximum number of customers allowed in the system (finite/infinite).
5. Calling Source (Population):
o Can be finite (small number of users) or infinite (large, unpredictable
population).

(c) Operating Characteristics of a Queuing System

These are performance metrics used to evaluate the system:

1. λ (lambda) = Average arrival rate (customers/unit time)

2. μ (mu) = Average service rate (customers served/unit time)

Key characteristics:

 L: Average number of customers in the system (waiting + being served)

 Lq: Average number of customers in the queue
 W: Average time a customer spends in the system
 Wq: Average time a customer spends waiting in queue
 ρ (rho): Server utilization (ρ = λ / μ), should be less than 1 for stability

(d) Queuing Models (Probabilistic)

We focus on the M/M/1 model:

(M = Poisson arrival, M = Exponential service, 1 = single server)

Assumptions of M/M/1 Model:

 Arrivals follow Poisson distribution

 Service times follow Exponential distribution
 One server
 Infinite population and system capacity
 First Come First Serve
Formulas for M/M/1 Model:

Let λ\lambdaλ = arrival rate, μ\muμ = service rate

1. Server Utilization: ρ=λμ\rho = \frac{\lambda}{\mu}ρ=μλ

2. Average number in system (L):
L=λμ−λL = \frac{\lambda}{\mu - \lambda}L=μ−λλ
3. Average number in queue (Lq):
Lq=λ2μ(μ−λ)L_q = \frac{\lambda^2}{\mu(\mu - \lambda)}Lq=μ(μ−λ)λ2
4. Average time in system (W):
W=1μ−λW = \frac{1}{\mu - \lambda}W=μ−λ1
5. Average waiting time in queue (Wq):
Wq=λμ(μ−λ)W_q = \frac{\lambda}{\mu(\mu - \lambda)}Wq=μ(μ−λ)λ

Example:

 Arrival rate (λ) = 5 customers/hour

 Service rate (μ) = 8 customers/hour
 ρ=5/8=0.625\rho = 5/8 = 0.625ρ=5/8=0.625
 L=5/(8−5)=1.67L = 5 / (8 - 5) = 1.67L=5/(8−5)=1.67 customers
 W=1/(8−5)=0.333W = 1 / (8 - 5) = 0.333W=1/(8−5)=0.333 hours (20 minutes)

Cost Analysis in Queuing

Objective: To find the balance between:

1. Cost of waiting (e.g., customers' time, dissatisfaction, opportunity loss)

2. Cost of service (e.g., number of servers, wages, infrastructure)

Total Cost = Waiting Cost + Service Cost

✅ Decision-making:
Increase number of servers if waiting cost is high, or reduce servers if service cost
dominates.