UNIT -3

Assignment Problem:

3.1 Introduction:
1. What is the Assignment Problem? Explain its importance and
applications.

Answer:

The Assignment Problem is a fundamental type of optimization problem in which the goal is to
assign a number of resources (like workers, machines, or jobs) to an equal number of tasks in the
most efficient way, often to minimize cost or maximize profit or efficiency.

It is a special case of the Transportation Problem in Operations Research where:

• The supply and demand at all sources and destinations is exactly one.
• The assignment must be one-to-one (each job is assigned to one person, and each person
gets only one job).

Importance:

• Helps in optimizing resources

• Reduces time, cost, or manpower waste
• Improves decision-making in operations and logistics
• Used in automated systems and AI for optimal task allocation

Applications:

• Assigning employees to tasks

• Allocating machines to jobs
• Scheduling nurses to shifts
• Assigning vehicles to routes
• Matching students to schools in educational systems

2. Explain the mathematical formulation of the Assignment Problem.

Answer:

Let there be n agents and n tasks. Let CijC_{ij}Cij be the cost of assigning agent iii to task jjj. Define a
binary variable:
3a.What are the characteristics of the Assignment Problem?

Answer:

The Assignment Problem has the following key characteristics:

1. One-to-One Matching: Each agent is assigned to only one task and vice versa.
2. Square Matrix: The cost matrix is always a square matrix. If not, dummy rows or
columns are added.
3. Deterministic: All parameters (costs, agents, tasks) are known and fixed.
4. Objective: Usually to minimize total cost or maximize total profit.
5. Binary Decision Variables: Each decision variable can be only 0 or 1.
6. Special case of Linear Programming: But solved more efficiently using specialized
methods.

b. Explain the difference between the Transportation Problem and the

Assignment Problem.

Answer:
3.2 Zero-One Programming Model:
4a. Solve this Zero-One Programming Problem:

A company is considering three projects. The profit and cost are as follows:
b.What are the main challenges in solving Zero-One Programming Models?

Answer:

• Non-convex solution space makes the problem hard to solve using traditional LP
methods.
• It often requires combinatorial optimization techniques like:
o Branch and Bound
o Cutting Planes
o Heuristic or Metaheuristic methods (e.g., Genetic Algorithms)
• Solving time increases exponentially with the number of variables.

c. How is Zero-One Programming different from Linear Programming?

5a. Formulate a Zero-One model for selecting products to advertise within
a fixed budget.

Answer:

Suppose a company wants to choose among 5 products (P1 to P5) to advertise. Each product has an
associated advertising cost and expected return (profit). The total budget is $100,000.
b. What is the role of constraints in a Zero-One Programming problem?

Answer:

Constraints define the feasibility and boundaries of a Zero-One Programming problem.

They ensure that the solution:

• Respects limits such as budgets, time, capacity, or resources.

• Enforces logical conditions like "only one project can be chosen" or "at least two
workers must be assigned."
• Prevents unrealistic or invalid combinations of 0-1 decisions.

For example, in a staffing model:

• Constraint: Each shift must be covered by exactly one staff member.

• Constraint: A staff member cannot be assigned to two overlapping shifts.

Thus, constraints make sure the selected binary choices result in a practical and
implementable solution.

6a. Explain with an example how a real-world scheduling problem can be

modeled using 0-1 programming.

Answer:

Let’s consider a nurse scheduling problem in a hospital.

There are 3 nurses (N1, N2, N3) and 3 shifts (Morning, Evening, Night). Each nurse can
work only one shift per day, and each shift must be assigned to exactly one nurse.

Let:
 • xij=1x_{ij} = 1xij=1 if Nurse iii is assigned to Shift jjj, else 0.

Objective: Minimize total cost or maximize preferences (if given).

Constraints:

• Each shift must be covered:

b.Why can't we just round off LP solutions to get 0 or 1 in binary decision

problems?

Answer:

Rounding off LP (Linear Programming) solutions to get 0 or 1 values does not guarantee a
feasible or optimal solution for the original binary problem. Here's why:

• Violates constraints: Rounding may produce a solution that breaks constraints,

such as exceeding budgets or assigning multiple jobs to one person.
• Inaccurate results: LP relaxation may find a solution with fractional values (e.g.,
0.6, 0.4), which don't make logical sense in binary decisions (e.g., assigning 0.6 of a
person to a task).
• Suboptimal: The rounded solution may not be the best or lowest-cost integer
solution. It can actually be worse than another valid binary combination.

3.4 Types of Assignment Problem:

7. Explain the types of Assignment Problems.

Answer:

The Assignment Problem is a type of optimization problem that involves assigning resources
(agents) to tasks in a way that minimizes cost or maximizes profit. The problem can be
classified into several types, depending on the nature of the cost or profit structure and the
constraints. The main types of Assignment Problems are:

1. Minimization Assignment Problem:

o Objective: Minimize the total cost of assigning tasks to agents.
o Example: A company wants to assign jobs to workers in a way that minimizes
the total wage cost, given a cost matrix of assigning each worker to a specific
job.
o Solution Method: Can be solved using the Hungarian Method or Branch
and Bound.
2. Maximization Assignment Problem:
o Objective: Maximize the total profit from assigning tasks to agents.
o Example: A sales company wants to assign salespeople to clients in a way
that maximizes the total sales profit, given a profit matrix.
o Solution Method: The problem can be converted into a Minimization
Problem by subtracting each value from the largest value in the matrix, and
then solving using the Hungarian method.
3. Unbalanced Assignment Problem:
o Objective: This type occurs when the number of agents does not match the
number of tasks.
o Solution: We balance the problem by adding dummy tasks or dummy
agents to make the matrix square.
o Example: A factory has more tasks than workers, so dummy workers are
added to the matrix with zero cost to balance the system.
4. Balanced Assignment Problem:
o Objective: The number of agents and tasks are equal, so no dummy agents or
tasks are needed.
o Solution: Solved using methods like the Hungarian Method or Simplex for
optimal task allocation.
o Example: A team of 5 engineers needs to be assigned to 5 projects.
5. Dual Assignment Problem:
o Objective: Involves assigning pairs of tasks to agents in such a way that both
task assignment and agent task pairing are optimized.
o Example: Pairing doctors to patients or assigning teachers to classes in a
balanced and efficient way.
6. Multi-Criteria Assignment Problem:
o Objective: Involves multiple objectives, such as minimizing cost while
maximizing quality or efficiency.
o Solution: This type often requires optimization techniques like Goal
Programming or Pareto Efficiency to handle multiple objectives
simultaneously.

Each type of Assignment Problem can be solved efficiently using specific algorithms suited
to the structure of the problem.

8a. What is the difference between a balanced and unbalanced Assignment

Problem? Provide examples.

Answer:
The balanced and unbalanced Assignment Problems differ primarily in the relationship
between the number of agents and tasks.

1. Balanced Assignment Problem:

o Definition: The number of tasks is equal to the number of agents. There is no
need for any adjustments.
o Example: A company has 5 workers, and 5 tasks. The cost matrix will be 5x5,
where each worker is assigned exactly one task.
o Solution: The Hungarian Method can be applied directly since the matrix is
already square.
2. Unbalanced Assignment Problem:
o Definition: The number of tasks is not equal to the number of agents. In this
case, the problem is unbalanced, and it needs to be balanced before solving.
o Example: A company has 6 workers but only 5 tasks. The cost matrix will be
6x5, which is unbalanced.
o Solution: The unbalanced problem can be made balanced by introducing
dummy tasks (or agents). These dummy entities are assigned zero cost, and
the matrix becomes square. After this, the problem can be solved using the
Hungarian Method.

b. Explain the difference between a minimization and maximization

Assignment Problem. How can you convert one to the other?

Answer:

In an Assignment Problem, the goal is to either minimize costs or maximize profits based
on the given cost/profit matrix. Here's the difference:

1. Minimization Assignment Problem:

o Objective: Minimize the total cost or total distance.
o Example: Assigning workers to jobs in such a way that minimizes the total
cost of assignments.
o Solution Method: The Hungarian Method is often used, where the cost
matrix is directly used without modification.
2. Maximization Assignment Problem:
o Objective: Maximize the total profit or total efficiency.
o Example: Assigning salespeople to clients in a way that maximizes total sales
revenue.
o Solution Method: To solve this type, we convert the problem into a
minimization problem:
▪ Subtract each element in the cost/profit matrix from the largest value in
that matrix.
▪ After conversion, apply the Hungarian Method as if it were a
minimization problem.

Example of Conversion: If the profit matrix is:

567856785\begin{matrix} 5 & 6 & 7 \\ 8 & 5 & 6 \\ 7 & 8 & 5 \end{matrix}

9a. How do you solve an unbalanced Assignment Problem? Provide a step-
by-step approach.

Answer:

To solve an unbalanced Assignment Problem, follow these steps:

1. Identify the imbalance: Check if the number of agents and tasks are unequal.
o If there are more agents than tasks, you will add dummy tasks.
o If there are more tasks than agents, you will add dummy agents.
2. Add Dummy Tasks or Agents:
o If there are more agents than tasks, add dummy tasks (with zero cost).
o If there are more tasks than agents, add dummy agents (with zero cost).

For example, if you have 6 agents and only 5 tasks, add one dummy task, making the
matrix square (6x6).

3. Update the Cost Matrix:

o Replace the added dummy entries with zero in the cost matrix.
4. Apply the Hungarian Method:
o Use the Hungarian Method or any other suitable method to solve the
balanced matrix.
5. Interpret the Solution:
o Ignore any assignments involving the dummy tasks or agents, as they
represent non-assignments or irrelevant tasks.

Example: If you have 6 workers (agents) and 5 tasks, your original matrix might look like
this:
b. Give an example of a real-world scenario where the Dual Assignment
Problem can be applied.

Answer:

A Dual Assignment Problem can arise in scenarios where you need to assign two sets of
agents to two sets of tasks while optimizing both assignments. This situation arises when
both task assignments and pairings must be optimized.

Example:

• Scenario: A hospital needs to assign doctors to patients, but at the same time,
specialty needs and patient preferences must be considered. Each doctor has a
specialty (e.g., Cardiology, Neurology, etc.), and each patient requires specific care
based on their condition. The hospital needs to pair doctors with patients such that
both the doctor’s specialty matches the patient’s needs, and the pairing maximizes the
total patient satisfaction while optimizing doctor allocation.

Solution:

• The dual assignment problem involves assigning both the doctors to tasks (patients)
and optimizing the pairing of specialties to maximize satisfaction, which can be
modeled as a dual assignment problem.

3.5 Hungarian Method:

10.What is the Hungarian Method? Explain its steps in detail.
Answer:

The Hungarian Method is an optimization algorithm used to solve Assignment Problems.

It finds the optimal way to assign n agents to n tasks such that the total cost is minimized (or
profit is maximized). The method is named after Hungarian mathematician Harold Kuhn,
who developed it based on earlier work by Dénes Kőnig and others.

Steps to Apply the Hungarian Method (Minimization Case):

Let’s say we have a cost matrix CCC of order n×nn \times nn×n:

1. Row Reduction:
o Subtract the smallest element of each row from all the elements in that row.
o This ensures each row has at least one zero.
2. Column Reduction:
o After row reduction, subtract the smallest element of each column from all
the elements in that column.
o This step ensures each column has at least one zero.
3. Cover All Zeros:
o Cover all zeros in the matrix using the minimum number of horizontal and
vertical lines.
o If the number of lines = nnn, then an optimal assignment exists among those
zeros.
4. Test for Optimality:
o If the number of lines used < nnn, proceed to the next step.
o Otherwise, proceed to assignment and solve.
5. Modify the Matrix:
o Find the smallest uncovered value (i.e., not covered by any line).
o Subtract it from all uncovered elements.
o Add it to all elements covered twice (intersection of lines).
o Repeat steps 3 and 4 until an optimal assignment is found.
6. Make the Optimal Assignment:
o Choose a zero in the matrix, assign it, and eliminate all other zeros in that row
and column.
o Repeat until all agents are assigned exactly one task.

11. Solve the following assignment problem using the Hungarian Method:

Cost Matrix (Minimization):

Solution:

Step 1: Row Reduction

Subtract the smallest element in each row:

• Row 1: [7, 0, 5, 6]
• Row 2: [3, 1, 0, 4]
• Row 3: [4, 7, 0, 7]
• Row 4: [3, 2, 5, 0]

Step 2: Column Reduction

Subtract the smallest element in each column:

Step 3: Cover Zeros

Use the minimum number of lines to cover all zeros:

• Cover Row 2 (0 at [2,1])

• Cover Column 2 (0 at [1,2])
• Cover Column 3 (zeros at [2,3] and [3,3])
• Cover Column 1 (zeros at [2,1] and [4,1])
• Total lines = 4 (same as matrix order) → Optimality achieved
 Step 4: Assignment

Select zeros without conflict:

• Man 1 → Job 2
• Man 2 → Job 1
• Man 3 → Job 3
• Man 4 → Job 4

Total minimum cost = 2 + 6 + 1 + 4 = 13

12a. How do you adapt the Hungarian Method for a Maximization

Assignment Problem?

Answer:

In a maximization problem, the objective is to maximize the total profit or efficiency.

Since the Hungarian Method is based on minimization, we convert the maximization matrix
to a minimization form before applying the method.

Steps:

1. Find the largest value in the matrix.

2. Subtract every element in the matrix from this maximum value to get a new cost
matrix.
3. Apply the Hungarian Method on the converted matrix.
4. The result will yield the maximum total value from the original matrix.

Example:

• Max value = 9
• Convert matrix:
b. What are the advantages and limitations of the Hungarian Method?

Answer:

Advantages:

• Provides an exact optimal solution.

• Can solve both balanced and unbalanced problems (with dummy rows/columns).
• Works well for problems with small to moderate size matrices.
• Efficient with time complexity O(n3)O(n^3)O(n3), suitable for many practical
applications.

Limitations:

• Not suitable for very large datasets due to memory and computational limits.
• Requires the problem to be in a square matrix format (unbalanced problems need
modification).
• The standard method only applies to one-to-one assignment problems (not suitable
for many-to-many assignments).
• Modifications are needed for problems involving multiple criteria or dynamic
changes.

c. Where is the Hungarian Method used in real life? Provide examples.

Answer:

The Hungarian Method is used in many real-life optimization problems where resource-
task assignment is required:

1. Workforce Assignment:
o Assigning employees to jobs/tasks based on skill, cost, or preference.
o Example: Scheduling shift workers in hospitals or factories.
2. College Seat Allocation:
o Assigning students to universities or courses to maximize preferences or cutoff
scores.
3. Machine Scheduling:
o Assigning machines to tasks in manufacturing processes to minimize time or
cost.
4. Airline Crew Scheduling:
 oAssigning pilots and crew members to flights to minimize layover costs and
maximize efficiency.
5. Matching in Online Platforms:
o Platforms like Uber or food delivery services use variations of the Hungarian
Method to assign riders to drivers.

3.6 Branch-and-Bound Technique for Assignment Problem:

13. What is the Branch-and-Bound Technique? How is it applied to
Assignment Problems?

Answer:

The Branch-and-Bound (B&B) technique is a systematic method for solving

combinatorial optimization problems, especially when the problem requires integer
solutions or has additional constraints. It is particularly useful for solving Assignment
Problems that are non-linear, involve complex constraints, or require binary decisions.

Steps to Apply Branch-and-Bound in Assignment Problems:

1. Initialization (Root Node):

o Begin with the original problem — this is your root node.
o Apply a relaxation technique (like the Hungarian Method) to find a lower
bound for minimization (or upper bound for maximization).
2. Branching:
o Create subproblems (nodes) by choosing an assignment (e.g., assign Task 1 to
Agent A) and creating two branches:
▪ One includes the assignment (force that pair).
▪ One excludes the assignment (disallow that pair).
3. Bounding:
o For each new node, solve the relaxed problem again to find a bound.
o If the bound is worse than the best known solution, prune that branch (do not
explore it further).
4. Fathoming (Pruning):
o If a node yields a feasible and complete assignment, update the best-known
solution.
o Discard (fathom) nodes whose lower bounds exceed the current best solution.
5. Repeat Until Optimal:
o Continue branching and bounding until all possible nodes are either fathomed
or yield optimal integer solutions.

14. Give an example of solving a small assignment problem using the

Branch-and-Bound method.
Answer:

Problem:

Step 1: Initial Node (Relaxed Problem)

Apply Hungarian method to get initial cost (e.g., 205). This becomes our initial lower
bound.

Step 2: Branch on Assignment (A assigned to Job 1)

Create two branches:

• Left Branch (Include A–Job 1): Remove Row A and Column 1. Solve subproblem
with remaining matrix.
• Right Branch (Exclude A–Job 1): Set cost of A–Job 1 to ∞. Solve new matrix.

Step 3: Compute Bounds

Solve each subproblem using the Hungarian method and compute new bounds (e.g., Left =
220, Right = 210). Choose branch with smaller bound to explore next.

Step 4: Continue Branching

Repeat the process, creating subproblems and evaluating bounds.

Final Step: Optimal Solution Found

Once all nodes are either fathomed or yield complete assignments, choose the one with
minimum cost — this is your optimal assignment.

15a. What are the advantages and disadvantages of the Branch-and-Bound

method for Assignment Problems?

Answer:

Advantages:

• Guaranteed optimality: Always finds the optimal solution.

• Flexible: Can handle complex constraints and integer requirements not supported
by methods like the Hungarian Method.
• Systematic pruning: Efficiently reduces the number of possibilities using bounds.

Disadvantages:
 • Computationally expensive: Becomes slow for large-scale problems due to
exponential growth of branches.
• Memory intensive: Requires tracking many subproblems in memory.
• Not ideal for real-time decision-making unless problem size is small or well-
structured.

b. In what scenarios is the Branch-and-Bound technique preferred over the

Hungarian Method?

Answer:

Branch-and-Bound is preferred in the following situations:

1. Binary/Integer Constraints: When assignments must be strictly 0 or 1 (e.g.,

assigning only one unique option to a task, or if task must be accepted or rejected).
2. Non-standard Constraints: If the assignment problem has additional constraints
like:
o Task A cannot be done by Person X.
o Person Y must be assigned at least 2 tasks.
o Some assignments are forbidden or conditionally allowed.
3. Unstructured Cost Matrices: When the cost matrix is irregular or contains penalties
for some assignments that need to be explicitly avoided.
4. Multi-objective Optimization: When the assignment must be made based on
multiple conflicting objectives (e.g., cost and time).
5. Resource-Constrained Assignment: When there are limitations like time,
manpower, or cost that cannot be modeled in a regular cost matrix.

3.7Integer Programming: Introduction:

16a.What is Integer Programming? Explain its importance and types.

Answer:

Integer Programming (IP) is a special class of mathematical optimization problems in

which some or all decision variables must take integer values. Unlike linear programming
(LP), which allows fractional values, IP ensures that the solutions are discrete, making it
ideal for real-world problems where partial decisions are not practical (e.g., number of
workers, machines, vehicles).

Types of Integer Programming:

1. Pure Integer Programming (PIP):

o All decision variables must be integers.
o Example: Assigning jobs to workers.
2. Mixed Integer Programming (MIP):
o Some variables are restricted to integers, others can be continuous.
 oExample: Number of trucks (integer) and amount of cargo (continuous).
3. Binary (Zero-One) Integer Programming:
o Variables can only take values 0 or 1.
o Used for decision-making problems (e.g., Yes/No choices,
inclusion/exclusion).

Importance of Integer Programming:

• Models real-world discrete decisions accurately.

• Used in resource allocation, scheduling, logistics, capital budgeting, and facility
location problems.
• Essential in scenarios where LP solutions are not feasible or realistic due to
fractional values.

b. What are the key differences between Linear Programming and Integer
Programming? Why is IP more complex?

Answer:

17a.Explain with an example how Integer Programming is applied in real-

life business decisions.

Answer:

Example: Product Mix Decision in a Factory

A factory produces two types of products: A and B. Each unit of A requires 3 hours of labor
and each unit of B requires 2 hours. The total labor available is 18 hours. Product A gives a
profit of $5, and product B gives $4.

Additionally:

• The factory can produce only whole units (no partial products).
• A minimum of 2 units of A must be produced.
Objective: Maximize profit
Decision Variables:

• x: Number of product A units

• y: Number of product B units

Formulate the Integer Program:

Maximize:

b. What are the main methods used to solve Integer Programming

problems? Explain briefly.

Answer:

Here are the main techniques:

1. Branch-and-Bound:
o Systematically explores branches of the decision tree.
o Solves LP relaxations and prunes suboptimal or infeasible branches.
2. Cutting Plane Method:
o Starts with LP relaxation.
o Adds linear constraints (cuts) to eliminate fractional solutions.
3. Gomory Cuts:
o Specific type of cutting plane used for pure IP problems.
4. Branch-and-Cut:
 oCombines Branch-and-Bound with Cutting Plane techniques.
5. Heuristic and Metaheuristic Methods:
o Includes Genetic Algorithms, Simulated Annealing, Tabu Search.
o Used for large, complex, or non-linear integer problems.
6. Dynamic Programming (for small cases):
o Breaks down problems into stages with decision rules.

18a. What are the challenges in solving Integer Programming problems?

Answer:

1. Computational Complexity:
IP problems are NP-Hard. Even small problems can have huge solution spaces.

2. Non-Convex Feasible Region:

The feasible region is discrete, making LP-based optimization difficult.

3. Lack of Efficient Exact Methods for Large Problems:

Exact methods (Branch-and-Bound, Cutting Planes) can become slow as problem size
increases.

4. Multiple Local Optima:

Heuristics may find local optima instead of global, especially in non-linear IP problems.

5. Model Formulation Difficulty:

IP models are harder to formulate and interpret compared to LP models, especially when
incorporating logical constraints.

b. How does Integer Programming apply to network design and

scheduling? Give examples.

Answer:

1. Network Design:

• Used to decide whether to build a link (0-1 decision).

• Example: Telecom Network Expansion – deciding which routes to establish to
minimize cost.

2. Project Scheduling:

• Assigning tasks to time slots or resources.

• Example: Job-shop Scheduling – assigning jobs to machines to minimize total
completion time.

3. Staff Scheduling:

• Deciding shifts for employees such that coverage is ensured.

 • Example: Nurse Scheduling in hospitals with labor constraints.

4. Capital Budgeting:

• Selecting investment projects under budget constraints.

• Projects are either selected (1) or not (0).

3.8 Integer Programming Formulations:

19a.What is an Integer Programming formulation? Why is it important in
operations research?

Answer:

An Integer Programming (IP) formulation is the process of translating a real-world

decision-making problem into a mathematical model that uses integer (whole number)
decision variables, along with an objective function and constraints. Unlike regular linear
programming, IP ensures decisions are feasible, realistic, and practical, especially when
fractional values are not acceptable.

Importance of IP Formulation:

1. Accurate modeling of real-life decisions, such as assigning workers, scheduling

jobs, or selecting projects.
2. Allows for yes/no (binary) decisions.
3. Ensures compliance with constraints, such as budget, capacity, or time.
4. Essential in resource allocation, facility location, production planning, and
network design.

b. Formulate an integer programming model for a capital budgeting

problem.

Answer:

Problem:

A company has $100,000 to invest in projects. There are 5 potential projects with different
costs and expected profits:
20. How can you model a facility location problem using Integer
Programming?

Answer:

Problem:

A company wants to decide whether to open warehouses in three cities (A, B, C) to serve four
markets. Each warehouse has a fixed cost, and serving markets from warehouses has variable
shipping costs. The goal is to minimize total cost while ensuring each market is served from one
open warehouse.
Formulation:

Let:

21. Formulate an IP model for a staff scheduling problem.

Answer:

Problem:

A company needs a minimum number of workers each day. Full-time staff work 5 consecutive days.
Determine the number of staff starting each day (Monday to Sunday) to meet daily requirements at
the minimum cost.

Formulation:
3.9 The Cutting-Plane Algorithm:
22. What is the Cutting-Plane Algorithm in Integer Programming? Explain
the concept with an example.

Answer:

The Cutting-Plane Algorithm is an exact method for solving integer programming (IP)
problems. It begins by solving the linear programming (LP) relaxation of the problem
(ignoring integer constraints), then iteratively adds new constraints (called cuts) to
eliminate the non-integer solutions without removing any feasible integer points.

Core Idea:

• Solve the LP relaxation of the IP problem.

• If the solution is integer, it’s optimal.
• If not, generate a cutting plane — a linear constraint that excludes the current non-
integer solution but includes all integer feasible solutions.
• Add this cut to the model and repeat.

This process continues until an integer solution is found.

Example:

Suppose we want to solve:

Maximize:

Z=3x+2y

Subject to:

23a. How are cutting planes generated? What is the Gomory Cut?
Answer:
Cutting planes are derived from the tableau of the simplex method used to solve the LP relaxation.
They are constructed to cut off non-integer solutions by analyzing the fractional parts of the basic
variables in the final tableau.

Gomory's Cutting Plane Method:

The most common type of cut is the Gomory cut, introduced by Ralph Gomory. It is derived from a
constraint in the final simplex tableau that corresponds to a basic variable with a fractional value.

Let’s say in the final tableau you get a row like:

b. Compare the Cutting-Plane Algorithm with the Branch-and-Bound

method. When is each more suitable?

Answer:

Feature Cutting-Plane Algorithm Branch-and-Bound Method

Adds constraints (cuts) to exclude non- Explores branches of decision
Approach
integer points tree
LP Relaxation
Continually refined with cuts Solved at each node
Used
Mixed IP problems, complex
Best For Smaller IP problems, pure integer
logic
Slower if branching grows
Speed Faster for tight LP relaxations
too large
Can be large due to tree
Memory Usage Generally efficient
storage
Can Be Yes, uses cutting planes
Yes, forms Branch-and-Cut
Combined? within B&B

When to Use Cutting-Plane:

• When the LP relaxation is tight, meaning its solution is close to integer.

• In pure integer problems with clear fractional patterns.

When to Use Branch-and-Bound:

• When the problem involves both integer and continuous variables.

• When constraints are nonlinear or logic-based.
3.10 Branch-and-Bound Technique:
24.What is the Branch-and-Bound Technique? Explain its working
principle with an example.

Answer:

The Branch-and-Bound (B&B) technique is a powerful algorithm used to solve

combinatorial optimization problems, particularly integer programming and mixed-
integer programming problems. It systematically explores branches of a decision tree by
dividing the problem into smaller subproblems (branching) and calculating bounds to
eliminate non-promising solutions (bounding).

Working Principle:

1. Relaxation:
o Solve the linear programming (LP) relaxation of the original integer problem
(ignore integer constraints temporarily).
2. Branching:
o If the solution contains non-integer values, create two subproblems by placing
additional constraints (e.g., x≤⌊x∗⌋), where x∗ is the fractional solution.
3. Bounding:
o Solve the LP relaxation of each subproblem to determine the lower bound
(minimization) or upper bound (maximization).
o Compare these bounds with the current best (incumbent) solution.
4. Fathoming (Pruning):
o Eliminate (prune) branches that:
▪ Yield infeasible solutions.
▪ Have bounds worse than the incumbent solution.
▪ Produce an integer solution (potentially updating the incumbent).
5. Iteration:
o Repeat the branching and bounding process until all branches are explored or
pruned.
6. Termination:
o The best feasible integer solution found becomes the optimal solution.

Example:
25.aWhat are the key components of the Branch-and-Bound technique?

Answer:

The Branch-and-Bound method consists of the following key components:

1. Branching:
o Dividing the problem into smaller subproblems by adding constraints.
o Typically done using floor and ceiling functions on fractional values of
decision variables.
2. Bounding:
o Calculating lower bounds (for minimization) or upper bounds (for
maximization) for each subproblem using LP relaxation.
o Used to assess whether further exploration of a branch is necessary.
3. Incumbent Solution:
o The best feasible (integer) solution found so far.
o Used as a reference to prune branches with worse bounds.
4. Fathoming (Pruning):
o Eliminating subproblems that:
 ▪Are infeasible.
▪Have worse bounds than the incumbent solution.
▪Have achieved integer optimality.
5. Search Strategy:
o Methods like depth-first search, breadth-first search, or best-first search to
explore branches efficiently.

b. Discuss the advantages and disadvantages of the Branch-and-Bound

method.

Answer:

Advantages:

1. Guaranteed optimality:
o Finds the optimal solution by systematically exploring feasible solutions.
2. Handles complex constraints:
o Suitable for problems with mixed-integer variables and additional logical
constraints.
3. Pruning improves efficiency:
o Eliminates unnecessary branches, reducing computational effort.
4. Flexibility:
o Can be applied to a wide range of combinatorial optimization problems (e.g.,
assignment, knapsack, traveling salesman).

Disadvantages:

1. Computationally expensive:
o The solution space can grow exponentially with more variables, increasing
computation time.
2. Memory intensive:
o Requires tracking and storing multiple subproblems simultaneously.
3. Inefficiency in large-scale problems:
o Even with pruning, solving very large or complex problems can be slow.
4. Depends on initial bounds:
o Weak initial bounds can result in less effective pruning.

26.. Explain how the Branch-and-Bound technique is used for solving

Assignment Problems with additional constraints.

Answer:

Assignment Problems typically involve assigning tasks to agents while minimizing costs or
maximizing profits. However, when additional constraints (e.g., specific tasks cannot be
assigned to certain agents) are introduced, the Hungarian Method may not be suitable,
making Branch-and-Bound a better option.
 Steps for Assignment Problem with Additional Constraints:

1. Formulate the problem:

o Define decision variables (usually binary xijx_{ij}xij, where xij=1x_{ij} =
1xij=1 if agent iii is assigned to task jjj).
2. Solve LP relaxation:
o Ignore binary restrictions and solve to get a lower/upper bound.
3. Check for integrality:
o If all xijx_{ij}xij are 0 or 1, the solution is optimal.
o If not, identify a fractional xijx_{ij}xij and create subproblems using
branching.
4. Apply branching:
o Add constraints such as xij=0x_{ij} = 0xij=0 (exclude the assignment) and
xij=1x_{ij} = 1xij=1 (force the assignment) to create two new branches.
5. Compute bounds for each subproblem:
o Solve LP relaxation for each subproblem and compare with the incumbent
solution.
6. Prune non-promising branches:
o Eliminate branches with bounds worse than the current best solution or
infeasible subproblems.
7. Repeat until optimal solution is found.

27a.. What are some real-world applications of the Branch-and-Bound

technique?

Answer:

The Branch-and-Bound technique is widely used in solving complex decision-making

problems across various industries, including:

1. Operations & Supply Chain:

o Facility location planning.
o Vehicle routing and logistics optimization.
o Scheduling production tasks.
2. Finance & Investment:
o Capital budgeting with binary project selection.
o Portfolio optimization with investment limits.
3. Telecommunications:
o Network design and routing optimization.
4. Manufacturing:
o Job-shop scheduling and machine assignment.
5. Resource Allocation:
o Workforce scheduling with specific constraints.
o Project selection with budget limitations.
6. Traveling Salesman Problem (TSP):
o Finding the shortest route visiting all cities.
7. Cutting Stock Problem:
o Determining optimal cutting patterns to minimize waste.
b. Differentiate between Branch-and-Bound and Cutting-Plane methods.
Can they be used together?

Answer:

Aspect Branch-and-Bound Cutting-Plane

Divides the problem into smaller
Adds linear constraints (cuts) to
subproblems (branching) and calculates
Approach remove non-integer solutions
bounds (bounding) to prune non-promising
from the LP relaxation.
branches.
Mixed-integer and combinatorial Pure integer problems where LP
Best for
problems. relaxation is tight.
Systematic tree search (depth-first,
Exploration Iterative constraint addition.
breadth-first, best-first).
Can handle complex logic constraints; may Can be computationally faster for
Efficiency
require more memory. smaller problems.
Can be combined with cutting planes to
Forms the "cutting" component in
Combination form Branch-and-Cut for improved
Branch-and-Cut.
efficiency.

Yes! These methods are commonly used together in modern solvers (Branch-and-Cut) to
leverage the advantages of both techniques, leading to faster convergence for large-scale
problems.

3.11 Zero-One Implicit Enumeration Algorithm:

28a.What is the Zero-One Implicit Enumeration Algorithm? Explain its
significance in solving 0-1 integer programming problems.

Answer:

The Zero-One Implicit Enumeration Algorithm is a specialized method for solving binary
(0-1) integer programming problems. These problems involve decision variables that can
take on only two values: 0 (no) or 1 (yes). The algorithm systematically explores all feasible
combinations of variable assignments, while implicitly eliminating large portions of the
solution space using bounding and feasibility checks.

Significance:

• Used when branch-and-bound or cutting-plane methods are not efficient due to the
problem structure.
• Focuses on logical variable fixing rather than solving LP relaxations.
• Helps solve problems like project selection, investment decisions, facility location,
and scheduling where decisions are binary.
b. Explain the steps involved in the Zero-One Implicit Enumeration
Algorithm.

Answer:

The algorithm follows these main steps:

Step 1: Generate All Nodes (Implicitly)

• Consider a binary tree where each level represents a decision variable.

• Each node corresponds to a partial solution (some variables fixed to 0 or 1, others
undecided).

Step 2: Branching

• At each node, fix the next unfixed variable to 0 or 1, creating two branches.
• For example, if current variable x1x_1x1 is unfixed:
o Branch 1: x1=0x_1 = 0x1=0
o Branch 2: x1=1x_1 = 1x1=1

Step 3: Feasibility Check

• At each node, check whether the partial assignment satisfies the problem constraints
(i.e., whether it's feasible).

Step 4: Bounding

• Calculate an upper bound (for maximization problems) or lower bound (for

minimization) for each partial solution.
• If the bound is worse than the current best (incumbent) solution, prune the node.

Step 5: Continue Enumeration

• Repeat the process recursively, moving down the tree.

• Update the best solution if a better complete feasible solution is found.

Step 6: Terminate

• The algorithm terminates when all feasible nodes are either explored or pruned.
• The best feasible solution found is the optimal solution.

29. Illustrate the Zero-One Implicit Enumeration Algorithm with an

example.

Answer:

Let’s solve the following problem:

30a.What are the advantages and limitations of the Zero-One Implicit
Enumeration Algorithm?

Answer:

Advantages:

1. Tailored for binary problems:

o Directly handles 0-1 variables, avoiding rounding issues from LP relaxation.
2. Efficient pruning:
 o Can eliminate large portions of the search tree early.
3. No need for LP solvers:
o Especially useful for small to medium-sized problems with complex logic.
4. Flexible and logical branching:
o Easier to encode logical relationships between binary decisions.

Limitations:

1. Exponential time in worst-case:

o For large problems, enumeration can still become computationally expensive.
2. No general-purpose solver implementation:
o Not commonly available in off-the-shelf solvers like branch-and-bound.
3. Performance is structure-dependent:
o Efficient pruning requires problems with strong constraint structures.
4. Memory usage can be high for deep recursion or large solution trees.

b. What are some applications of the Zero-One Implicit Enumeration

Algorithm?
Answer:

The algorithm is especially useful in problems where decisions are naturally binary and constraints
are easily evaluated. Key application areas include:

1. Project Selection:
o Selecting a subset of projects under budget or resource constraints.
2. Capital Budgeting:
o Choosing investments where each project is either accepted or rejected.
3. Facility Location:
o Deciding whether to open facilities (yes/no).
4. Knapsack Problems:
o Selecting items under weight and value constraints.
5. Task Assignment and Scheduling:
o Assigning tasks to resources with binary compatibility.
6. Network Design:
o Choosing which links to include in a network.

c.How does the Zero-One Implicit Enumeration Algorithm compare to

Branch-and-Bound?
Both algorithms can complement each other. For example, branch-and-bound can handle problems
with continuous variables, while implicit enumeration is often better for binary logic-based
problems with tight feasibility checks.

Filling the blanks:

1. The assignment problem is a special case of the __________ problem.

Answer: Linear programming

2. The main objective of the assignment problem is to minimize the total __________ or
maximize the total __________.

Answer: Cost, profit

3. The assignment problem deals with assigning _________ to _________ on a one-to-

one basis.

Answer: Tasks, agents (or jobs, persons)

4. In a balanced assignment problem, the number of _________ equals the number of

_________.

Answer: Tasks, agents

5. The solution to the assignment problem is represented using a _________ matrix.

Answer: Cost

6. In an assignment problem, each agent can be assigned to _________ task(s) only.

Answer: One

7. If the number of agents is not equal to the number of tasks, we add a _________ to
balance the matrix.
Answer: Dummy row or column

8. The assignment problem assumes that all assignments are made simultaneously and
_________.

Answer: Independently

9. The Hungarian method is one of the most widely used methods to solve the
_________ problem.

Answer: Assignment

10. The assignment problem is a special case of the transportation problem where all
supplies and demands are equal to _________.

Answer: One

11. The Hungarian Method starts by subtracting the _________ element from each row.

Answer: Minimum

12. After row reduction, the next step in the Hungarian Method is _________ reduction.

Answer: Column

13. The solution is optimal when the minimum number of lines covering all zeros is equal
to the number of _________.

Answer: Rows (or columns)

14. If an optimal assignment cannot be made, we adjust the matrix by subtracting the
smallest _________ from all uncovered elements.

Answer: Uncovered value

15. The Hungarian Method helps in achieving the _________ solution for the assignment
problem.

Answer: Optimal

16. The assignment problem is used when the resources are to be assigned in a ________
manner to tasks.

Answer: One-to-one

17. In the cost matrix of an assignment problem, each cell represents the ________ of
assigning one agent to one task.

Answer: Cost (or profit)

 18. The assignment problem assumes that each job must be assigned to exactly one
________.

Answer: Worker (or agent)

19. When profits instead of costs are given, we convert the problem into a ________
problem by subtracting values from the highest profit.

Answer: Minimization

20. If there are more tasks than agents, a ________ agent is introduced with zero cost
values.

Answer: Dummy

Choose the correct answer:

1. The Assignment Problem is a special case of:

A) Integer programming
B) Non-linear programming
C) Transportation problem
D) Simulation

Correct Answer: C) Transportation problem

2. In the Assignment Problem, each task is assigned to:

A) Multiple agents
B) Only one agent
C) At least two agents
D) None of the above

Correct Answer: B) Only one agent

3. Which method is commonly used to solve the assignment problem?

A) Big M method
B) Simplex method
C) Hungarian method
D) Newton-Raphson method

Correct Answer: C) Hungarian method

4. In the Hungarian method, the first step is:

A) Column reduction
B) Subtracting the largest value in each row
C) Row reduction
D) Adding dummy rows

Correct Answer: C) Row reduction

5. If there are more agents than tasks, what is added to the matrix to balance
it?

A) New constraints
B) Dummy task (column)
C) Slack variable
D) Artificial variable

Correct Answer: B) Dummy task (column)

6. What values are assigned to the dummy row/column in an assignment

matrix?

A) Maximum cost
B) Minimum cost
C) Zero
D) Random values

Correct Answer: C) Zero

7. What is the main objective of the assignment problem?

A) Maximize the time taken

B) Minimize total transportation
C) Minimize cost or maximize profit
D) Balance supply and demand

Correct Answer: C) Minimize cost or maximize profit

8. In an assignment problem, the number of optimal assignments equals:

A) Total number of jobs

B) Total number of agents
C) The smaller of the number of jobs or agents
D) Number of dummy variables

Correct Answer: C) The smaller of the number of jobs or agents

9. Which of the following is true about the assignment problem?

A) It always results in a degenerate solution

B) It requires fractional variable values
C) It can be solved using the Hungarian Method
D) It has no optimal solution

Correct Answer: C) It can be solved using the Hungarian Method

10. What type of cost values does the Hungarian Method operate on?

A) Only positive values

B) Any real number
C) Zero and positive values after reduction
D) Negative values only

Correct Answer: C) Zero and positive values after reduction

11. Integer Programming is a type of:

A) Non-linear programming
B) Linear programming with continuous variables
C) Linear programming with integer constraints
D) Dynamic programming

Correct Answer: C) Linear programming with integer constraints

12. If all decision variables in a model must be integers, it is known as:

A) Mixed Integer Programming

B) Binary Integer Programming
C) Pure Integer Programming
D) Linear Programming

Correct Answer: C) Pure Integer Programming

13. If only some of the variables in an optimization model must be integers,

the model is called:

A) Nonlinear programming
B) Pure Integer Programming
C) Mixed Integer Programming
D) Binary Integer Programming

Correct Answer: C) Mixed Integer Programming

14. In Binary Integer Programming, decision variables can take on values:

A) 0 to 9
B) Any positive integer
C) 0 or 1 only
D) Negative integers only
 Correct Answer: C) 0 or 1 only

15. Which of the following is NOT a method used to solve integer

programming problems?

A) Branch and Bound

B) Simplex Method
C) Cutting Plane Method
D) Graphical Method for two variables

Correct Answer: D) Graphical Method for two variables

(Graphical method is for LP problems with continuous variables, not IP)

16. The main challenge in integer programming is:

A) Too many variables

B) Discrete nature of variables makes the solution space non-convex
C) Linear constraints
D) Objective function not being quadratic

Correct Answer: B) Discrete nature of variables makes the solution space non-convex

17. Which type of integer programming is typically used in project selection

or facility location problems?

A) Mixed Integer Programming

B) Binary Integer Programming
C) Non-linear Programming
D) Dynamic Programming

Correct Answer: B) Binary Integer Programming

18. The Branch and Bound method is used to:

A) Solve non-linear equations

B) Convert fractional variables into integer
C) Handle integer constraints in optimization problems
D) Generate random solutions

Correct Answer: C) Handle integer constraints in optimization problems

19. Cutting Plane method starts with:

A) Integer solution
B) Binary solution
C) Linear programming relaxation
D) Mixed integer values
 Correct Answer: C) Linear programming relaxation

20. In Integer Programming, rounding off the LP solution may:

A) Always give the correct answer

B) Lead to optimal integer solution
C) Give an infeasible or suboptimal solution
D) Improve the solution

Correct Answer: C) Give an infeasible or suboptimal solution