Part 1 Waiting line: Customers who arrive when the service facility is busy must wait in a queue

until service becomes available. The waiting line can be finite or infinite.
1a) The formula for the coefficient of variation is:
Service facility: The service facility processes the customers who are waiting in the queue.
Coefficient of variation = (standard deviation / mean) * 100% The service rate is the average number of customers that can be served per unit of time.

We have X = 120 and N = 10, so the mean is X/N = 12. Queue discipline: The order in which customers are served can be determined by various
rules, such as first-come, first-served or priority based on the type of customer or the nature
We also have X-1530, which means X - mean = 1530 - 120 = 1410. of the service.

So, the standard deviation is sqrt((X-X̄)²/(N-1)) = sqrt(1410²/9) ≈ 165.99. Service time distribution: The time required to serve each customer can be modeled using a
probability distribution. The service time distribution affects the overall performance of the
Therefore, the coefficient of variation is (165.99 / 12) * 100% ≈ 1383.25%. queuing system.

1b) The formula for mean deviation is: 1i) The arrival rate is 12 per hour, which means the arrival rate per minute is 12/60 = 0.2.

Mean deviation = (Σ |Xi - X̄ |) / N The average service time is 4 minutes per file.

We know that the standard deviation is 60. The relationship between standard deviation The utilization factor, ρ, is the ratio of the arrival rate to the service rate, which is ρ = 0.2 /
and mean deviation is: (1/4) = 0.8.

Mean deviation = standard deviation * (pi / 2), where pi is the value of pi (3.1416). Using the M/M/1 queuing model, the average length of the waiting line, Lq, can be
calculated as:
Therefore, Mean deviation = 60 * (3.1416 / 2) ≈ 94.25.
Lq = ρ² / (1-ρ) = 0.8² / (1-0.8) ≈ 3.2.
The formula for quartile deviation is:
Therefore, the average length of the waiting line is 3.2.
Quartile deviation = (Q3 - Q1) / 2, where Q1 and Q3 are the first and third quartiles,
respectively. 1j) Kendall's notation is a standard notation used to describe queuing models. It consists of
three parts:
There is no direct relationship between standard deviation and quartile deviation, so we
cannot find Q1 and Q3 based on the given information. A/B/C, where:

1c) The geometric mean is the nth root of the product of n numbers. In this case, we have A represents the distribution of inter-arrival times (e.g., M for exponential, D for
four numbers: 3, 9, 27, and 81. deterministic).

Their geometric mean is (34)(93)(272)(811))^(1/10) ≈ 13.262. B represents the distribution of service times (e.g., M for exponential, D for deterministic).

1d) An unbalanced transportation problem occurs when the total supply of goods from C represents the number of service channels (e.g., 1 for a single server, 2 for two servers).
sources is not equal to the total demand for goods at destinations. In other words, there is
either a surplus or a shortage of goods. To balance the problem, we can introduce a dummy For example, the notation M/M/1 represents a queuing model with exponential inter-arrival
source or dummy destination with zero cost, which represents the excess or shortfall in times, exponential service times, and a single server.
supply or demand.
Part 2
1e) The mathematical model for the assignment problem can be written as follows:
2d) Let the other two observations be x and y. Then we have:
Let C be an n x n cost matrix, where C(i,j) represents the cost of assigning the ith job to the
jth worker. The objective is to minimize the total cost of assignment. Mean of 5 observations = (2.6+10+x+y)/5 = 6

Let x(i,j) be a binary variable, where x(i,j) = 1 if the ith job is assigned to the jth worker, and 0 => x + y = 9.4
otherwise.
Also, Variance of 5 observations = [(2-6)^2 + (6-6)^2 + (10-6)^2 + (x-6)^2 + (y-6)^2]/5 = 8
The mathematical model can be written as:
=> (x-6)^2 + (y-6)^2 = 12
Minimize Σ Σ C(i,j) * x(i,j)
Now, we have two equations and two unknowns. Solving them, we get:
subject to:
x = 1.4 and y = 8
Σ x(i,j) = 1 for all i (each job must be assigned to exactly one worker)
Therefore, the other two observations are 1.4 and 8.
Σ x(i,j) = 1 for all j (each worker can be assigned to at most one job)
2e) Let the correct difference in ranks be d. Then we have:
x(i,j) ∈ {0,1} for all i,j.
Given coefficient of rank correlation = 0.5
1f) To determine whether a coefficient of correlation of 0.8 is significant, we need to
calculate the t-statistic and compare it to the critical value at the desired level of => (1 - (6 * sum(d^2))/(n * (n^2 - 1)))^(1/2) = 0.5
significance. The formula for the t-statistic is:
where n = 10 (number of students)
t = r * sqrt(n-2) / sqrt(1-r²)
Solving the above equation, we get:
where r is the sample correlation coefficient and n is the sample size.
sum(d^2) = 7.2
Assuming a significance level of 0.05 and a two-tailed test, the critical value for a sample size
of 5 is 2.571. Now, the incorrect difference in ranks is given as 3, which should have been 7. Therefore, for
each pair of ranks (i, j) where i > j and i and j are the ranks in X and Y respectively, we need
If the absolute value of t is greater than the critical value, then the correlation is significant to subtract 4 from |i - j| to get the correct difference in ranks.
at the 0.05 level.
Using this, we can find the correct sum of squared differences in ranks as follows:
In this case, t = 2.83, which is greater than the critical value. Therefore, we can conclude that
the correlation is significant at the 0.05 level. sum(correct d^2) = sum((|i - j| - 4)^2) = sum(i^2 + j^2 - 8|i - j| + 16) = 2(sum(i^2) + sum(j^2))
- 8(sum|i - j|) + 160
1g) Two properties of regression coefficient are:
where the sums are taken over all pairs (i, j).
The regression coefficient measures the strength and direction of the linear relationship
between two variables. A positive regression coefficient indicates a positive correlation, Using the given coefficient of rank correlation and the correct sum of squared differences in
while a negative regression coefficient indicates a negative correlation. ranks, we can find the correct value of sum|i - j| as follows:

The regression coefficient can be used to predict the value of the dependent variable based (1 - (6 * sum(correct d^2))/(n * (n^2 - 1)))^(1/2) = 0.5
on the value of the independent variable. The slope of the regression line represents the
change in the dependent variable for a one-unit increase in the independent variable. => sum|i - j| = 13.6

1h) The basic components of a queuing system are: Therefore, the correct coefficient of rank correlation is given by:

Arrivals: Customers arrive at the system according to some probability distribution. The (1 - (6 * sum(correct d^2))/(n * (n^2 - 1)))^(1/2) = (1 - (6 * 7.2)/(10 * (10^2 - 1)))^(1/2) = 0.76
arrival rate is the average number of arrivals per unit of time.
2f) (i) Let the total number of customers be T. Then, A has 0.8T loyal customers and B has Outcome 3: Rs. 3 lakhs from Contract 1 and Rs. 5 lakhs from Contract 2 with probability 0.4 *
0.2T loyal customers. Out of the remaining 0.2T customers who were not loyal to A, 60% or 0.6 = 0.24
0.12T switched back to A. Therefore, the number of customers retained by A is given by:
Outcome 4: Rs. 3 lakhs from Contract 1 and Rs. 3 lakhs from Contract 2 with probability 0.4 *
0.8T + 0.12T = 0.92T 0.4 = 0.16

Similarly, the number of customers gained by A is given by: We can now construct the decision tree as shown below:

0.2T - 0.12T = 0.08T / Rs. 10 lakhs (0.6)

The number of customers lost by B is given by: /

0.2T - 0.12T = 0.08T A

(ii) The probability of a customer purchasing A's product at the end of the second period can \
be calculated as follows:
\ Rs. 6 lakhs (0.4)
Let P(A) be the probability of purchasing A's product and P(B) be the probability of
purchasing B's product. Then \

we have: P(A) = P(A|A)P(A) + P(A|B)P(B) \

where P(A|A) is the probability of staying with A given that the customer is currently using |
A's product, P(A|B) is the probability of switching to A given that the customer is currently
using B's product, P(A) is the overall probability of using A's product, and P(B) is the overall |
probability of using B's product.
B----/ \----B
From the given information, we have:
| |
P(A|A) = 1 (since A has 80% loyal customers)
/\ /\
P(A|B) = 0.6 (since B's customers switched back to A's product 60% of the time)
/ \ / \
P(A) = 0.92 (since A retained 92% of its total customers)
| | | |
P(B) = 0.08 (since B lost 92% of its total customers)
Rs. 5L Rs. 3L Rs. 3L Rs. 1L (0)
Substituting these values, we get:
To use EMV (Expected Monetary Value) criteria, we need to calculate the expected value of
P(A) = 1(0.92) + 0.6(0.08) = 0.944 each outcome by multiplying the value of each outcome with its probability and then adding
up the results. The optimal decision is the one with the highest expected value.
Therefore, the probability of a customer purchasing A's product at the end of the second
period is 0.944 or approximately 94.4%. For Course of Action A:

2h) LPP stands for Linear Programming Problem. It is a mathematical optimization technique EMV = (10 lakhs * 0.6) + (6 lakhs * 0.4) = 8.4 lakhs
that is used to determine the best outcome for a given linear objective function, subject to a
set of linear constraints. The objective is to optimize a linear function of a number of For Course of Action B:
variables, subject to constraints that are also linear functions of the same set of variables.
EMV = (5 lakhs * 0.36) + (3 lakhs * 0.24) + (3 lakhs * 0.24) + (1 lakh * 0.16) = 3.6 lakhs
Assumptions of LPP:
Therefore, the optimal decision using EMV criteria is to choose Course of Action A, which
The problem must be formulated in terms of linear equations or inequalities. has an expected monetary value of Rs. 8.4 lakhs.

The objective function and the constraints must be linear and continuous. 2l) Explain the concept of regression analysis and also differentiate between correlation and
regression.
The decision variables must be continuous, meaning that they can take any value within a
certain range. Regression analysis is a statistical method used to examine the relationship between a
dependent variable and one or more independent variables. The main objective of
The problem must have finite and non-empty feasible region, which means there must be at regression analysis is to estimate the values of the dependent variable based on the values
least one solution that satisfies all constraints. of one or more independent variables. Regression analysis is often used to make predictions
or forecasts about future outcomes based on historical data.
The coefficients of the objective function and constraints must be known with certainty.
There are different types of regression analysis, such as simple linear regression, multiple
Advantages of LPP: regression, logistic regression, etc. In simple linear regression, there is only one independent
variable, and the relationship between the independent variable and the dependent
It is a very flexible and widely used optimization technique that can be applied to a wide variable is assumed to be linear. In multiple regression, there are multiple independent
range of problems in different fields. variables, and the relationship between the independent variables and the dependent
variable is also assumed to be linear.
It provides an efficient method for identifying the best possible solution to a problem.
Correlation and regression are often used together, but they are not the same thing.
It allows the user to consider multiple objectives and constraints simultaneously, making it Correlation measures the strength and direction of the linear relationship between two
useful for complex decision-making problems. variables, whereas regression is used to estimate the values of one variable based on the
values of one or more other variables. Correlation does not involve any prediction or
It is capable of handling both deterministic and stochastic problems. estimation, whereas regression is primarily used for prediction or estimation.

LPP provides a quantitative basis for decision-making, which can be very useful in many Another difference between correlation and regression is that correlation coefficients can
applications. range from -1 to 1, whereas the values of the dependent variable estimated by regression
can be any value on the continuous scale. Correlation is a symmetric measure, which means
2j) To construct the decision tree for the given problem, we need to first identify the that the correlation coefficient between two variables is the same regardless of which
different courses of action and the outcomes associated with each course of action: variable is considered the dependent or independent variable. In regression, the dependent
variable is always the variable being estimated, and the independent variable(s) are used to
Course of Action A: Risky contract with two possible outcomes make the estimate.

Outcome 1: Rs. 10 lakhs with probability 0.6 To summarize, correlation measures the strength and direction of the linear relationship
between two variables, whereas regression estimates the values of one variable based on
Outcome 2: Rs. 6 lakhs with probability 0.4 the values of one or more other variables.

Course of Action B: Diversified portfolio with two independent contracts, each with two Part 3
possible outcomes
Q6)(a) Simulation refers to the process of modeling a real-life system or process using a
Outcome 1: Rs. 5 lakhs from Contract 1 and Rs. 5 lakhs from Contract 2 with probability 0.6 * computer program. The simulation process involves developing a mathematical model that
0.6 = 0.36 describes the behavior of the system being simulated, implementing the model in a
computer program, and running the program to generate results that simulate the behavior
Outcome 2: Rs. 5 lakhs from Contract 1 and Rs. 3 lakhs from Contract 2 with probability 0.6 * of the system over time.
0.4 = 0.24
The two major reasons for using simulation to solve a problem are:
Cost-effectiveness: Simulation allows for experimentation with different scenarios without
incurring the costs associated with physical experimentation.

Risk reduction: Simulation enables the evaluation of potential outcomes and their
associated risks in a controlled environment, allowing for better decision-making and risk
management.

(b) Decision making refers to the process of selecting a course of action from among
multiple options. The steps involved in the decision-making process are as follows:

Define the problem: Clearly identify the problem or decision to be made, and gather
relevant information about it.

Generate alternatives: Brainstorm and generate a list of potential options or alternatives to

Evaluate alternatives: Assess the pros and cons of each alternative, and consider the
potential risks and benefits of each.

Select an alternative: Choose the best alternative based on the evaluation criteria and the
goals or objectives of the decision.

Implement the decision: Develop a plan to implement the chosen alternative, and take
action to carry out the plan.

Evaluate the results: Monitor and evaluate the outcomes of the decision, and adjust the plan
as necessary to ensure its effectiveness.