NOTES FOR OPERATIONS RESEARCH

WHAT IS OPERATIONS RESEARCH?

Operations Research is the representation of real-world systems by mathematical models together with the
use of quantitative methods (algorithms) for solving such models, with a view to optimizing an Objective.
Operations Research is a scientific approach to decision making that seeks to best design and operate a
system, usually under conditions requiring the allocation of scarce resources.
It is a collection of methods and algorithms to solve problems that are formulated mathematically.
A mathematical model as consisting of:
➢ Decision variables, which are the unknowns to be determined by the solution to the model.
➢ Constraints to represent the physical limitations of the system
➢ An objective function
➢ An optimal solution to the model is the identification of a set of variable values which are feasible
(satisfy all the constraints) and which lead to the optimal value of the objective function.
An optimization model seeks to find values of the decision variables that optimize (maximize or minimize)
an objective function among the set of all values for the decision variables that satisfy the given constraints.
The steps involved in applying OR methods involve the following:
Formulate the Problem
Observe the System
Formulate a Mathematical Model of the Problem
Verify the Model and Use the Model for Prediction
Select a Suitable Alternative
Present the Results and Conclusions of the Study
Implement and Evaluate Recommendation

DISCUSS SOME OF THE AREAS OF APPLICATION OF OR

1. Production and operations

• Product mix decisions to maximize profits or minimize costs
• Forecasting
• Inventory management
• Scheduling and sequencing of production runs
• Transportation and distribution planning
• Allocation of resources (men, matls, machines, money, space etc)
• Factory size and location decision.
2. Project Management
• Project scheduling, monitoring and control
• Optimal allocation of resources
• Deployment of proper workforce
3. Marketing:
• Advertising budget allocation.
• Product introduction timing.
• Selection of advertising media.
• Selection of product mix.
4. Organization Behaviour:
• Selection of personnel, determination of retirement age and skills.
• Recruitment policies and assignment of jobs.
• Recruitment of employees.
• Scheduling of training programs.
5. Finance:
• Capital requirements, cash flow analysis.
• Credit policies, credit risks etc.
• Investment decision.
6. Research and Development:
• Control of R&D projects.
• Determination of areas for research and development.
• Selection of projects and preparation of their budgets.
LINEAR PROGRAMMING
Linear programming is a special class of mathematical programming models in which the objective function
and the constraints can be expressed as linear functions of the decision variables.
An objective function represents some principal objective criterion or goal that measures the effectiveness of
the system (such as maximizing profits or productivity or minimizing cost or consumption). There is always
some practical limitation on the availability of resources (time, materials, machines, energy, or manpower)
for the system, and such constraints are expressed as linear inequalities or equations involving the decision
variables.
Thus Linear Programming problem can be considered as a mathematical model formulation of a real life
problem when the following conditions are met:
The formulation consists of:
A single Objective function to be maximized or minimized
A set of constraints in the forms of equalities or inequalities
Constraints on the signs that variables in the formulation are allowed to take.
All variables should be of continuous type (i.e. can take fractional values)
The objective function and constraints are linear i.e. any term is either a constant or a constant
multiplied by an unknown.
Some of the major areas to which LP can be applied are:
• Work scheduling
• Production planning & Production process
• Capital budgeting
• Financial planning
• Blending (e.g. Oil refinery management)
• Farm planning
• Distribution
• Multi-period decision problems
• Inventory model
• Financial models
• Work scheduling

THE SIMPLEX ALGORITHM

The Simplex method is an iterative algorithm that begins with an initial feasible solution, repeatedly
moves to a better solution, and stops when an optimal solution has been found and, therefore, no
improvement can be made.
Steps
1. Convert the LP to standard form
2. Obtain a basic feasible solution from the standard form
3. Determine whether the current basic feasible solution is optimal. If it is optimal, stop.
4. If the current basic feasible solution is not optimal, determine which non basic variable should
become a basic variable and which basic variable should become a non basic variable to find a
new basic feasible solution with a better objective function value
5. Go back to Step 3.
Converting a Linear Programming to the standard form:
• A minimization problem can be converted to a maximization problem by multiplying the objective
function by −1. For example, the problem of minimizing z = 3x1 − 5x2 is equivalent to maximizing
z = −3x1 + 5x2

• Negative right hand sides of the constraints can be made positive by multiplying the constraint by
−1 (reversing the sign of the inequality).
• Equality constraints require no modification.
• Inequality constraints can be converted to equalities through the introduction of additional variables
that make up the difference in the left and right sides of the inequalities.
• Less than or equal to (≤) inequalities require the introduction of variables called slack variables. For
example, a constraint such as 3x1 + 4x2 ≤ 7 becomes the equality 3x1 + 4x2 + s1 = 7 when we
introduce the slack variable s1, where s1 ≥ 0.
• Greater than or equal to (≥) constraints are modified by introducing surplus variables. For example,
the constraint 14x1 + 3x2 ≥ 12 becomes the equality 14x1 + 3x2 − s2 = 12, where s2 is the non-
negative surplus variable.
• Finally, all variables are required to be non-negative in the standard form.
 Obtain a basic feasible solution from the standard form
• A basic solution to the system of m equations and n unknowns is obtained by setting (n − m) of the
variables to zero, and solving for the remaining m variables. The m variables that are not set equal
to zero are called basic variables, and the (n − m) variables set to zero are non-basic variables.
Determine whether the current basic feasible solution is optimal.
• Corresponding to each variable in the Objective function, we calculate Cj-Zj values.
• If all the Cj-Zj values are <=0 in case of a maximization problem, the solution is optimal.
• If all the Cj-Zj values are >= 0 in case of a minimization problem, the solution is optimal.
Iteration to improve the solution when not optimal
• This involves identifying a variable that should be replaced in the basis matrix.
• In a maxn problem, the variable that has the largest +ve value of Cj-Zj is called the entering
variable. This column are called the pivot column. The RHS of the constraint equations are divided
by the corresponding elements in the pivot column to find the ratios, and the row with the minimum
+ve ratio is the pivot row. The variable in the pivot row is the leaving variable.
• The simplex matrix is revised with the leaving variable removed and replaced with the entering
variable.
• The new Cj- Zj values are calculated, and optimality checked.
• The process is repeated until optimality is attained.

The Big M Method

The Big-M method is an approach to solving Linear Programming problems when the Objective function is
a Minimization function.
In this case, usually the constraints will be of the >= type.
If an LP has any ≥ or = constraints, a surplus variable is added to the LHS of the inequality to convert into
an equality. In this case an initial feasible solution is not possible as the variables will take -ve values
violating the non-negativity constraints.
For this purpose, an artificial variable is added to the LHS.
For ex: a constraint of the form 𝑥1 + 𝑥2 >= 5 will be converted into equality as
𝑥1 + 𝑥2 – 𝑥3 + A1 = 5, where 𝑥3 is the surplus variable and A1 is an artificial variable.
The contribution of the artificial function in the Objective function, ie the Objective function coefficient of
the artificial variable, is entered as M, which is a vary large value.
For example, an Objective function may look like
Minimize Z = 3𝑥1 + 2𝑥2 + 0𝑥3 + MA1
As the problem is one of minimization, the variable with a large value will be eliminated during the iteration
process and hence the solution obtained will contain only the other decision and slack or surplus variables.

DUALITY
Associated with any LP is another LP called the dual.. When taking the dual of any LP, the given LP is
referred to as the primal. If the primal is a max problem, the dual will be a min problem and vice versa.
In general, the dual is constructed following the rules as below:
The dual problem is a minimization problem.
• For every variable xi in the primal problem, there is a constraint in the dual problem. If xi ≥ 0 in the
primal, the constraint is a ≥ inequality in the dual.
• If xi is unrestricted in sign, the i-th constraint is an equality in the dual.
• For every constraint in the primal problem, there is a variable yi in the dual.
• If the constraint is ≤, then yi ≥ 0 in the dual problem.
• If the constraint is an equality, then yi is unrestricted in sign in the dual.
• The right hand sides in the primal are the objective function coefficients in the dual.
• The objective function coefficients in the primal are the right hand sides in the dual.
• The coefficient matrix in the primal is transposed to form the coefficient matrix for the dual

Economic Interpretation of the dual variable values

When the primal is a normal max problem, the dual variables are related to the value of resources available
to the decision maker. For this reason, dual variables are often referred to as resource shadow prices.
SENSITIVITY ANALYSIS
Sensitivity analysis in Linear Programming is the study of the types, ranges, and magnitude of changes in
problem parameters whose effects can be determined relatively easily, without the need for solving a new
linear programming problem.
Sensitivity analysis addresses several different kinds of changes to a linear programming formulation,
including:
• Changes in objective function coefficients
• Increases or decreases in the right-hand side of a constraint
• Adding a new variable
• Adding a constraint
• Changes in constraint coefficients

TRANSPORTATION PROBLEMS
In general, a transportation problem is specified by the following information:
• A set of m supply points from which a good/service is shipped. Supply point i can supply at most si
units.
• A set of n demand points to which the good/service is shipped. Demand point j must receive at least
dj units.
• Each unit produced at supply point i and shipped to demand point j incurs a variable cost of cij.
The Objective is to minimize the cost of transportation so as to meet the demands and supply constraints.

There are three methods that can be used to find in initial feasible solution for a balanced transportation
problem:
1. Northwest Corner method
2. Minimum cost method
3. Vogel’s method

Northwest Corner Method

We begin in the upper left corner of the transportation tableau and set x11 as large as possible (clearly, x11
can be no larger than the smaller of s1 and d1).
• If x11=s1, cross out the first row of the tableau. Also change d1 to d1-s1.
• If x11=d1, cross out the first column of the tableau. Change s1 to s1-d1.
• If x11=s1=d1, cross out either row 1 or column 1 (but not both!).
o If you cross out row, change d1 to 0.
o If you cross out column, change s1 to 0.
Continue applying this procedure to the most northwest cell in the tableau that does not lie in a crossed out
row or column.
Eventually, we will come to a point where there is only one cell that can be assigned a value. Assign this
cell a value equal to its row or column demand, and cross out both the cell’s row or column.
initial feasible solution is now obtained.

Minimum Cost Method

To begin the minimum cost method, find the variable with the smallest shipping cost (call it xij). Then assign
xij its largest possible value, min {si, dj}.
As in the NWC method, cross out row i or column j and reduce the supply or demand of the noncrossed-out
of row or column by the value of xij.
Continue like NWC method (instead of assigning upper left corner, the cell with the minimum cost is
assigned). See Northwest Corner Method for the details!

Vogel’s Method
Begin by computing for each row and column a penalty equal to the difference between the two smallest
costs in the row and column.
Next find the row or column with the largest penalty.
In the lowest cost cell in this row or column, allocate the maximum possible quantity,
cross out row or column, and change the supply or demand figures in the row or column as the case may be.
Now recomputed new penalties (using only cells that do not lie in a crossed out row or column)
Repeat the procedure until only one uncrossed cell remains.
Set this variable equal to the supply or demand associated with the variable, and cross out the variable’s row
and column.
PROJECT MANAGEMENT
The planning and coordination of large complex projects, consisting of many tasks or activities. These
activities are interconnected and interdependent defined through Precedence and Successor relationships.
A project network provides a graphical representation of the precedence relations among all the activities in
a project. Each activity is represented by an arc in the network. The nodes in the network denote events
corresponding to points in time when one or more activities are completed. Directions on the arcs indicate
the sequence in which events must occur. Additionally, a node is added at the beginning of the network to
represent the start event for the entire project. Similarly, a final node is introduced to denote the finish event
for the project.
Project activity networks are used to plan and coordinate large complex projects consisting of many tasks.
Critical paths in networks determine the minimum project completion time, and identify those tasks or
activities whose timely completion is critical to achieving this minimum project duration.

PERT and CPM

Two methodologies for project management are the Critical Path Method (CPM), and Program Evaluation
and Review Technique (PERT).
CPM is usually applied to the management of construction and production activities. CPM is based on
deterministic specifications of task durations, and is therefore appropriate for production projects in which
previous experience with the subtasks allows management to make reliable time estimates.
PERT, on the other hand, is based on probabilistic estimates of task durations, and thus is most useful in a
research and development environment where task completion times cannot be known in advance.
In PERT, the task duration is estimated based on estimates of an
a optimistic duration (a),
a pessimistic estimate (b)
a most likely duration (m).
The expected time duration of the activity is calculated as (a+4m+b)/6.
Because both PERT and CPM approach project scheduling using similar network models and methods, the
terms PERT and CPM are sometimes used interchangeably or collectively as PERT-CPM methods.

Crashing a Project Network

The CPM helps in scheduling activities in order to achieve a minimum project duration, and no
consideration has been given to the cost of the project.
In addition to direct costs associated with each individual activity, there are typically indirect costs that may
be viewed as overhead costs and that are proportional to the duration of the entire project. These costs may
include such expenses as administrative or supervisory costs, equipment and facilities rental, and interest on
capital. It may be possible to add resources, involving some direct expense, to certain activities in order to
reduce the duration of those activities, and thereby to reduce the project duration and the attendant indirect
costs.
Assuming that the normal duration and the budgeted cost associated with completing the activity under
normal circumstances is known, and also, through additional expenditures, the duration of each activity can
be reduced. This is known as crashing.

When the cost of crashing an activity in the critical path by a day is less than the indirect cost per day, then
crashing will reduce both the costs and duration of the project.

QUEING
Waiting lines represent people waiting for service, machines waiting for a repairman, parts waiting to be
assembled, and so on
Queues or Waiting lines inherently create inconvenience, inefficiency, delay, or other problems. and these
situations cost time and money.
Waiting lines can be reduced or eliminated by simply adding lots of servers, but this can be very expensive.
Therefore, the management of the queuing problem involves balancing these costs and optimizing.
Queueing models provide a set of tools by which we can analyze the behavior of systems involving waiting
lines, or queues. Queueing systems are characterized by the distribution of customers entering the system
and the distribution of times required to service the customers.

THE KENDAL NOTATION OF QUEING SYSTEM

A queue system is described using a set of 6 symbols in the Kendall Notation.
A / B / C / D / E / F where:
A and B are letters that denote the interarrival time distribution and the service time distribution,
respectively.
C is a number that denotes the number of parallel servers or channels.
 D and E denote the system capacity and size of the calling source, respectively.
F is an abbreviation identifying the queue discipline.
For example, a queueing system described as M / M / 1 / ∞/ ∞ / FCFS means:
Arrival process follows a Markovian or Poisson Distribution ( first M)
Service process follows a Poisson Distribution or Exponential time distribution ( 2nd M)
The number 1 denotes the number of severs.
∞ denotes that the system capacity has no limitations
∞ denotes that the source population is infinite
FCFS means the service discipline is First Come First Served.

SIMULATION
Simulation is the process of studying the behavior of an existing or proposed system by observing the
behavior of a model representing the system.
Simulation is the imitation of a real system or process operating over a period of time.
By simulating a system, we may be able to make observations of the performance of an existing system,
hypothesize modifications to an existing system, or even determine the operating characteristics of a
non-existent system.
Through simulation, it is possible to experiment with the operation of a system in ways that would be too
costly or dangerous or otherwise infeasible to perform on the actual system itself.
Many systems are so complex that mathematical methods are inadequate to model the intricate
interaction among system elements. In these cases, simulation techniques may provide a framework for
observing, predicting, modifying, and even optimizing a system.

MONTE CARLO SIMULATION

Monte Carlo Simulation is a mathematical technique that generates random variables for modelling risk or
uncertainty of a certain system.
The random variables or inputs are modelled on the basis of probability distributions such as normal, log
normal, etc. Different iterations or simulations are run for generating paths and the outcome is arrived at by
using suitable numerical computations.
Monte Carlo Simulation is most often used when a model has uncertain parameters or a dynamic complex
system needs to be analysed. It is a probabilistic method for modelling risk in a system.
The method is used extensively in a wide variety of fields such as physical science, computational biology,
statistics, artificial intelligence, and quantitative finance.
Monte Carlo Simulation uses probability distribution for modelling a stochastic or a random variable.
Different probability distributions are used for modelling input variables such as normal, lognormal,
uniform, and triangular. From probability distribution of input variable, different paths of outcome are
generated.
Compared to deterministic analysis, the Monte Carlo method provides a superior simulation of risk. It gives
an idea of not only what outcome to expect but also the probability of occurrence of that outcome. It is also
possible to model correlated input variables.

GAME THEORY
Game theory addresses possible approaches to decision-making under the assumption of complete
ignorance.
It is described in terms of players, payoffs, and strategies.
For example, in a two-person game: the decision-maker (player one) selects an alternative and then player
two selects a state. The payoff is given by the corresponding entry in a payoff matrix.
Player two is assumed to be indifferent to the choices of player one
Player one will make a selection based on some strategy intended to make the most of the opportunity.
Game theory provides a framework for classifying and analyzing general types of human behavior in the
presence of uncertainty.

A variety of strategies can be employed in making the selection of alternatives.

The selection depends on the decision-maker’s attitude toward risk. Because each choice has a different
degree of risk associated with it, different people will make different selections.

Maximin Strategy
For each alternative, aj , pick the worst possible outcome (the minimum). Choose the alternative that
has the maximum value of this minimum.
 The Maximin strategy is pessimist approach. The person who believes that, whatever they do will
always turn out badly, and that nature is working directly against them, and is characteristic of the
conservative decision-maker. The given decision has the lowest risk.

Maximax Strategy
For each alternative, pick the best possible outcome (the maximum).
Choose the alternative which has the maximum value of this maximum.
The Maximax player represents the optimist approach. Such people believe that anything they do will
turn out right, and are willing to take risks for a chance at the greatest possible return.

Laplace Principle (Principle of Insufficient Reason)

Assume that every state is equally likely and calculate the expected payoff for each alternative.
The alternative with the highest expected payoff is selected.
The expected payoff for a given alternative is computed by taking each element in the corresponding
column of the payoff matrix, and multiplying each payoff by the corresponding state probability. The
expected payoff is the sum of these values.
This approach is often referred to as the approach of a Rational Decision Maker.

Hurwicz Principle
Define 0 ≤ α ≤ 1 to be the Decision-maker’s Degree of Optimism between the two extremes of
Maximin (α = 0) and Maximax (α = 1).
For each alternative, the Hurwicz measure: α*maximum + (1 – α)* minimum
Select the alternative with the highest value.
The Hurwicz principle is based on the assumption that the decision-maker is neither totally
pessimistic (as with the Maximin strategy), nor totally optimistic (as with Maximax).
Each individual decision-maker can select his own degree of optimism some where between these
two extremes.

Savage Minimax Regret

For each choice, a regret row is (the highest pay off – the payoff in the other cells)
The regret matrix is constructed for all the rows.
For each alternative, find the maximum regret.
Select the alternative that minimizes this maximum regret.
This strategy is associated with insecure decision-makers. Such people are not primarily interested in
making the highest profit; they are more concerned with how disappointed they are going to feel after
the fact.