Scribd
Upload
10 views

Big M

How to deal with customers for higher retention

Uploaded by

Rabbani Momin
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
10 views

Big M

How to deal with customers for higher retention

Uploaded by

Rabbani Momin
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

The Big M Method is a technique used in linear programming to handle constraints that are

either not in the standard form (equalities) or have surplus or artificial variables. This method
is particularly useful for solving problems involving inequalities and ensuring a feasible
solution is found before proceeding to optimize the objective function. Here's a detailed
theoretical explanation:

Theory of the Big M Method

1. Introduction to Linear Programming and Standard Form

In linear programming, we typically want to maximize or minimize an objective function


subject to a set of constraints. The standard form of a linear programming problem requires
that all constraints be equalities and all variables be non-negative. The Big M Method
transforms problems with inequalities or specific constraints into this standard form.

2. Conversion to Standard Form

To apply the Big M Method, follow these steps:

1. Convert Inequality Constraints:


o Less-than-or-equal-to (≤\leq≤) constraints: Introduce a slack variable. For a
constraint a1x1+a2x2≤ba_1x_1 + a_2x_2 \leq ba1x1+a2x2≤b, introduce a
slack variable sss to convert it to an equality: a1x1+a2x2+s=ba_1x_1 +
a_2x_2 + s = ba1x1+a2x2+s=b, where s≥0s \geq 0s≥0.
o Greater-than-or-equal-to (≥\geq≥) constraints: Introduce a surplus variable
and an artificial variable. For a constraint a1x1+a2x2≥ba_1x_1 + a_2x_2 \geq
ba1x1+a2x2≥b, convert it to an equality: a1x1+a2x2−s+a=ba_1x_1 + a_2x_2 -
s + a = ba1x1+a2x2−s+a=b, where s≥0s \geq 0s≥0 (surplus) and a≥0a \geq
0a≥0 (artificial).
2. Objective Function Adjustment:
o In the standard form, the objective function must be in terms of maximization
or minimization without artificial variables. If the problem involves artificial
variables, modify the objective function to include a penalty term for these
artificial variables to ensure they are driven out of the solution.

3. Incorporating the Big M

The Big M Method introduces a large penalty (M) for artificial variables to ensure they do
not remain in the final solution. The key steps include:

1. Modify the Objective Function:


o If artificial variables are introduced, include a large penalty term in the
objective function. For a maximization problem, this is represented as −M⋅ai-
M \cdot a_i−M⋅ai for each artificial variable aia_iai. For a minimization
problem, it is +M⋅ai+M \cdot a_i+M⋅ai.
2. Formulate the Initial Simplex Tableau:
o Create an initial tableau with the modified objective function that includes
these penalties. This ensures that in the simplex method iterations, the
artificial variables will be driven out of the solution as their coefficients are
large and negative (or positive for minimization).
3. Perform the Simplex Method:
o Apply the simplex algorithm as usual. During the iterations, ensure that
artificial variables are eliminated from the basis. The solution will be feasible
if the artificial variables are zero in the final tableau.

4. Solution Interpretation

After applying the Big M Method and solving the problem using the simplex algorithm:

1. Check the Final Tableau:


o Ensure that all artificial variables have been driven to zero in the final
solution. If any artificial variable has a non-zero value, it indicates that the
problem is infeasible.
2. Interpret the Optimal Solution:
o The remaining variables and their values, along with the optimal value of the
objective function, provide the solution to the original problem.
The Two-Phase Method is an approach used in linear programming to solve problems where
the initial feasible solution is not obvious, especially when dealing with constraints that are
not in standard form or involve artificial variables. This method is used to find a feasible
starting point for the simplex algorithm and to ensure that the solution satisfies all constraints.
Here's a detailed explanation of the Two-Phase Method:

Theory of the Two-Phase Method

1. Introduction

The Two-Phase Method is employed when:


• The linear programming problem includes constraints that are not in equality form or
have artificial variables.
• The initial feasible solution is not readily apparent.

The method involves two main phases:

1. Phase 1: Finding a Feasible Solution


2. Phase 2: Optimizing the Objective Function

2. Phase 1: Finding a Feasible Solution

Objective: Determine if there is a feasible solution to the constraints and, if so, find an initial
basic feasible solution for the simplex method.

Steps:

1. Convert Constraints:
o Convert all constraints to equality constraints if they are not already in this
form.
o For constraints of the form a1x1+a2x2≥ba_1x_1 + a_2x_2 \geq ba1x1+a2x2
≥b, introduce a surplus variable and an artificial variable.
o For constraints of the form a1x1+a2x2≤ba_1x_1 + a_2x_2 \leq ba1x1+a2x2
≤b, introduce a slack variable.
o For equality constraints a1x1+a2x2=ba_1x_1 + a_2x_2 = ba1x1+a2x2=b,
introduce an artificial variable.
2. Formulate Phase 1 Objective Function:
o Create a new objective function (Phase 1 objective function) to minimize the
sum of artificial variables.
o This is expressed as: Minimize W=∑iai\text{Minimize } W = \sum_{i}
a_iMinimize W=∑iai
o where aia_iai are the artificial variables.
3. Construct Phase 1 Simplex Tableau:
o Set up the simplex tableau for Phase 1 with the Phase 1 objective function as
described above.
4. Solve Phase 1 Problem:
o Use the simplex method to solve for Phase 1.
o If the objective value WWW in Phase 1 is zero, then a feasible solution to the
original problem exists. If WWW is positive, the problem is infeasible.
5. Check Feasibility:
o If W=0W = 0W=0, proceed to Phase 2. Otherwise, the problem has no feasible
solution.

3. Phase 2: Optimizing the Objective Function

Objective: Optimize the original objective function using the feasible solution found in
Phase 1.

Steps:

1. Prepare Phase 2 Simplex Tableau:


o Replace the Phase 1 objective function with the original objective function in
the tableau.
o The feasible solution from Phase 1 is used as the starting point for Phase 2.
2. Adjust the Tableau:
o Update the tableau to reflect the original objective function and ensure that the
feasible solution from Phase 1 is properly incorporated.
3. Solve Phase 2 Problem:
o Apply the simplex method to optimize the original objective function using
the feasible solution obtained from Phase 1.
o Continue iterations until no negative coefficients remain in the objective
function row for maximization (or no positive coefficients for minimization).
4. Interpret the Final Solution:
o The final tableau will provide the values of the decision variables and the
optimal value of the original objective function.

You might also like