15
Big M Method A Variant of Simplex Method Presented By : Luckshay Batra [email protected]

Big m method

Embed Size (px)

Citation preview

Page 1: Big m method

Big M MethodA Variant of

Simplex MethodPresented By : Luckshay Batra

[email protected]

Page 2: Big m method

Content

• Introduction• Algorithm• Points to Remember• Example• Analysis of Big M Method• Drawbacks• Conclusion• References

Page 3: Big m method

Introduction

A method of solving linear programming problems.

It is one of the oldest LP techniques.

Big M refers to a large number associated with the artificial variables.

The Big M method introduces surplus and artificial variables to convert all inequalities into standard form.

Page 4: Big m method

Algorithm

Add artificial variables in the model to obtain a feasible solution.

Added only to the ‘>’ type or the ‘=‘ constraints.

A value M is assigned to each artificial variable.

The transformed problem is then solved using simplex eliminating the artificial variables.

Page 5: Big m method

Points To Remember

Solve the modified LPP by simplex method, untilany one of the three cases may arise:-

If no artificial variable appears in the basis and the optimality conditions are satisfied.

If at least one artificial variable in the basis at zero level and the optimality condition is satisfied .

If at least one artificial variable appears in the basis at positive level and the optimality condition is satisfied, then the original problem has no feasible solution.

Page 6: Big m method

Example Maximize Z = x1 + 5x2

Subject to 4x1 + 4x2 ≤ 6

x1 + 3x2 ≥ 2

x1 , x2 ≥ 0

Solution : Introducing slack & surplus variables :

4x1 + 4x2 + S1 = 6

x1 + 3x2 - S2 = 2

whereS1 is a slack variable S2 is a surplus variable

The surplus variable S2 represents the extra units.

Page 7: Big m method

Now if we let x1 & x2 equal to zero in the initial solution , we will have S1=6 , S2=-2 , which is not possible because a surplus variable cannot be negative . Therefore , we need artificial variables.

Introducing an artificial variable , say A1.

Standard Form :

Maximize Z = x1 + 5x2+ 0s1 + 0s2 – M(A1)

Subject to 4x1 + 4x2 +S1 = 6

x1 + 3x2 –S2 +A1 = 2

x1 , x2 , S1 , S2 , A1≥0

Page 8: Big m method
Page 9: Big m method
Page 10: Big m method

Analysis of Big M Method Problem P : Minimize cx

Subject to Ax = b x≥ 0

Problem P(M) : Minimize cx + M s

Subject to Ax + s = b x , s ≥ 0 where, “s” is an artificial variable

Page 11: Big m method

Analysis of Big M Method

Solve P(M) for a very large positive M

Optimal is finite

s=0. Optimal solution of P

is found

s≠0. P has no feasible solutions

Optimal is unbounded

s=0. Optimal solution of P

is unbounded

s≠0. P is infeasible

Page 12: Big m method

Drawbacks How large should M be?

If M is too large, serious numerical difficulties in a computer.

Big-M method is inferior than 2 phase method.

Here feasibility is not known until optimality.

Never used in commercial codes.

Page 13: Big m method

Conclusion

The application of the M technique requires that M approaches infinity but to computerize the solution algorithm , M must be finite while being “sufficiently large.”

The pitfall in this case is, however, if M is too large it can lead to substantial round-off error yielding an incorrect optimal solution . For this reason, most commercial LP solvers do not apply the M-method but use, rather, an artificial variable method called the two-phase method.

Page 14: Big m method

References http://cbom.atozmath.com/CBOM/Simplex.aspx

http://businessmanagementcourses.org/Lesson09TheBigMMethod.pdf

http://www.slideshare.net/NiteshSinghPatel/big-m-32360766

http://en.wikipedia.org/wiki/Big_M_method

Linear Programming & Network Flows by Mokhtar S. Bazaraa , Hanif D. Sherali , John J. Jarvis

Operation Research An Introduction by H. A. Taha

Page 15: Big m method