Linear Programming
Big M Method
Mohamed Haridy
[email protected]
Acknowledgment & Credit
Shokoufeh Mirzaei (GREAT Plasy list)
Simplex Method
Initial Tableu_Iteration 0
Raw Number Basic Variable X1 X2 y1 y2 Z RHS Ratio Test
R1 y1 1 1 1 0 0 12 (12/1)=12
R2 y2 2 1 0 1 0 16 (16/2)=8
Z Z -40 -30 0 0 1 0
All inequalities must be
RHS values must be positive
Non negativity condition
Coefficients of all slack variables are +1
• Used when there are constraints with equal or signs in a LPP
Subject to:
One Feasible solution (satisfies all constraints) is 00
The problem is that can not be a basic variable as needed:
• It has a negative coefficient
• Even if the equation is multiplied with -1, the RHS will be –ve which would result in an infeasible solution at the
initial table (when =0) since e= -150 and negative DV is an infeasible solution.
• The solution to this situation is the artificial variable, which will be the BV
Artificial Variables
• Are added to the constraints with
• Add the A.V. to the LHS of the constraint with a coefficient of +1
• The A.V. doesn't exist and it is an auxiliary which we need to get rid of ,through
the solution
• In the O.F. add the A.V. with very big –ve coefficient (in case of max. problems)
• -Ma4 , where M is a very big number and the goal of the solution is to get rid of the A.V. to
eliminate its negative effect of Z. Goal is to set a4 to zero if possible
The Standard For is :
Subject to:
Initial Table
Basic Variables x1 x2 s1 s2 s3 a4 e4 RHS
s1 2 1 1 0 0 0 0 600
s2 1 1 0 1 0 0 0 225
s3 5 4 0 0 1 0 0 1000
a4 1 2 0 0 0 1 -1 150
z -3 -4 0 0 0 M 0 0
Must be zero for a4 to be B.V
so, we need to use the EROs
znew = -MR4+z
Initial Table
Basic Variables x1 x2 s1 s2 s3 a4 e4 RHS Min
test
s1 2 1 1 0 0 0 0 600 600
s2 1 1 0 1 0 0 0 225 225
s3 5 4 0 0 1 0 0 1000 250
a4 1 2 0 0 0 1 -1 150 75
z -3-M -4-2M 0 0 0 0 M 0 -150M
P.C.
Pivot element
Iteration 2
B.V. x1 x2 s1 s2 s3 a4 e4 RHS
-R4n+R1o s1 3/2 0 1 0 0 -1/2 1/2 525 1150
-R4n+R2o s2 ½ 0 0 1 0 -1/2 1/2 150 300
-2R4o+R3o s3 3 0 0 0 1 -2 2 700 350
R4/2 x2 ½ 1 0 0 0 1/2 -1/2 75 __
(2M+4)R4n+zo z -1 0 0 0 0 M+2 -2 300
P.C.
-ve values , still need to be optimized
Pivot element
Iteration 2
B.V. x1 x2 s1 s2 s3 a4 e4 RHS
-R2o+R1o s1 1 0 1 -1 0 0 0 375
2R2o e4 1 0 0 2 0 -1 1 300
-R2n+R3o s3 1 0 0 -4 1 0 0 100
R2o+R4o x2 1 1 0 1 0 0 0 225
2R2n+zo z 1 0 0 4 0 M 0 900
Optimality Check: All Positive