where C and A are coefficient matrices of X and b is the constant vectors in constraint equations. That means
X must satisfy
z CX ¼ 0;
AX ¼ b.
The above equation can be written as follows:
1 C
0 A
z
X
¼ 0
b
.
Assume that B is a feasible basis. Let XB be the corresponding set of basic variables with CB as its associated
objective vector [8].
z CBXB ¼ 0
BXB ¼ b
)
1 CB
0 B
z
XB
¼ 0
b
.
The corresponding solution might be obtained [8]:
z
XB
¼ 1 CB
0 B
1 0
b
¼ 1 CBB1
0 B1
! 0
b
¼ CBB1
b
B1
b
!.
Then, we take the original equation to multiply 1 CBB1
0 B1
in both sides yielding the following equation [8]:
1 CBB1
0 B1
! 1 C
0 A
z
X
¼ 1 CBB1
0 B1
! 0
b
)
1 CBB1
A C
0 B1
A
! z
X
¼ CBB1
b
B1
b
!.
Let Vj be the jth vector of A, that means A = (V1
p
p
p
V2
p
p
p
Vj Vn). The simplex tableau column for variable
xj is shown.
From Table 1, it is easy to know that the inverse B1 is really important. The entire tableau can be generated
once the inverse B1 is known. The optimum solution can be computed in iterations as follows:
z ¼ CBB1
b;
XB ¼ B1
b.
(
A main feature of the generalized simplex me