L16 LP part2 Homework Review N design variables, m equations Summary 1.
LP- a review
-
Upload
fahriamirullah -
Category
Documents
-
view
216 -
download
0
Transcript of LP- a review
-
8/3/2019 LP- a review
1/22
Linear
Programminghttp://endrayanto.staff.ugm.ac.id/courses/MMS2302
-
8/3/2019 LP- a review
2/22
LP: Standard Formc1x1 + c2x2 + + cnxn = z (Min)
a11x1 + a12x2 + + a1nxn = b1
a21x1 + a22x2 +
+ a2nxn = b2...
......
......
am1x1 + am2x2 + + amnxn = bm,
and x1 0, x2 0, . . . , xn 0.
-
8/3/2019 LP- a review
3/22
LP: Matrix NotationMinimize cTxsubject to Ax = b, A : m n,
x 0,
c : column vector (n x 1)
x : column vector (n x 1)
b : column vector (m x 1)
-
8/3/2019 LP- a review
4/22
-
8/3/2019 LP- a review
5/22
-
8/3/2019 LP- a review
6/22
-
8/3/2019 LP- a review
7/22
-
8/3/2019 LP- a review
8/22
-
8/3/2019 LP- a review
9/22
-
8/3/2019 LP- a review
10/22
Definitions
Basic Solutions
Degeneracy Basis
Basic Columns
-
8/3/2019 LP- a review
11/22
Improving a non-optimal BFS
Definition: Reduced Costs (Relatives Cost Factors)
(z) + 2x1 5x3 + x5 = 11+ 2x1 + 3x3 + x4 x5 = 3
x1 + x2 + 2x3 + 2x5 = 4.
z = 11, xB
= (x4, x2) = (3, 4), xN = (x1, x3, x5) = (0, 0, 0).
z = 11 5x3.Ifx3 is increased to any positive value
the value of z would be reduced
-
8/3/2019 LP- a review
12/22
Improving a non-optimal BFS(z) + 2x1 5x3 + x5 = 11
+ 2x1 + 3x3 + x4 x5 = 3 x1 + x2 + 2x3 + 2x5 = 4.
to make x3 as large as possible, z = 11 5x3.
However, x3 cannot be increased indefinitely
x4 = 3
3x3x2 = 4 2x3.
while the other nonbasic variables remain zero
-
8/3/2019 LP- a review
13/22
Improving a non-optimal BFSx4 = 3
3x3
x2 = 4 2x3.if x3 increases beyond 3 3, then x4 becomes negative
x3 increases beyond 4 2 then x2 also becomes negative.
value of x3 is the smaller of these, namely x3 = 1
z = 11 5x3.z = 6, x3 = 1, x2 = 2, x1 = x4 = x5 = 0.
z = 11, xB
= (x4, x2) = (3, 4), xN = (x1, x3, x5) = (0, 0, 0).
new feasible basic solutions
compared to previous solutions
-
8/3/2019 LP- a review
14/22
Improving a non-optimal BFS(z) + 2x1 5x3 + x5 = 11
+ 2x1 + 3x3 + x4 x5 = 3 x1 + x2 + 2x3 + 2x5 = 4.
Choose a pivot term s.t.
basic variables, x2 and x4, could be adjusted without becoming negative,
Pivoting on 3x3,
(z) + 163x1 +
5
3x4
2
3x5 = 6
+2
3x1 + x3 +1
3x4
1
3x5 = 1
7
3x1 + x2
2
3x4 +
8
3x5 = 2.
-
8/3/2019 LP- a review
15/22
Improving a non-optimal BFS(z) + 16
3x1 + 53x4
23x5 = 6
+ 23x1 + x3 +
1
3x4
1
3x5 = 1
7
3x1 + x2
2
3x4 +
8
3x5 = 2.
the basic solution,
z = 6, xB
= (x3, x2) = (1, 2), xN = (x1, x4, x5) = (0, 0, 0),
z = 6 + 23x5
x3 = 1 +1
3x5
x2 = 2 8
3x5. the pivot term,
83x5
-
8/3/2019 LP- a review
16/22
Improving a non-optimal BFS
(z) + 194x1 +
1
4x2 +
3
2x4 =
11
2
+ 38x1 +
1
8x2 + x3 +
1
4x4 =
5
4
7
8x1 +
3
8x2
1
4x4 + x5 =
3
4
he basic feasible solution
z =11
2, x3 =
5
4, x5 =
3
4, x1 = x2 = x4 = 0.
(z) + 16
3x1 + 53x4
23x5 =
6
+ 23x1 + x3 +
1
3x4
1
3x5 = 1
7
3x1 + x2
2
3x4 +
8
3x5 = 2.
-
8/3/2019 LP- a review
17/22
Key component of SM
The optimality test
Introducing Non Basic Variable (NBV) intobasis
-
8/3/2019 LP- a review
18/22
Exercises Exercise 3.2 (Infeasible Problem)
Exercise 3.3 (Unique Minimum)
Exercise 3.4 (Multiple Minima)
Exercise 3.5 (Unbounded Class of Solutions)
-
8/3/2019 LP- a review
19/22
Exercise 3.2It is obvious that the linear program shown below is infeasible.Show algebraically by generating an infeasible inequality thatthis is indeed the case
Minimize x1 + x2subject to x1 + x2 = 2
x1 0
x2 0.
-
8/3/2019 LP- a review
20/22
Exercise 3.3Minimize 7
2x1 = z + 15
subject to x1
+ x2
= 332
x1 + x3 = 4
and x1 0, x2 0, x3 0.
Determine by inspection the basic solution to
Why is it feasible, optimal, and unique?
-
8/3/2019 LP- a review
21/22
Exercise 3.4Prove that the basic solution
is a feasible optimal solution to
Minimize 0x3 = z + 15
subject to x2 2
3x3 =
1
3
x1 +2
3x3 =
8
3
and x1 0, x2 0, x3 0,
z = 15, x1 = 8/3, x2 = 1/3
but that it is not unique.Can you find another optimal basic feasible solution?Are there any nonbasic feasible solutions that are also optimal?
-
8/3/2019 LP- a review
22/22
Exercise 3.5
to feasible canonical form and generate a class of solutions
that in the limit cause the objective function to go to -infinity.
Minimize x1 x2 = zsubject to x1 x2 = 1
x1 0x2 0
Reduce