1. Run the Dijkstra method to find the shortest paths to every node from A! What will be the shortest path to B?



A

B

C

D

E

F

G

0

-

-

-

-

-

-

|

7

-

-

1

-

-

|

A



A



|

7

-

5

|

4

2

|

A


E

|

E

E

|

7

-

5

|

3

|

|

A


E

|

G

|

|

7

4

5

|

|

|

|

A

F

E

|

|

|

|

5

|

5

|

|

|

|

C

|

E

|

|

|

|

|

|

5

|

|

|

|

|

|

E

|

|

|



The shortest path to B in backward direction: B – C – F – G – E – A.























2. Run the modified Dijkstra method to find the widest path among the shortest ones to each node from A (the black numbers are the distances, and the red ones are the widths). What is the best path to D?




A

B

C

D

E

F

G

0,-

-

-

-

-

-

-

|

4,5

-

-

1,4

-

-

|

A



A



|

4,5

-

5,1

|

3,2

2,4

|

A


E

|

E

E

|

4,5

-

5,1

|

3,3

|

|

A


E

|

G

|

|

4,5

4,2

5,1

|

|

|

|

A

F

E

|

|

|

|

|

4,2

5,5

|

|

|

|

|

F

B

|

|

|

|

|

|

5,5

|

|

|

|

|

|

B

|

|

|



The shortest path to D in backward direction: D – B – A.






















3. Solve the following linear programming problems with graphic tools.

a) Max{x-y: y >= 0; 2x-y >= 0; x+y<=9}

b) Max{-x+y: y >= 0; 2x-y >= 0; x+y<=9}


The optimal solution for a) will be on the left node of the triangle: x=9,y=0, with objective function value 9.

The optimal solution for b) will be on the top node of the triangle: x=3,y=6, with objective function value 3.




















4. What is the dual of the following linear programming problems:

a) The problems of the previous example

b) Min{2x+3y: x+y+z >= 2; 2x-y >= 3; x+y-z<=4}

c) Min{3x+2y: 2x+y = 2; 3x-2y = 3; x-2y=4 x>=0; y>=0}


a) a) Max{x-y: y >= 0; 2x-y >= 0; x+y<=9}

Max{x-y: -y <=0; -2x+y <= 0; x+y<=9}


The dual will be:

Min{9x3: -2x2 + x3 = 1; -x1+x2+x3=-1; x1>=0; x2>=0; x3>=0}


b) The same way as in the previous part of the example:


The dual will be:

Min{9x3: -2x2 + x3 = -1; -x1+x2+x3=1; x1>=0; x2>=0; x3>=0}



b) Min{2x+3y: x+y+z >= 2; 2x-y >= 3; x+y-z<=4} <=> -Max{-2x-3y: x+y+z >= 2; 2x-y >= 3; x+y-z<=4} <=> -Max{-2x-3y: -x-y-z <= -2; -2x+y <= -3; x+y-z<=4}

The dual will be -Min{-2x-3y+4z: -x-2y+z=-2; -x+y+z=-3; -x-z=0; x>=0; y>=0; z>=0 }

That is Max{2x+3y-4z: -x-2y+z=-2; -x+y+z=-3; -x-z=0; x>=0; y>=0; z>=0 }



c) Observe that it is in the form of the dual LP in the theorem, so its dual will be in the form of the original primal one:so the dual of Min{3x+2y: 2x+y = 2; 3x-2y = 3; x-2y=4 x>=0; y>=0}


Is Max{2x+3y+4z: 2x+3y+z<=3; x-2y-2z<=2}.




















5. Write down the linear/integer programming formulation of the following problems:

a) The knypsack problem with limit 103, objects 21-46, 38-12, 53-26, 49-38, 25-25

max{46x1+12x2+26x3+38x4+25x5}

With the conditions:

For all i xi>=0, xi<=1 and xi integer

21x1+38x2+53x3+49x4+25x5 <=103












b) The network flow problem of this graph:


We label the edges with the red numbers:


max{x7+x8+x9}

With the conditions:

For all i xi>=0,

x1-x7=0

x2+x4+x5-x8=0

x3-x4-x6=0

-x5+x6-x9=0

x1<=3

x2<=3

x3<=3

x4<=4

x5<=4

x6<=4

x7<=5

x8<=5

x9<=5












c) The minimum assignment (this means a perfect matching with minimum total value) problem of this graph:


We label the edges from left to right.

min{5x1+4x2+3x3+7x4+x5+5x6+2x7}

With the conditions:

For all i xi>=0, xi<=1 and xi integer

x1+x2=1

x3+x4+x5=1

x6+x7=1

x1+x3=1

x2+x4+x6=1

x5+x7=1