Norges teknisk-naturvitenskapelige universitet

TIØ4130 Optimeringsmetoder 2022-11-29

Exercise 1 (20 %) Consider the following maximization problem

(1)
X
max Ci x i
i∈N

(2)
X
Aij xi ≤ Bj j∈M
i∈N
xi ≥ 0 i∈N (3)
where the A matrix is given as
 
2 2 −1
 1 3 −1  (4)
2 1 4

The problem is solved using the simplex method and the optimal tableau is shown below
z x1 x2 x3 x4 x5 x6 b
z 1 0.00 5.00 0.00 1.00 0.00 1.00 16.00
x1 0 1.00 0.90 0.00 0.40 0.00 0.10 3.40
x5 0 0.00 1.90 0.00 −0.60 1.00 0.10 2.40 ⋆
x3 0 0.00 −0.20 1.00 −0.20 0.00 0.20 0.80

a) State B −1 .
b) Recreate the tableau from the previous iteration given that the last entering variable was x3 .
c) Calculate the Gomory cut from the constraint marked ⋆. Express the Gomory cut in the
original variables.
d) Assume that the cut found in c) is added as a variable to the dual problem. Calculate the
reduced cost of this variable given that it is added to the optimal tableau of the dual problem.

Exercise 2 (25 %)
Chris Mai is an entrepreneur. Chris sees business opportunities everywhere. Chris likes good food,
fancy cars and machine scheduling. Chris' current business adventures include the production
of ornaments for the big holidays and he is currently planning the production of ornaments for
December. Chris is also a fun guy and has named this collection of ornaments Chris Mai's Tree
Ornaments. Chris has received a set of D orders that must be produce before Christmas and has
a deal with a local producer to use a set of M machines for the production. Chris knows that the
cost of producing order d on machine m is Cdm that the time it takes is Tdm time units. He also
knows that he can use at most T m time units on machine m. Chris has formulated the following
model to support him in making production decisions.
Model X

(5)
X X
min Cdm xdm
d∈D m∈M

(6)
X
Tdm xdm ≤ T m m∈M
d∈D

(7)
X
xdm = 1 d∈D
m∈M

Side 1 av 7
Norges teknisk-naturvitenskapelige universitet
TIØ4130 Optimeringsmetoder 2022-11-29

xdm ∈ {0,1} d ∈ D, m ∈ M (8)

where xdm = 1 if order d is scheduled on machine m and 0 otherwise. Chris has discovered that
the solutions to the model cannot be used because of some technical considerations and asks you
to help him improve the model.

a) Assume that there is precedence between pair of orders, such that Pij = 1 if the production
of order j cannot start before order i is nished. Extend Model X with this requirement.
Clearly dene all sets, parameters and variables used.
b) The local producer informs Chris that the Italian machines used to produce the orders can
be run in P dierent modes. For each mode p ∈ P = {1, . . . , P }, the production time and
production cost for order d on machine m is Tdmp and Cdmp , respectively. According to
the local producer, the modes cannot change during the production of Chris Mai's Tree
Ornaments. Extend Model X with the possibility to choose modes for the machines.
c) Chris realizes that only using one mode per machine is inecient. He calls the Italian machine
producer Alva Julkuli and ask her what to do. Dr. Julkuli tells Chris that it is possible to
change modes, but that there is a setup time T for each time the mode of a machine is
changed. Extend Model X with the possibility to change modes for the machines.
d) Chris tested his model on a small example to check that it is correct. In this example, con-
straint (6) for machine 1 is
5x11 + 11x21 + 3x31 + 7x41 + 6x51 + 9x61 ≤ 20 (9)
Help Chris nd the largest minimum cover of this constraint, i.e., the minimum cover with
the most variables in it.
e) Given that the solution to the LP relaxation of the small example is
x11 = 2/5, x21 = 1, x31 = 1/3, x41 = 0, x51 = 1, x61 = 0 (10)
Formulate and solve the separation problem. Find the cover inequality given by the solution
to the separation problem and verify that it cuts away the solution to the LP relaxation.

Exercise 3 (25 %) Assume that we are analyzing a production planning problem with two
continuous variables, two binary variables, and four variables representing the breakpoints of a
piecewise linear function between the cost and capacity of a machine. The last four variables are
modeled as a special ordered set of type 2 (SOS2). The piecewise linear cost/capacity function
is shown in Figure 1.
The table below shows all solutions that can be created by xing one or more variables. Each
row in the table shows: Row number, the status of the two binary variables, (# means not xed,
0 means xed to 0, and 1 means xed to 1), the status of the four variables in SOS2, (# means
not xed, and 0 means that the variable belongs to the set of variables xed to 0), the value of
all eight variables, and the objective value for this solution.

1 : # # : # # # # : 60.00, 0.00, 1.00, 0.00, 0.16, 0.00, 0.00, 0.84 : 179.20

2 : # # : 0 # # # : 71.43, 0.00, 1.00, 0.29, 0.00, 0.00, 0.00, 1.00 : 168.57
3 : # # : # # 0 0 : 0.00, 16.67, 0.00, 0.00, 0.67, 0.33, 0.00, 0.00 : 108.33
4 : # # : 0 0 # # : 71.43, 0.00, 1.00, 0.29, 0.00, 0.00, 0.00, 1.00 : 168.57
5 : # # : # # # 0 : 0.00, 16.67, 0.00, 0.00, 0.83, 0.00, 0.17, 0.00 : 130.83
6 : # # : # 0 0 0 : 0.00, 0.00, 0.00, 0.00, 1.00, 0.00, 0.00, 0.00 : 0.00
7 : # # : 0 0 0 # : 71.43, 0.00, 1.00, 0.29, 0.00, 0.00, 0.00, 1.00 : 168.57

Side 2 av 7
Norges teknisk-naturvitenskapelige universitet
TIØ4130 Optimeringsmetoder 2022-11-29

1,620

1,215

Cost
675

0
0 50 100 150
Capacity

Figur 1: The piecewise linear cost/capacity function.

8 : # # : 0 # # 0 : 47.62, 0.00, 0.69, 0.00, 0.00, 0.00, 1.00, 0.00 : 40.24

9 : # 1 : # # # # : 71.43, 0.00, 0.29, 1.00, 0.00, 0.00, 0.00, 1.00 : 132.86
10 : # 1 : 0 # # # : 71.43, 0.00, 0.29, 1.00, 0.00, 0.00, 0.00, 1.00 : 132.86
11 : # 1 : # # 0 0 : 0.00, 50.00, 0.00, 1.00, 0.00, 1.00, 0.00, 0.00 : -45.00
12 : # 1 : 0 0 # # : 71.43, 0.00, 0.29, 1.00, 0.00, 0.00, 0.00, 1.00 : 132.86
13 : # 1 : # # # 0 : 0.00, 50.00, 0.00, 1.00, 0.50, 0.00, 0.50, 0.00 : 22.50
14 : # 1 : # 0 0 0 : 0.00, 0.00, 0.00, 1.00, 1.00, 0.00, 0.00, 0.00 : -370.00
15 : # 1 : 0 0 0 # : 71.43, 0.00, 0.29, 1.00, 0.00, 0.00, 0.00, 1.00 : 132.86
16 : # 1 : 0 # # 0 : 39.47, 17.11, 0.00, 1.00, 0.00, 0.00, 1.00, 0.00 : -19.21
17 : # 0 : # # # # : 60.00, 0.00, 1.00, 0.00, 0.16, 0.00, 0.00, 0.84 : 179.20
18 : # 0 : 0 # # # : 60.00, 0.00, 1.00, 0.00, 0.00, 0.24, 0.00, 0.76 : 146.80
19 : # 0 : # # 0 0 : 0.00, 16.67, 0.00, 0.00, 0.67, 0.33, 0.00, 0.00 : 108.33
20 : # 0 : 0 0 # # : 60.00, 0.00, 1.00, 0.00, 0.00, 0.00, 0.48, 0.52 : 114.40
21 : # 0 : # # # 0 : 0.00, 16.67, 0.00, 0.00, 0.83, 0.00, 0.17, 0.00 : 130.83
22 : # 0 : # 0 0 0 : 0.00, 0.00, 0.00, 0.00, 1.00, 0.00, 0.00, 0.00 : 0.00
23 : # 0 : 0 0 0 # : 60.00, 0.00, 1.00, 0.00, 0.00, 0.00, 0.00, 1.00 : -80.00
24 : # 0 : 0 # # 0 : 47.62, 0.00, 0.69, 0.00, 0.00, 0.00, 1.00, 0.00 : 40.24
25 : 1 # : # # # # : 60.00, 0.00, 1.00, 0.00, 0.16, 0.00, 0.00, 0.84 : 179.20
26 : 1 # : 0 # # # : 71.43, 0.00, 1.00, 0.29, 0.00, 0.00, 0.00, 1.00 : 168.57
27 : 1 # : # # 0 0 : 0.00, 50.00, 1.00, 0.00, 0.00, 1.00, 0.00, 0.00 : 5.00
28 : 1 # : 0 0 # # : 71.43, 0.00, 1.00, 0.29, 0.00, 0.00, 0.00, 1.00 : 168.57
29 : 1 # : # # # 0 : 0.00, 50.00, 1.00, 0.00, 0.50, 0.00, 0.50, 0.00 : 72.50
30 : 1 # : # 0 0 0 : 0.00, 0.00, 1.00, 0.00, 1.00, 0.00, 0.00, 0.00 : -320.00
31 : 1 # : 0 0 0 # : 71.43, 0.00, 1.00, 0.29, 0.00, 0.00, 0.00, 1.00 : 168.57
32 : 1 # : 0 # # 0 : 39.47, 17.11, 1.00, 0.00, 0.00, 0.00, 1.00, 0.00 : 30.79
33 : 1 1 : # # # # : 52.63, 39.47, 1.00, 1.00, 0.00, 0.00, 0.00, 1.00 : 111.05
34 : 1 1 : 0 # # # : 52.63, 39.47, 1.00, 1.00, 0.00, 0.00, 0.00, 1.00 : 111.05
35 : 1 1 : # # 0 0 : 0.00, 50.00, 1.00, 1.00, 0.00, 1.00, 0.00, 0.00 : -365.00
36 : 1 1 : 0 0 # # : 52.63, 39.47, 1.00, 1.00, 0.00, 0.00, 0.00, 1.00 : 111.05
37 : 1 1 : # # # 0 : 0.00, 83.33, 1.00, 1.00, 0.17, 0.00, 0.83, 0.00 : -35.83
38 : 1 1 : # 0 0 0 : 0.00, 0.00, 1.00, 1.00, 1.00, 0.00, 0.00, 0.00 : -690.00
39 : 1 1 : 0 0 0 # : 52.63, 39.47, 1.00, 1.00, 0.00, 0.00, 0.00, 1.00 : 111.05
40 : 1 1 : 0 # # 0 : 13.16, 72.37, 1.00, 1.00, 0.00, 0.00, 1.00, 0.00 : -49.74
41 : 1 0 : # # # # : 60.00, 0.00, 1.00, 0.00, 0.16, 0.00, 0.00, 0.84 : 179.20
42 : 1 0 : 0 # # # : 60.00, 0.00, 1.00, 0.00, 0.00, 0.24, 0.00, 0.76 : 146.80
43 : 1 0 : # # 0 0 : 0.00, 50.00, 1.00, 0.00, 0.00, 1.00, 0.00, 0.00 : 5.00
44 : 1 0 : 0 0 # # : 60.00, 0.00, 1.00, 0.00, 0.00, 0.00, 0.48, 0.52 : 114.40
45 : 1 0 : # # # 0 : 0.00, 50.00, 1.00, 0.00, 0.50, 0.00, 0.50, 0.00 : 72.50
46 : 1 0 : # 0 0 0 : 0.00, 0.00, 1.00, 0.00, 1.00, 0.00, 0.00, 0.00 : -320.00
47 : 1 0 : 0 0 0 # : 60.00, 0.00, 1.00, 0.00, 0.00, 0.00, 0.00, 1.00 : -80.00
48 : 1 0 : 0 # # 0 : 39.47, 17.11, 1.00, 0.00, 0.00, 0.00, 1.00, 0.00 : 30.79
49 : 0 # : # # # # : 20.00, 0.00, 0.00, 0.00, 0.72, 0.00, 0.00, 0.28 : 166.40
50 : 0 # : 0 # # # : 60.00, 0.00, 0.00, 1.00, 0.00, 0.24, 0.00, 0.76 : 96.80
51 : 0 # : # # 0 0 : 0.00, 16.67, 0.00, 0.00, 0.67, 0.33, 0.00, 0.00 : 108.33
52 : 0 # : 0 0 # # : 60.00, 0.00, 0.00, 1.00, 0.00, 0.00, 0.48, 0.52 : 64.40
53 : 0 # : # # # 0 : 0.00, 16.67, 0.00, 0.00, 0.83, 0.00, 0.17, 0.00 : 130.83
54 : 0 # : # 0 0 0 : 0.00, 0.00, 0.00, 0.00, 1.00, 0.00, 0.00, 0.00 : 0.00
55 : 0 # : 0 0 0 # : 60.00, 0.00, 0.00, 1.00, 0.00, 0.00, 0.00, 1.00 : -130.00
56 : 0 # : 0 # # 0 : 23.81, 0.00, 0.00, 0.10, 0.00, 1.00, 0.00, 0.00 : 27.86
57 : 0 1 : # # # # : 60.00, 0.00, 0.00, 1.00, 0.16, 0.00, 0.00, 0.84 : 129.20
58 : 0 1 : 0 # # # : 60.00, 0.00, 0.00, 1.00, 0.00, 0.24, 0.00, 0.76 : 96.80
59 : 0 1 : # # 0 0 : 0.00, 50.00, 0.00, 1.00, 0.00, 1.00, 0.00, 0.00 : -45.00

Side 3 av 7
Norges teknisk-naturvitenskapelige universitet
TIØ4130 Optimeringsmetoder 2022-11-29

60 : 0 1 : 0 0 # # : 60.00, 0.00, 0.00, 1.00, 0.00, 0.00, 0.48, 0.52 : 64.40

61 : 0 1 : # # # 0 : 0.00, 50.00, 0.00, 1.00, 0.50, 0.00, 0.50, 0.00 : 22.50
62 : 0 1 : # 0 0 0 : 0.00, 0.00, 0.00, 1.00, 1.00, 0.00, 0.00, 0.00 : -370.00
63 : 0 1 : 0 0 0 # : 60.00, 0.00, 0.00, 1.00, 0.00, 0.00, 0.00, 1.00 : -130.00
64 : 0 1 : 0 # # 0 : 39.47, 17.11, 0.00, 1.00, 0.00, 0.00, 1.00, 0.00 : -19.21
65 : 0 0 : # # # # : 20.00, 0.00, 0.00, 0.00, 0.72, 0.00, 0.00, 0.28 : 166.40
66 : 0 0 : 0 # # # : 20.00, 0.00, 0.00, 0.00, 0.00, 1.00, 0.00, 0.00 : -55.00
67 : 0 0 : # # 0 0 : 0.00, 16.67, 0.00, 0.00, 0.67, 0.33, 0.00, 0.00 : 108.33
68 : 0 0 : 0 0 # # : 20.00, 0.00, 0.00, 0.00, 0.00, 0.00, 1.00, 0.00 : -595.00
69 : 0 0 : # # # 0 : 0.00, 16.67, 0.00, 0.00, 0.83, 0.00, 0.17, 0.00 : 130.83
70 : 0 0 : # 0 0 0 : 0.00, 0.00, 0.00, 0.00, 1.00, 0.00, 0.00, 0.00 : 0.00
71 : 0 0 : 0 0 0 # : 20.00, 0.00, 0.00, 0.00, 0.00, 0.00, 0.00, 1.00 : -1000.00
72 : 0 0 : 0 # # 0 : 20.00, 0.00, 0.00, 0.00, 0.00, 1.00, 0.00, 0.00 : -55.00

a) Chris believes this is a minimization problem. Is he right? Motivate your answer.

b) Solve the problem using branch-and-bound and draw the branching tree. Use depth-rst as
search strategy. For the binary variables, branch on the most fractional value and start with
the = 0 branch. For the SOS2, use the value of the capacity of each breakpoint as the reference
value. Clearly show your calculations and motivate the branching choice. Write down the row
numbers from the table corresponding to the nodes that you solve in the order you solve them.
HINT: Variables can be excluded from either S 0 or S 1 . HINT2: Neither S 0 nor S 1 should be
empty.
c) What is the best dual bound after you have solved ve nodes including the root node?
d) Assume that the last four variables are modeled as a special ordered set of type 1 (SOS1). Solve
the problem using branch-and-bound and draw the branching tree. Use the same reference
values as in b).

Side 4 av 7
Norges teknisk-naturvitenskapelige universitet
TIØ4130 Optimeringsmetoder 2022-11-29

Exercise 4 (30 %)
AgriCoop is a retailer specializing on importing and selling agricultural goods to farmers in the
Trondheim area. After the last harvesting season, the CEO of AgriCoop decided to expand the
assortment within hay harvesting machines. Recently, AgriCoop signed an agreement with a new
distributor GrassTools Ltd. that can serve AgriCoop with two products: the MowerX and the
BaleWrapperPro. The harvesting season is approaching and the sales manager of AgriCoop has
made an estimate of the demand for each product for the next four weeks, which is shown in
Table 1. Currently, there are no units in storage, and the CEO has asked the sales manager to
make a plan for how much to order of each product for the next four weeks, considering that the
demand estimates should be met and that products can be stored from one period to the next. It
is assumed that AgriCoop has unlimited storage capacity, but the sales manager has estimated
that the cost of storing one unit of MowerX from one period to the next amounts to 5 000 NOK.
Similarly, the cost of storing BaleWrapperPro from one period to next amounts to 7 000 NOK.
Furthermore, the ordering cost consists of two parts: the unit cost of purchasing, which equals
100 000 NOK for MowerX and 80 000 NOK for BaleWrapperPro ; and a xed cost for handling
orders and freight, which amounts to 40 000 NOK. The xed cost occurs whenever AgriCoop
orders at least one unit of any product. This means that AgriCoop can reduce costs by jointly
planning the ordering of the products. GrassTools Ltd. has informed AgriCoop that up to 3 units
of each product can be ordered in each week. The sales manager assumes that the CEO wants a
plan that minimizes the total costs, that is, the sum of setup, purchasing and storage costs, and
which results in zero units in storage of both products at end of the four week period.
Tabell 1: Demand estimates
Weeks
Products 1 2 3 4
MowerX 1 1 2 2
BaleWrapperPro 0 1 1 2

Based on the problem description above, the sales manager formulates the ordering problem as
a dynamic program, and starts solving the problem based on backwards recursion. The compu-
tations of the sales manager can be found in the appendix. The computations are summarized in
one table per stage, and the tables list for each relevant state the value of performing the dierent
decisions (the value of decisions that would lead to a non-relevant state in the next stage is not
computed). The formulation of the sales manager assumes that the demand is fullled in the
end of each week, and that units ordered in a week can cover the demand in the same week.
Furthermore, the storage cost incurs in the end of each week, after the demand has been fullled.

a) Provide a dynamic programming formulation in correspondence with the computations of

the sales manager. Clearly dene stages, states, decisions, transition functions, cost function,
recursion function and initial conditions.

b) Some values in the computation table for Stage 2 have been replaced by the letters A, B, and
C. Find these values. Show your calculations.

c) What is the optimal objective value, and what is the optimal ordering quantity in each week?

Assume now the distributor GrassTools Ltd. oers three additional products with the same
agreement on xed costs, i.e., the xed costs occurs whenever AgriCoop orders at least one unit
of any of the ve products. To reduce unnecessary xed costs, the CEO of AgriCoop wants the
sales manager to include the additional products in the same planning problem. Also assume

Side 5 av 7
Norges teknisk-naturvitenskapelige universitet
TIØ4130 Optimeringsmetoder 2022-11-29

that the CEO wants the sales manager to increase the planning period to twelve weeks, where
it can be assumed that the sales manager is able to make good estimates on the demand for all
products over the twelve week period.

d) What consequences do the increased number of products and increased planning period have
on the computational burden of the dynamic programming approach? Justify your answer.

The sales manager proposes to formulate the ordering problem as a mixed integer linear program
instead.

e) Formulate the ordering problem for P products and T weeks as a mixed integer linear program.
Clearly dene indices, sets, parameters and variables.

Assume now that GrassTools Ltd. oers a new agreement to AgriCoop, which species a regime
for quantity discounts. In the regime, there is dened a number of discount levels L, and if
AgriCoop's total unit costs in any week is between Kℓ NOK and Kℓ+1 NOK, AgriCoop will have
a discount rate at level ℓ of Rℓ % on the unit costs. There is no discount for orders below K1
NOK, and you can assume that Rℓ < Rℓ+1 .

f) Formulate the problem extension with the quantity discount regime as a mixed integer linear
program. Clearly dene additional indices, sets, parameters and variables.

Side 6 av 7
Norges teknisk-naturvitenskapelige universitet
TIØ4130 Optimeringsmetoder 2022-11-29

Appendix: DP tables

Tabell 2: Stage 1
Value of making decision, f1 (x1 ,s1 )
State Decisions Opt. value Opt. decision
s1 (0,0) (0,1) (0,2) (0,3) (1,0) (1,1) (1,2) (1,3) (2,0) (2,1) (2,2) (2,3) (3,0) (3,1) (3,2) (3,3) f1 (s1 ) x∗1
(0,0) 1057 1064 1071 1092 1062 1052 1066 1087 1067 1046 1060 1074 1046 (3,1)

Tabell 3: Stage 2
Value of making decision, f2 (x2 ,s2 )
State Decisions Opt. value Opt. decision
s2 (0,0) (0,1) (0,2) (0,3) (1,0) (1,1) (1,2) (1,3) (2,0) (2,1) (2,2) (2,3) (3,0) (3,1) (3,2) (3,3) f2 (s2 ) x∗2
(0,0) 940 947 961 929 936 943 934 917 931 917 (3,2)
(0,1) 860 867 881 895 849 856 863 870 854 837 851 865 837 (3,1)
(0,2) 787 801 815 776 783 790 757 771 785 757 (3,0)
(0,3) 721 735 703 710 691 705 691 (3,0)
(1,0) 840 847 861 829 836 843 834 817 831 839 827 841 817 (2,2)
(1,1) 720 767 781 795 749 756 763 770 754 C 751 765 759 747 761 775 720 (0,0)
(1,2) 647 701 715 676 683 690 657 671 685 667 681 695 647 (0,0)
(1,3) 581 635 603 610 591 605 601 615 581 (0,0)
(2,0) 729 736 743 734 717 731 739 727 741 744 737 751 A B
(2,1) 609 656 663 670 654 637 651 665 659 647 661 675 664 657 671 645 609 (0,0)
(2,2) 536 583 590 557 571 585 567 581 595 577 591 565 536 (0,0)
(2,3) 463 510 491 505 501 515 511 485 463 (0,0)

Tabell 4: Stage 3
Value of making decision, f3 (x3 ,s3 )
State Decisions Opt. value Opt. decision
s3 (0,0) (0,1) (0,2) (0,3) (1,0) (1,1) (1,2) (1,3) (2,0) (2,1) (2,2) (2,3) (3,0) (3,1) (3,2) (3,3) f3 (s3 ) x∗3
(0,0) 720 727 734 725 732 739 720 (2,1)
(0,1) 640 647 654 645 652 659 640 (2,0)
(0,2) 567 574 572 579 567 (2,0)
(0,3) 494 499 494 (2,0)
(1,0) 620 627 634 625 632 639 630 637 604 604 (3,3)
(1,1) 540 547 554 545 552 559 550 557 524 524 (3,2)
(1,2) 467 474 472 479 477 444 444 (3,1)
(1,3) 394 399 364 364 (3,0)
(2,0) 520 527 534 525 532 539 530 537 504 504 (2,3)
(2,1) 400 447 454 445 452 459 450 457 424 400 (0,0)
(2,2) 327 374 372 379 377 344 327 (0,0)
(2,3) 254 299 264 254 (0,0)
(3,0) 425 432 439 430 437 404 404 (1,3)
(3,1) 305 352 359 350 357 324 305 (0,0)
(3,2) 232 279 277 244 232 (0,0)
(3,3) 159 164 159 (0,0)
(4,0) 330 337 304 304 (0,3)
(4,1) 210 257 224 210 (0,0)
(4,2) 137 144 137 (0,0)
(4,3) 24 24 (0,0)

Tabell 5: Stage 4
Value of making decision, f4 (x4 ,s4 )
State Decisions Opt. value Opt. decision
s4 (0,0) (0,1) (0,2) (0,3) (1,0) (1,1) (1,2) (1,3) (2,0) (2,1) (2,2) (2,3) (3,0) (3,1) (3,2) (3,3) f4 (s4 ) x∗4
(0,0) 400 400 (2,2)
(0,1) 320 320 (2,1)
(0,2) 240 240 (2,0)
(1,0) 300 300 (1,2)
(1,1) 220 220 (1,1)
(1,2) 140 140 (1,0)
(2,0) 200 200 (0,2)
(2,1) 120 120 (0,1)
(2,2) 0 0 (0,0)

Side 7 av 7