Assignment Week 4
February 4, 2025
Convention: In a zero-sum game, the column player maximises his payoff. In a
non-zero-sum game, both players are maximizing their payoff.
1 Assignments
Question 1. A consequence of von Neumann’s Minimax Theorem is that a mixed saddle
point strategy always exists for a two-player zero-sum game, even when the strategy sets
are infinite.
(a). True.
(b). False.
Answer 1. False.
Question 2. Let
T 1 -4
A=
B -1 3
be the payoff matrix for a two-player zero-sum game. Suppose we compute the value of
V̄m (A) by formulating it as an LP problem. Then which of the following point(s) of the
form (t, y) (refer to notes/lectures for notation) lies in the feasible region of the LP?
(a). (0, 0.75, 0.25).
(b). (1, 0, 1).
(c). (8, 0.5, 0.5).
(d). (12, 1.1, −0.1).
(e). None of these.
Answer 2. (8, 0.5, 0.5).
Question 3. Consider a primal linear programming (LP) problem and its corresponding
dual. Suppose we take the dual of the dual problem. Then, the resulting LP problem is
(select the correct option(s)).
1
2
(a). Is always the same as the original primal problem.
(b). May differ from the original primal problem.
(c). Is always infeasible.
(d). Has the same feasible region as the dual.
(e). None of these.
Answer 3. Is always the same as the original primal problem.
Question 4. Let
P 1\P 2 L R
A= T 0 -3
B -1 2.5
be the payoff matrix of a two-player zero-sum game. What is the value of the game (up
to 2 decimal places)? If there is no value, then enter 999.
Answer 4. -0.46.
Question 5. Let
P 1\P 2 L R
T 2 3
A=
M 5 4
B 7 4
be the payoff matrix of a two-player zero-sum game. What is the value of the game? If
no value exists, then enter 999.
Answer 5. 5.2.
Question 6. Let A = (aij ) and B = (bij ) be two n×m matrices representing two-player
zero-sum games in strategic form. Select the correct option(s).
(a). The absolute difference between the values of the two games is less than or equal
to
n m
max max |aij − bij | .
i=1 j=1
(b). The absolute difference between the values of the two games is greater than
n m
max max |aij − bij | .
i=1 j=1
(c). The absolute difference between the values of the two games is greater than or equal
to n m
min min |aij − bij | .
i=1 j=1
(d). The absolute difference between the values of the two games is less than
n m
min min |aij − bij | .
i=1 j=1
3
(e). None of these
Answer 6. The absolute difference between the values of the two games is less than or
equal to
n m
max max |aij − bij | .
i=1 j=1
Question 7. Let Mn,m be the space of matrices of order n × m representing two-player
zero-sum games in which (row) Player I has n pure strategies and (column) Player II has
m pure strategies. Select the correct option(s).
(a). The function that associates with every matrix A = (aij ) ∈ Mn,m the value in mixed
strategies of the game that it represents is continuous in (aij ).
(b). The function that associates with every matrix A = (aij ) ∈ Mn,m the value in mixed
strategies of the game that it represents is continuous in (aij ) if and only A is a
square matrix, i.e., m = n.
(c). The function that associates with every matrix A = (aij ) ∈ Mn,m the value in mixed
strategies of the game that it represents need not be continuous in (aij ).
(d). None of these.
Answer 7. The function that associates with every matrix A = (aij ) ∈ Mn,m the value
in mixed strategies of the game that it represents is continuous in (aij ).
Question 8. Consider the following payoff matrix of a zero-sum game where P1 is the
row player who wants to minimize his payoff and P2 is the column player who wants to
maximize his payoff.
P 1\P 2 L R
T a b
B c d
Which of the following condition(s) ensures that there is NO pure saddle point strategy?
Select the correct option(s).
(a). max{min(a, b), min(c, d)} < max{min(a, c), min(b, d)}.
(b). min{min(a, b), min(c, d)} < max{max(a, c), max(b, d)}.
(c). min{max(a, b), max(c, d)} < min{max(a, c), max(b, d)}.
(d). max{min(a, b), min(c, d)} < min{max(a, c), max(b, d)}.
(e). None of these.
Answer 8. max{min(a, b), min(c, d)} < min{max(a, c), max(b, d)}.
