Computer Network Assignment Help

For any Assignment related queries, Call us at : - +1 678 648 4277
You can mail us at : - support@computernetworkassignmenthelp.com or
reach us at : - https://www.computernetworkassignmenthelp.com/

Problem 1: Consider the game below.
(a) What is player 1’s minimax payoff?
Suppose player 2 chooses action A with probability q and B with probability 1-q.
Player 1 earns in expectation 2q from playing A and 1-q from playing B. To find the
minimax payoff, we solve
Player 2 delivers the worst possible payoff by choosing q = 1/3 , and player 1 is then
indifferent between A and B.
(b) Describe the set of all Nash equilibria (pure and mixed).
computernetworkassignmenthelp.com

There are two pure straegy Nash equilibria (A, A) and (B, B). There is also exactly one
mixed strategy Nash equilibria. Using the solution from the last problem, we see that if
player 2 chooses action A with probability 1 3 , player 1 is indifferent between the two
actions. By symmetry, if player 1 chooses action B with probability 1 3 , player 2 is
indifferent between the two actions. The mixed strategy equilibrum involves player 1
choose A with probability 2 3 and B with probability 1 3 , and player 2 choosing A with
probability 1 3 and B with probability 2/3.
Suppose player 1 (the row player) is uncertain about player 2’s preferences. The actual
game is one of the following two games:
Player 2 learns her preferences before making a choice, while player 1 must make a
choice without further information. The common prior is that the first game is played
with probability p, and the second with probability 1 - p.
(c) As a function of p, describe the set of all (pure strategy) Bayes-Nash equilibria.

With two types of player 2, there are exactly 8 pure strategy profiles. Notice for the
second type that action B dominates action A. Therefore, we need only consider four
possible strategy profiles: (A,(A, B)), (A,(B, B)), (B,(A, B)), and (B,(B, B)). The middle
two are clearly not best responses for the first type of player 2.
In the profile (A,(A, B)), player 2 is clearly playing a best response. Player 1 earns 2p
from action A and 1- p from action B. It is a best response for player 1 to choose action
a if and only if 2p ≥ 1- p or p ≥ 1/3 . In the profile (B,(B, B)), player 1 is clearly playing
a best reply, so this is always an equilibrium.
Problem 2: Each player i in a population of size N makes a non-negative contribution
xi ∈ R to a public good. If the vector of investments is (x1, x2, ..., xN ), player i’s
payoff is
(a) Write down player i’s best response map. What is the total level of investment
in any Nash equilibrium? What is the efficient level of investment?
Taking the first order condition, we have

with equality whenever player i invests effort. Player i’s best response is therefore
In equilibrium, the total level of investment must be exactly one, since no one invests
above this amount in a best reply, and someone will invest 1 if no one else invests.
To find the efficient level of investment, we maximize
The first order condition gives
The efficient level of investment is much higher than the equilibrium level and increases
in with N.

Suppose instead that spillovers from other players’ investments are imperfect, and player
i’s payoff is
for some δ ∈ (0, 1).
(b) Write down player i’s best response map. What is the total level of investment
in a symmetric Nash equilibrium? What is the efficient level of investment?
The first order condition gives
with equality whenever player i invests. Player i’s best response is
In a symmetric strategy profile, we have xi = x∗ for all i. This is an equilibrium if and
only if

implying a total level of investment
The efficient level of investment maximizes
Taking a derivative with respect to xi , we get
Taking advantage of symmetry, the efficient investment for an individual player x0 satisfies

implying
and efficient total investment of N times this quantity.
Suppose N = 2 and now the game takes place in two stages. In the first stage, players 1
and 2 invest efforts s1, s2 ≥ 0 at constant marginal cost c to establish a relationship.
These investments result in the tie strength δ(s1, s2) = min{s1 +s2, 1}. Once these
investments are made, the resulting tie strength is observed, and we move to the second
stage. In the second stage, the players invest as above. The payoff to player i is
(c) As a function of c, what tie strength forms in equilibrium? What is the efficient outcome?
We use our result from the last part. Fixing the tie strength, player i invests 1
in equilibrium, yielding a total payof

The derivative of δ with respect to si is exactly x-i if δ < 1 and 0 if δ ≥ 1. Therefore, if
c < x-i, player i will invest up to the point where δ = 1, and if c > x-i, player i will
choose si = 0.
We therefore have 2 possible cases. If δ = 1 in equilibrium, then from part (a) we know
players are willing to invest up to 1. From our above analysis, in order to have δ = 1 in
equilibrium, we must have at least one of the players investing an amount at least c. If c
≤ 1, we can have such an equilibrium, with possibly asymmetric strategies. If δ = 0 in
equilibrium, then both players invest xi = 1. For this to be an equilibrum, we must have
c ≥ 1, otherwise the players would wish to invest in a link. Consequently, we have δ = 1
in any equilibrum with c < 1, and δ = 0 in equilibrium if c > 1.
In an efficient strategy profile, we get a similar all-or-nothing result on link investment.
If δ = 1, it is efficient for each player to invest 2, yielding total welfare 6 -c. If δ = 0,
each player efficiently invests 1, yielding total welfare 2. It is efficient to invest in the
link as long as c ≤ 4.
Problem 3: Consider a variant of the mean-field diffusion model from the first lecture
on diffusion. Each agent in a large population chooses between two actions 0 and 1.
Agents have degrees drawn independently from the distribution D and private values
drawn independently

from a uniform distribution on [0, 1]. If an agent has degree d and value v, and a
neighbors adopt, the payoff to adoption is
u(d, v, a) = av - c.
That is, the payoff to adoption increases linearly in the number of neighbors who
end up adopting. Recall the neighbor degree distribution D˜ that corrects for the
friendship paradox:
Time is discrete. Let σt,d denote the fraction of degree d agents adopting at time t. At
time t + 1, each agent chooses an action to maximize expected utility, assuming that
neighbors will each adopt with independent probability
(a) Suppose the degree distribution D takes the value 3 with probability one. Compute
σt+1,3 as a function of σt,3. Find the steady state adoption levels. Which are stable?

First note that having no players adopt is always an equilibrium. If D is 3 with
probability one, then so is D˜. For a player with value v, the expected value of a is just
The expected payoff from adoption is 3σt,3v -c, so a player is adopts in period
t + 1 if v > Since v is uniform on [0, 1], the best response map is then
To compute the steady states, we solve for fixed points of the best response map. We solve
Positive steady state adoption levels are possible if c ≤ 3/4.
For c strictly below 3/4 , there are two equilibria. Notice that in between the two steady
where the inequality follows
be- cause c < 3/4 . Best response dynamics lead to increasing adoption for σt,3
between the steady states, and decreasing adoption above the high equilibrium or
below the low equilibrium. This means the high equilibrum is stable, and the low
equilibrium is a tipping point. The non-adoption equilibrium is also stable.

(b) Suppose now that D takes the values 2 and 4 with equal probability. Write down the
neighbor degree distribution D˜. Compute the best response maps σt+1,2 and σt+1,4 as a
function of σt,2 and σt,4. Find the steady state adoption levels.
From the definition we have
We then have P(D˜ = 4) = 2/3 .
This means that for a degree 2 player with value v, the expected payoff to adoption is
and for a degree 4 player it is

To find steady state adoption levels, we can compute
and solve for the fixed points. Note the fixed points are exactly the same as in part
(a). Given the fixed point σ, a fraction
of degree 2 players adopt, and a fraction
degree 4 players adopt.

Computer Network Assignment Help

More Related Content

Similar to Computer Network Assignment Help (20)

More from Computer Network Assignment Help (20)

Recently uploaded (20)

Computer Network Assignment Help