DOC PREVIEW
Stanford MS&E 246 - Pure strategy Nash equilibrium

This preview shows page 1-2-3-18-19-36-37-38 out of 38 pages.

Save
View full document
View full document
Premium Document
Do you want full access? Go Premium and unlock all 38 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 38 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 38 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 38 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 38 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 38 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 38 pages.
Access to all documents
Download any document
Ad free experience
View full document
Premium Document
Do you want full access? Go Premium and unlock all 38 pages.
Access to all documents
Download any document
Ad free experience
Premium Document
Do you want full access? Go Premium and unlock all 38 pages.
Access to all documents
Download any document
Ad free experience

Unformatted text preview:

MS&E246:Lecture3PurestrategyNashequilibriumRamesh JohariJanuary 16, 2007Outline• Best response and pure strategyNash equilibrium• Relation to other equilibrium notions• Examples• Bertrand competitionBestresponsesetBest response set for player n to s-n:Rn(s-n) = arg maxsn∈ SnΠn(sn, s-n)[ Note: arg maxx ∈ Xf(x) is theset of x that maximize f(x) ]NashequilibriumGiven: N-player gameA vector s = (s1, …, sN) is a (pure strategy) Nash equilibrium if:si∈ Ri(s-i)for all players i.Each individual plays a best response to the others.NashequilibriumPure strategy Nash equilibrium is robust to unilateral deviationsOne of the hardest questions ingame theory:How do players know to play a Nash equilibrium?Example:Prisoner’sdilemmaRecall the routing game:(-2,-2)(-5,-1)far(-1,-5)(-4,-4)nearfarnearAT&TMCIExample:Prisoner’sdilemmaHere (near,near) is the unique (pure strategy) NE:(-2,-2)(-5,-1)far(-1,-5)(-4,-4)nearfarnearAT&TMCISummaryofrelationshipsGiven a game:• Any DSE also survives ISD, and is a NE.(DSE = dominant strategy equilibrium; ISD = iterated strict dominance)Example:biddinggameRecall the bidding game from lecture 1:$4$3$2$1$0$4.00$4.57$5.33$6.40$8.00$4$4.43$5.00$5.80$7.00$9.00$3$4.67$5.20$6.00$7.33$10.00$2$4.60$5.00$5.67$7.00$11.00$1$4.00$4.00$4.00$4.00$4.00$0Player 2’s bidPlayer 1’s bidExample:biddinggameHere (2,2) is the unique (pure strategy) NE:$4$3$2$1$0$4.00$4.57$5.33$6.40$8.00$4$4.43$5.00$5.80$7.00$9.00$3$4.67$5.20$6.00$7.33$10.00$2$4.60$5.00$5.67$7.00$11.00$1$4.00$4.00$4.00$4.00$4.00$0Player 2’s bidPlayer 1’s bidSummaryofrelationshipsGiven a game:• Any DSE also survives ISD, and is a NE.• If a game is dominance solvable, the resulting strategy vector is a NEAnother example of this: the Cournot game.• Any NE survives ISD (and is also rationalizable).(DSE = dominant strategy equilibrium; ISD = iterated strict dominance)Example:CournotduopolyUnique NE: (t/3 , t/3)t00s2R1(s2)R2(s1)tNash equilibrium =Any point where thebest response curvescross each other.s1Example:coordinationgameTwo players trying to coordinate their actions:(1,2)(0,0)r(0,0)(2,1)lRLPlayer 2Player 1Example:coordinationgameBest response of player 1:R1(L) = { l }, R1(R) = { r }(1,2)(0,0)r(0,0)(2,1)lRLPlayer 2Player 1Example:coordinationgameBest response of player 2:R2(l) = { L }, R2(r) = { R }(1,2)(0,0)r(0,0)(2,1)lRLPlayer 2Player 1Example:coordinationgameTwo Nash equilibria: (l, L) and (r, R).Moral: NE is not a unique predictor of play!(1,2)(0,0)r(0,0)(2,1)lRLPlayer 2Player 1Example:matchingpenniesNo pure strategy NE for this gameMoral: Pure strategy NE may not exist.(1,-1)(-1,1)T(-1,1)(1,-1)HTHPlayer 2Player 1Example:Bertrandcompetition•In Cournot competition, firms choosethe quantity they will produce.•In Bertrand competition, firms choosethe prices they will charge.Bertrandcompetition:model• Two firms• Each firm i chooses a price pi ≥ 0• Each unit produced incurs a cost c ≥ 0• Consumers only buy from the producer offering the lowest price• Demand is D > 0Bertrandcompetition:model• Two firms• Each firm i chooses a price pi• Profit of firm i:Πi(p1, p2) = (pi- c)Di(p1, p2)where0, if pi> p-iDi(p1, p2) = D, if pi< p-i½ D, if pi= p-iBertrandcompetition:analysisSuppose firm 2 sets a price = p2< c.What is the best response set of firm 1?Firm 1 wants to price higher than p2.R1(p2) = (p2, ∞)Bertrandcompetition:analysisSuppose firm 2 sets a price = p2> c.What is the best response set of firm 1?Firm 1 wants to price slightly lower than p2… but there is no best response!R1 (p2) = ∅Bertrandcompetition:analysisSuppose firm 2 sets a price = p2= c.What is the best response set of firm 1?Firm 1 wants to price at or higher than c.R1 (p2) = [c, ∞)Best response of firm 1:cBertrandcompetition:analysis00p2R1(p2)p1cBest response of firm 2:cBertrandcompetition:analysis00p2p1cR2(p1)Where do they “cross”?cBertrandcompetition:analysis00p2p1cR2(p1)R1(p2)Thus the unique NE is where p1= c, p2= c.cBertrandcompetition:analysis00p2p1cR2(p1)R1(p2)Unique NEBertrandcompetitionStraightforward to show:The same result holds ifdemand depends on price, i.e.,if the demand at price p is D(p) > 0.Proof technique:(1) Show pi< c is never played in a NE.(2) Show if c < p1< p2, then firm 2 prefers to lower p2.(3) Show if c < p1= p2, then firm 2prefers to lower p2BertrandcompetitionWhat happens if c1< c2?No pure NE exists; however, an ε-NE exists:Each player is happy as long as they arewithin ε of their optimal payoff.ε-NE : p2= c2, p1= c2- δ(where δ is infinitesimal)Bertrandvs.CournotAssume demand is D(p) = a - p.Interpretation: D(p) denotes the total number of consumers willing to payat least p for the good.Then the inverse demand isP (Q) = a - Q.This is the market-clearing price at which Q total units of supply would be sold.Bertrandvs.CournotAssume demand is D(p) = a - p.Then the inverse demand isP (Q) = a - Q.Assume c < a.Bertrand eq.: p1= p2= cCournot eq: q1= q2= (a - c)/3⇒ Cournot price = a/3 + 2c/3 > cCournottotal profits(Producersurplus)Bertrandvs.Cournot00pP (Q)QcaBertrand eq.(perfectly competitive)Cournot eq.ConsumersurplusCournottotal profits(Producersurplus)Bertrandvs.Cournot00pP (Q)QcaConsumersurplusDeadweight loss:Consumers arewilling to pay, andfirms could havemade a profit by sellingBertrandvs.Cournot• Cournot eq. price > Bertrand eq. price• Bertrand price =marginal cost of production• In Cournot eq., there is positive deadweight loss.This is because firms have market power:they anticipate their effect on prices.Questionstothinkabout•Can a weakly dominated strategy be played in a Nash equilibrium?•Can a strictly dominated strategy be played in a Nash equilibrium?• Why is any NE rationalizable?• What are real-world examples ofBertrand competition?Cournot competition?Summary:FindingNEFinding NE is typically a matter of checking the definition.Two basic approaches…FindingNE:Approach1First approach to finding NE:(1) Compute the complete best response mapping for each player.(2) Find where they intersect each other (graphically or otherwise).FindingNE:Approach2Second approach to finding NE:Fix a strategy vector (s1, …, sN).Check if any player has a profitable deviation.If so, it cannot be a NE.If not, it is an


View Full Document

Stanford MS&E 246 - Pure strategy Nash equilibrium

Download Pure strategy Nash equilibrium
Our administrator received your request to download this document. We will send you the file to your email shortly.
Loading Unlocking...
Login

Join to view Pure strategy Nash equilibrium and access 3M+ class-specific study document.

or
We will never post anything without your permission.
Don't have an account?
Sign Up

Join to view Pure strategy Nash equilibrium 2 2 and access 3M+ class-specific study document.

or

By creating an account you agree to our Privacy Policy and Terms Of Use

Already a member?