## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

GAME THEORY (GT)

Thus far, we have considered situations where a single DM chooses an optimal decision without reference to the effect his decision has on other DM's Ðand without reference to the effect the decision of others may have on himÑ. A computer company, for example, must determine an advertising policy and pricing policy for its computers and each computer manufactures's decision will effect the revenues & profits of other computer manufacturers. Noncooperative GT is useful for making decisions in cases where two or more DM's have conflicting interests. Mostly, we shall be concerned with two decision makers. However, GT extends to more than two DM's. 1. TWO PERSON ZERO SUM & CONSTANT SUM GAMES Characteristics of 0-sum two person Games: 1. 2. 3. There are two players Ðcalled row player and column playerÑ The row player must choose 1 of m strategies. Simultaneously, the column player must choose 1 of n strategies. If the row player chooses his/hers ith strategy and the column player chooses his/her jth strategy the row player receives a reward of aij and the column player looses an amount aij . Thus, we may think of the row player's reward of aij as coming from the column player. This is a 2 person 0 sum game : the matrix aij is the game's reward matrix. Row Player's Column player's strategy

strategy Col 1 Col 2 ... Col n _____________________________________________________ Row 1 a11 a12 ... a1n Row ã Row m Example 1: 1 2 2 3 1 1 2 0 a21 ã am1 a22 ã am2 ... ... a2n ã amn Reward Matrix

Row player receives 2 units if the row player chooses the second strategy and the column player chooses the first strategy. 2 person 0-sum game: for any choice of strategies the sum of rewards to the players is zero. Every £ one player wins comes out of the other player's pocket. Thus the two players have totally conflicting interests no cooperation can occur. John-von Neumann and Oscar Morgenstern (Theory of Games and Economic Behaviour, J. Wiley 1943) developed the theory of 0-sum 2 person games and how they should be played. Basic Assumption of 2 person 0-sum games: Each player chooses a strategy that enables him to do the best he can given that the opponent knows the strategy he is following.

GT (2003) 2

Row Player's Column player's strategy Row Strategy Col 1 Col 2 Col 3 minimum ______________________________________________________________ Row 1 4 4 10 4 Row 2 2 3 1 1 Row 3 6 5 7 5 ______________________________________________________________ Column maximum 6 5 10 How should the Row Player (RP) play this game? If RP chooses R1, the assumption implies that the Column Player (CP) will choose C1 or C2 and hold the RP to a reward of 4 Ðthe smallest number in row 1 of the game matrixÑ. If RP chooses R2, CP will choose C3 and hold the RP's reward to 1 (the smallest-minimum in the second rowÑ. If RP chooses R3 then CP will allow him 5. Thus, the assumption Ê RP should choose the row having the largest minimum. Since max Ð4, 1, 5Ñ œ 5, RP chooses R3. This ensures a win of at least max Ðrow minimumÑ œ 5. If the CP chooses C1, the RP will choose R3 Ðto maximise earningsÑ. If CP chooses C2 the RP will choose R3. If the CP chooses C3 the RP will choose R1 (10 œ max Ð10, 1, 7ÑÑ. Thus the CP can hold his losses to min Ðcolumn maxÑ œ min Ð6, 5, 10Ñ œ 5 by choosing C2. Thus, the RP can ensure at least 5 ÐwinÑ and the CP can hold the RP's gains to at most 5. Thus, the only rational outcome of this game is for the RP to win 5. The RP cannot expect to win more because the CP Ðby choosing C2Ñ can hold RP's win to 5. The game matrix we have analysed satisfies the SADDLE POINT CONDITION property maximum minimum all Ð row minimum Ñ œ over all Ð column maximum Ñ over rows columns (1)

Any 2 person 0-sum game Ð2p0sgÑ satisfying Ð1Ñ is said to have a SADDLE POINT. If a 2p0sg has a saddle point the RP should choose any R strategy attaining the maximum on the LHS of Ð1Ñ and a CP should choose a C strategy attaining the minimum on the RHS. In the game considered a saddle point occurred at R3 and C2. If the game has a saddle point we call the common value to both sides of Ð1Ñ the VALUE ÐvÑ of the game. In the above case v œ 5. An easy way of identifying saddle points is to observe that the reward for a saddle point must be the smallest number in its row and the largest number in its column. Like the centre point of a horse's saddle, a saddle point for a 2p0sg is a local minimum in one direction Ðlooking across the rowÑ and local maximum in another direction Ðdown the columnÑ. A saddle point can also be seen as an EQUILIBRIUM POINT in that neither player can benefit from a unilateral change from the optimal strategy Ðof Row 3 to either R1 or 2Ñ since his reward would decrease. If the column player changed from the optimal strategy Ðof C2 to C1 or 3Ñ RP's reward would increase. Ê A saddle point is stable in that neither player has an incentive to deviate from it. Many 2p0sg's do not have saddle points. Example 2: 1 ” 1 1 1•

Max Ðrow minÑ œ 1 min Ðcol. maxÑ œ 1 TWO PERSON CONSTANT SUM GAMES Ð2PCSGÑ Two players can still be in total conflict. Definition: A 2pcsg is a two player game in which, for any choice of both players' strategies, the RP's reward and the CP's reward add up to a constant c. Note: 2p0sg is a 2pcsg with c œ 0. 2pcsg maintains the total conflict between RP and CP. A unit increase in RP's reward Ê a unit decrease in CP's reward.

1 † 4Ñ œ 4 † 5m viewers. Neither side will do better if it unilaterally changes strategy Ðcheck thisÑ. 2. it is unclear how to determine the value of the game and the optimal strategies. We discuss how one can find the value and optimal strategies for a 2p0sg that does not have a saddle point. N2 can hold N1 at most min Ð4 † 5. Reward matrix: R. the value of the game to N1 œ 4 † 5m. Column player : E. maximumÑ œ 4 † 5 Ð1Ñ is satisfied. viewers and the value of the game to N2 œ 10 † 0 4 † 5 œ 5 † 5m The optimal strategy for N1 is soap and N2 is western. 2P0SG RANDOMISED STRATEGIES NOT ALL 2P0SG HAVE SADDLE POINTS. O can be sure of a reward of 1 Ðat leastÑ and E can hold O to a reward of at most 1. N2 gets (10 3. choosing a western.GT (2003) 3 The optimal strategies and value of a 2pcsg can be found by the same methods used to solve a 2p0sg. maxÑ 1. Example 3: TV. N1 choosing a soap and N2 choosing a western yield a saddle point. Looking at column maxima. If the sum of the fingers put out by both players is odd. O wins £1 from E. Since max Ðrow minimumÑ œ Ðcol. 5 † 8. Possible choices for N1 & N2 and the number of N1 viewers ÐmillionsÑ for each choice are: N2 Western Soap Comedy Row minimum ____________________________________________________ W 3†5 1†5 6†0 1†5 S 4†5 5†8 5†0 4†5 N1 C 3†8 1†4 7†0 1†4 ____________________________________________________ Column 4†5 5†8 7†0 max. Thus. If the sum of the fingers is even. Player: E row min 1 1 1 Finger 2 Fingers ________________________________________ 1 Finger 1 1 2 Fingers 1 1 ________________________________________ Col. Two channels competing for an audience of 10 million. Each channel ÐN1 & N2Ñ must simultaneously announce their programme. Player: O C. Row player: Odd. 7 † 0Ñ œ 4 † 5m viewers. For any choice of strategies by both . E wins £1 from O. Value of the game for N1? b a saddle point? If both N1 and N2 show W then N1 gets 3 † 5m viewers. max 1 1 This is a 0 sum game: the amount gained by one player = . Since max Ðrow minÑ œ 1 and min Ðcol. Ð1Ñ is not satisfied Ê no saddle point. Thus. Example 4: ODDS and EVENS 2 Players: Odd ÐOÑ and Even ÐEÑ simultaneously choose the number of fingers Ð1 or 2Ñ to put out.the amount lost by the other. 2pcsg with c œ 10m. N1 can be sure of at least max Ð1 † 5. 8 9 pm slot. Looking at the row minima: choosing a soap.5 œ ) 6 † 5m. Thus. 4 † 5.

0. We continue to assume that both players play a 2p0sg in accordance of the basic assumption on page 1. Similarly if y1 .. any mixed strategy Ðy1 . A pure strategy is a special case of a mixed strategy in which a player always chooses the same action.. yn Ñ for C player is a pure strategy if for any. b − ‘m ÐD Ñ .. x2 0 and x1 x2 œ 1. 1. Thus. RANDOMISED OR MIXED STRATEGIES Until now we have assumed that each time a player plays a game. FURTHER LINEAR PROGRAMMING: DUALITY & COMPLEMENTARY SLACKNESS 3.. A − ‘m‚n . x % ‘n . y2 Ñ is a mixed strategy for E.Ñ and Ðy1 .GT (2003) 4 players. Resources needed for each D. the player will choose the same strategy. no choice of strategies by both players is stable. CH Ð PÑ c. x 0 › .. Any mixed strategy Ðx1 ... 0Ñ. strategy for O. The manufacture of each type of furniture requires wood ÐWÑ and two types of labour: finishing ÐFIÑ and carpentry ÐCAÑ. xm Ñ for R player is a PURE STRATEGY if any of the xi œ 1. E. then O can increase O's reward from 1 to 1 by changing from 1 to 2 fingers. x2 . y2 0 and y1 y2 œ 1. y2 . Ð " #. x2 Ñ is a randomized.G.. or mixed. 3.1 DUALITY Consider the LP Ðcalled the PRIMAL problemÑ max š cT x ¹ A x Ÿ b . Example 6: In the example on page 2. The DUAL of this problem is defined as min š bT y ¹ AT y c.) to maximise their own expected reward. Why not allow each player to select a probability of playing each strategy? Example 5: x1 x2 y1 y2 œ œ œ œ probability that O puts out 1 fingers probability that O puts out 2 fingers probability that E puts out 1 finger probability that E puts out 2 fingers " If x1 0. tables ÐTÑ and chairs ÐCHÑ.. there is a player who can benefit by unilaterally changing his/her strategy.. T. Ðx1 . 1Ñ and C's strategy was the pure strategy Ð0. Example 7: Furniture company ÐFCÑ manufactures desks ÐDÑ. Each player knows that his/her opponent will choose the probabilities Ðx1 . i.. x2 . R's optimal strategy could be represented as the pure strategy Ð0... Hence the DUAL of the DUAL is the PRIMAL. The choice of the strategy itself is done according to the probabilities that have been computed. yi œ 1. the game had a value of 5 Ðsaddle pointÑ. e.g. on a particular play of the game E is not assumed to know O's actual strategy choice until the game is played. In the context of randomised strategies the assumption Ðfrom the point of view of the oddÑ may be stated as follows: Odd should choose x1 and x2 to maximise O's expected reward under the assumption that Even knows the values of x1 . Similarly. y 0 › y − ‘m NOTE: The DUAL of ÐDÑ is ÐPÑ. #Ñ could be realised by O if O tossed a coin before each play of the game and put out 1 finger for heads and 2 fingers for tails. if both players put out 1 finger. Ðy1 .

Similarly the second dual constraint is associated with tables and the third with chairs. ConstraintÑ 4x1 2 x2 1 † 5x3 Ÿ 20 ÐFI. x1 œ number of D.GT (2003) 5 Resources D T CH ________________________________________________ W 8 ft 6 ft 1 ft FI CA 4 hrs 2 hrs 2 hrs 1 † 5 hrs 1 † 5 hrs 0 † 5 hrs Available: 48 ft wood. x3 0 The DUAL problem: min S. FC wishes to maximise revenue. y2 . Also. x3 œ number of CH A Desk sells £60. 60x1 30x2 20x3 8x1 6 x2 x3 Ÿ 48 ÐW. y1 is associated with wood y2 with finishing hours and y3 with carpentry hours. The PRIMAL problem: max S. Resource/Product Resources D T CH ____________________________________________ W 8ft 6ft 1ft FI 4 2 1†5 resources Available 48 ft 20 hrs 8 hrs CA 2 1†5 †5 ____________________________________________ Price ÐsellingÑ £60 £30 £20 To interpret the Dual: You are an entrepreneur who wants to purchase all of FC's resources.D & CH unlimited. 48y1 20y2 8y3 8y1 4y2 2y3 60 ÐDesk constraintÑ 6y1 2y2 1 † 5y3 30 ÐTable constraintÑ y1 1 † 5y2 0 † 5y3 20 ÐChair constraintÑ y1 .T. a Table for £30 . Then you must determine the price you are willing to pay for a unit of each of FC's resources.T. y3 0 The first constraint corresponds to x1 ÐDESKSÑ because each number in the first dual constraint comes from the x1 ÐDESKÑ column of the primal. You wish to minimise the cost of your purchase: . constraintÑ 2x1 1 † 5x2 † 5x3 Ÿ 8 ÐCA constraintÑ x1 . The total price you must pay for all of these resources is 48y1 20y2 8y3 . x2 . Demand for T. x2 œ number of T. 20 F1 hours and 8 CA hours. y3 should be determined by solving the DUAL above. Define y1 : price paid for 1ft of wood y2 : price paid for 1 finishing hour y3 : price paid for 1 carpentry hour Now we show that the resource price y1 . a Chair for £20. y2 .

t. In the FC problem x1 œ x2 œ x3 œ 1 is feasible. x. A − ‘m‚n . to the ith primal constraint. The sign restrictions y1 . in a natural way. if it wished. y.r.. DUAL THEOREM ÐDTÑ 1 P: max š cT x¹ A x Ÿ b. the third ÐchairÑ constraint ÐdualÑ y1 1 † 5y2 0 † 5y3 20 states that you must pay at least £20 Ðprice of a chairÑ for the resources needed to produce a chair Ð1ft. This has an objective function value of T T T xT AT y xT c .5 y3 30 Similarly. 0 † 5 CA hrsÑ. This means that y1 .5 CA hrs). y 0 ›. Thus. x 0 ›. y be feasible w. multiplying A x Ÿ b by yT yields yT Ax Ÿ yT b Since x 0 multiplying AT y c by xT yields Since y A x œ x A y. 2 of CA hrs since FC could. Lemma 1 Let x. x: Ax Ÿ b. 4 of FI hrs. W. use these to produce a desk & sell it for £60. 1 † 5 FI hrs. y 0 Then cT x Ÿ bT y Proof Since y 0. c − ‘n D: min š bT y¹ AT y c. we have xT c Ÿ xT AT y œ yT Ax Ÿ yT b. x 0 and y: AT y c. Example 8: If a feasible solution to either the primal or the dual is known.. y2 . y3 must satisfy 6y1 2y2 1. it can be used to obtain a bound on the optimal value of the objective function of the other problem. y2 y3 0 must also hold.e. 1. The ith dual variable thus corresponds. 8y1 4y2 2 y3 60 Similar reasoning shows that you must pay at least £30 for the resources used to produce a table (6 ft of W. the primal and the dual problems respectively i. 2 FI hrs. b − ‘m DT states that P and D have equal optimal objective function values.GT (2003) 6 min 48y1 20y2 8y3 In setting resource prices what constraints do you face? You must set the resource prices high enough to induce FC to sell you its resources: You must offer at least £60 for a combination of the resources that includes 8ft of W.

and y ^ − ‘m : AT y c.e. y Õ8Ø Õ 10 Ø cT x œ 280 œ bT y. any primal feasible x must yield an objective function value cT x that does not exceed bT y. y corresponds the smallest value b y can take. T T x0 = cB xB + cN xN Ê B1 b ( Ê xB œ B1 b . Let x ^ − ‘n : A x Ÿ b. ^ x ^ corresponds to the largest value cT x can take. then ^ If cT x x is optimal for ÐPÑ and ^ y is optimal for ÐDÑ.GT (2003) 7 60Ð1Ñ 30Ð1Ñ 20Ð1Ñ œ 110 Lemma 1 Ê Any dual feasible solution must satisfy 48y1 20y2 8y3 110 Lemma 2 ^ be a feasible solution to the primal i. Example 9: ^ œ For the FC problem x Ô2× Ô 0 × ^ œ 10 Ê 0 . Recall: Ax œ b x0 = cT x xB B1 N xN œ Ê B xB N xN œ b. ^ œ bT y. Hence y is an optimal solution for the dual. primal feasible and has objective function value cT x ^ is optimal. Proof Lemma 1 Ê for any feasible point x ^ cT x Ÿ bT y ^ Since x ^ is Thus. Lemma 4: If ÐDÑ is unbounded then ÐPÑ is infeasible. y 0 › and the dual Suppose B is an optimal basis for the primal.e. DUAL THEOREM 2 Given the primal problem max š cT x ¹ A x Ÿ b. x 0 x ^ be feasible solution to the dual i. Lemma 3: If ÐPÑ is unbounded then ÐDÑ is infeasible. T T^ ^ T ^ ^ y b œ c x. x 0 › min š bT y ¹ AT y c. xN œ 0) T 1 cB B b ( = yT b) 1 T x0 + ( cT N cN ) xN œ B B . Similarly Lemma 1 Ê Hence x ^ Ÿ yT b cT x ^ Since y ^ is dual feasible and objective value and for any dual feasible y the objective function yT b exceeds cT x. y 0 y ^ œ y ^T b.

We have shown that y œ B cB satisfies all n constraints of the dual problem and that all elements of B-T cB are nonnegative. 0. . We now need to show that dual objective function value for B-T cB = primal objective value for B. 1 T T 1 1 From x0 + ( cT N cN ) xN œ cB B b we know that the primal objective value is x0 œ cT B B B B -T b. For the case in which a slack variable is basic. note that it can be shown that the corresponding element of B-T cB is zero. Having found a primal feasible solution from B and a dual feasible solution. for the optimal values of x and y. m.. Since B is primal optimal. In the case of the ith slack. B cB -T -T 1 T ( cT N cN ) B B 1 T ( cT N cN ) 0 B B T ( yT N cN ) 0 satisfies the n dual constraints: T or T N T y cN .. Show that the optimal primal objective function value = the dual objective function value for B-T cB . the coefficient of the ith slack is given by cB B1 ei .GT (2003) 8 T 1 Then y = B-T cB (and yT = cB B ) is an optimal solution to the dual and bT y œ cT x. bT y œ bT B-T cB 3. s2 œ 0. 2. yi is the ith element of B-T cB . for i œ -T 1. 3... B is an optimal basis for the primal.the ith element of (cB B1 N cN ) is in general given by 1 ( cT ni ci ) B B where ni is the ith column of N. (The reason for this is that in the initial simplex tableau the coefficient of a slack variable is zero and the corresponding (i th) column of N is ei -the column of the ith slack which is null everywhere except in T T the ith row where it has a unit entry. = BT B-T cB = cB ’ cN “ = ” cN • B y B AT y = c B ã N dT y = ’ N T “ y = ’ NT y “ Since. which is the ith element of y.) Thus. But the dual objective value for the dual feasible solution B cB is which is the required result. we also know that the coefficient of the ith slack variable in 1 x0 + ( cT N cT B B N ) xN œ T 1 cB B b is the ith element (yi ) of y œ B-T cB ..) 2. Thus.. Thus: 1. We now invoke Lemma 2 to establish the Dual Theorem. the original objective function coefficient ci = 0 and ni T œ ei . s3 œ 0 Example 10: FC x Basic Variable x0 5x2 2x2 2x2 x3 s1 10s2 10s3 2s2 8s3 2s2 4s3 œ 280 œ 24 œ 8 x0 s1 x3 œ 280 œ œ 24 8 . yi 0. Let B be an optimal basis and let y œ B-T cB ( y = [y1 . B-T cB that have equal objective values. the coefficient of each variable in the reduced cost must be nonnegative: using y œ B cB Thus. . ^ œ 2. Proof Plan: 1. (For simplicity. Thus. We use the fact that B is primal optimal to show that B-T cB is feasible for the dual. 8 SLACKS: s1 œ 24. we assume that all slacks of the primal problem are non-basic at the optimum solution. 2. Thus y œ B-T cB is dual feasible. ym ]T ). to show that y = B-T cB is dual feasible. We use the fact that B is an optimal basis for the primal. we invoke Lemma 2 to conclude that B-T cB is optimal for the dual and bT x œ cT y.

m ej xj œ 0 j œ 1.. e œ Ô e1 × ã . c − ‘n A − ‘m‚n b − ‘m Let S − ‘ be the slack variables for ÐPÑ m Ô s1 × Ös Ù i. 2. n From Ð2Ñ and Ð3Ñ it follows that ith primal slack 0 Ê ith dual œ 0 ith dual 0 Ê ith primal slack œ 0 jth dual slack 0 Ê jth primal œ 0 jth primal 0 Ê jth dual slack œ 0 (4) (5) (6) (7) (2) (3) Ð4Ñ and Ð6Ñ Ê if a constraint in either primal or dual is not satisfied as an equality Ðhas either si 0 or ej 0Ñ then the corresponding variable of the other Ðor complementaryÑ problem must equal zero. Then x is primal optimal and y is dual optimal iff si yi œ 0 i œ 1. s œ Ö 2 Ù.. Õ en Ø Let e − ‘n be the dual slack variables.. x 0 › P: x. . Example 11: FC x œ Ô x1 œ 2 × x2 œ 0 Õ x3 œ 8 Ø s1 œ 48 Ð8Ð2Ñ 6Ð0Ñ s œ s2 œ 20 Ð4Ð2Ñ 2Ð0Ñ s3 œ 1 Ð8ÑÑ œ 24 0 0 1 † 5Ð8ÑÑ œ 8 Ð2Ð2Ñ 1 † 5Ð0Ñ 0 † 5Ð8ÑÑ œ e1 = Ð8Ð0Ñ 4Ð10Ñ 2Ð10ÑÑ 60 œ 0 . Õ 10 Ø œ 3.GT (2003) 9 x1 1 † 25x2 † 5s2 1 † 5s3 œ 2 x1 œ 2 We may also compute the dual optimal solution directly: BT cBV Ô 1 2 Õ 8 0 2 4 0 × Ô 0× †5 20 œ 1 † 5 Ø Õ 60 Ø Ô 0× 10 . . ã Õ sm Ø D: min š bT y ¹ AT y c. y 0 › y − ‘m i.e. . . 2.2 COMPLEMENTARY SLACKNESS Relates the primal and dual optimal solutions. THEOREM: COMPLEMENTARY SLACKNESS Let x be feasible to P and y be feasible to D..e. Let max š cT x ¹ A x Ÿ b. Hence complementary slackness.

xw2 .GT (2003) 10 y œ Ô 0× 10 Õ 10 Ø e œ e2 = Ð6Ð0Ñ 2Ð10Ñ 1 † 5Ð10ÑÑ 30 œ 5 e3 = Ð1Ð0Ñ 1 † 5Ð10Ñ 0 † 5Ð10ÑÑ 20 œ 0 s1 y1 œ s2 y2 œ s3 y3 œ 0 e1 x1 œ e2 x2 œ e3 x3 œ 0 s1 0 Ê y1 œ 0 : A positive slack in the wood constraint Ê wood must have zero price. s2 œ 0Ñ. xww 2 0 We can now find the dual. xi 0. we have x2 Ä w xww x2 2 max x0 œ 2x1 xw2 xww 2 w ST x1 x2 xww 2 Ÿ 2 x1 xw2 xww 2 Ÿ 2 2x1 xw2 xww 2 Ÿ 3 w x1 x2 xww 2 Ÿ 1 x1 . the ith dual constraint is an = constraint. This can only occur if we are using all available finishing hours Ði. max x0 min y0 x1 0 x1 x* 2 unrs x2 . This converts into a Ÿ constraint. y2 0 Ê s2 œ 0 : y2 0 Ê an extra finishing hour has some value.e.3 FINDING THE DUAL OF AN LP NOT IN NORMAL FORM Example 12: Maximisation problem max x0 œ 2x1 x2 ST x1 x2 œ 2 2x1 x2 3 x1 x2 Ÿ 1 x1 0 x2 unrestricted To convert to NF: Ð1Ñ Multiply each constraint by 1. Ð1 Ñ Ð2 Ñ (3) Rules for finding the Dual directly If the ith primal constraint is a constraint the ith dual variable must satisfy yi Ÿ 0 The ith primal constraint œ constraint. an extra foot of wood would indeed be worthless. the dual variable yi is unrestricted in sign The ith primal variable is unrestricted. 2x1 x2 3 Ä 2x1 x2 Ÿ 3 Ð2Ñ Replace each œ constraint by two inequality constraints: a Ÿ and a constraint. Then convert the constraint to a Ÿ constraint using Ð1Ñ. 3. x1 x2 œ 2 Ä x 1 x 2 2 Ä x 1 x 2 Ÿ 2 x1 x2 Ÿ 2 Ÿ Two constraints ww w ww Ð3Ñ Replace each unrestricted variable xi by xw i xi where xi . Since slack in the wood constraint means that extra wood will not be used.

GT (2003) 11 _____________________________________ y1 y2 y3 0 1 2 1 1 œ 2* 3* y3 1 1 Ÿ 1 _____________________________________ 2 1 max x0 min y0 x1 0 x2 unrs x1 x2 ______________________________________ y1 unrs y2 Ÿ 0 y3 0 y1 y2 1 2 1 1 œ 2 3 y3 1 1 Ÿ 1 ______________________________________ 2 œ 1 DUAL PROBLEM: min ST y0 œ 2y1 3y2 y3 y1 2y2 y3 2 y1 y2 y3 œ 1 y1 unrs. yi 0. min y0 œ 2yw1 2yww 1 4y2 6y3 ST yw1 yww 1 2y2 y3 2 yw1 yww y3 1 1 y2 y3 1 . y3 0 Example 13: minimisation problem min y0 œ 2y1 4y2 6y3 ST y1 2y2 y3 2 y3 1 y1 y2 y3 œ 1 Ð œ Ñ 2y1 y2 Ÿ 3 ÐŸÑ y1 ÐunrsÑ y2 . y3 0 Ð1Ñ Change each Ÿ constraint to by multiplying Ÿ constraint by 1: 2y1 y2 Ÿ 3 Ä 2y1 y2 3 Ð2Ñ Replace each œ constraint by a constraint and a Ÿ constraint. Then transform Ÿ to a constraint: y2 y3 œ 1 Ä y2 y3 1 y2 y3 Ÿ 1 Ä y2 y3 1 Ÿ Two constraint w ww Ð3Ñ Replace unrs variable yi by ywi yww i . yi . y2 Ÿ 0.

PAPER ÐPÑ AND SCISSORS ÐSRÑ Each of the two players simultaneously utters one of the three words S. Find the value and optimal strategies for this 2p0sg. x3 unrs. Row R Player S P Sc min __________________________________________________ S 0 1 1 1 P 1 0 1 1 C Player SC 1 1 0 1 __________________________________________________ Col. Rules for finding the Dual Ðfor min problemÑ directly max x0 min y0 x1 0 x2 0 x3 unrs x4 Ÿ 0 x1 x2 x3 x4 _____________________________________________________________ y1 unrs* y2 0 y3 0 y1 y2 y3 1 2 1 1 0 -1 0 1 1 2 1 0 œ 2 Ÿ 4 Ÿ 6 _____________________________________________________________ 2 1 = 1* Ÿ 3* Ð1 Ñ Ð2 Ñ Ð3 Ñ If the ith primal constraint is a Ÿ constraint. the dual variable xi must satisfy xi Ÿ 0 If the ith primal c. If the ith primal variable yi is unrestricted. S defeats ÐbreaksÑ SC. P defeats ÐcoversÑ S. P or SC. Example 14: STONE ÐSÑ. 3x4 2 x4 x3 x4 x2 x3 0. yww . the game is a draw. x2 œ 2 Ÿ 4 Ÿ 6 0 4. y . y 0 2 3 1 now find the dual. the ith dual constraint is an equality constraint. is an œ constraint. Let . LINEAR PROGRAMMING & ZERO SUM GAMES LP can be used to find the value and optimal strategies Ðfor the row ÐRÑ and column ÐCÑ playersÑ for any 2p0sg. max 1 1 1 The game does not have a saddle point. the dual variable xi will be unrs. If both players utter the same word. x4 Ÿ x2 x2 + x3 max x0 œ 2x1 ST x1 2x1 x1 x1 . Otherwise one player wins £1 from the other according to the rules: SC defeats ÐcutsÑ P.GT (2003) 12 y2 y3 1 2yw1 2yww y2 3 1 yw1 .

The Column Player's LP v in optimal solution is R's “floor". For each of R's strategies we may compute R's expected reward if C has chosen Ðy1 . x2 . x0 œ v v Ÿ x2 x3 ÐStone ConstraintÑ ÐPaper ConstraintÑ ÐScissors ConstraintÑ x1 . y3 Ñ. x1 x3 . x2 . y3 Ñ. observe that for any values of x1 . x3 Ñ then R's expected reward against each of the C's strategies R's expected reward if C Chooses R chooses Ðx1 . C will choose a strategy that makes R's expected reward equal to min Ðx2 x3 .T. x1 x2 . y2 . v unrestricted v Ÿ x1 x3 v Ÿ x1 x2 x1 x2 x3 œ 1 v. x2 . y2 . x2 . Row Player's expected reward R. x1 x2 Ñ is just that largest number Ðsay vÑ that is simultaneously less than or equal to x2 x3 . x3 the largest value of min Ðx2 x3 . x1 x3 . y2 . x3 Ñ _____________________________________________ S P SC x2 x3 x1 x3 x1 x2 By the basic assumption. R's expected reward is at least Suppose that C has chosen the mixed strategy Ðy1 . probabilities x1 . Also. x3 Ñ to make min Ðx2 x3 . x2 . x3 0 and x1 x2 x3 œ 1 Max S. x1 x2 Ñ as large as possible. No matter what strategy played by C. x3 0. x2 . x1 x2 Ñ Then R should choose Ðx1 .GT (2003) 13 x1 œ probability that row player chooses S x2 œ probability that row player chooses P x3 œ probability that row player chooses Sc y1 œ probability that column player chooses S y2 œ probability that column player chooses P y3 œ probability that column player chooses Sc The Row Player's LP are: If R chooses the mixed strategy Ðx1 . To obtain an LP formulation ÐR's LPÑ that will yield R's optimal strategy. Chooses if C chooses (y1 . x1 x3 . y3 ) __________________________________________ S y2 y3 . x1 x3 .

y1 y3 .T. y2 . y3 Ñ to make max ( y2 y3 . y2 . y3 Ñ. S. x3 Ÿ Ÿ Ÿ x3 œ 0 v unrestricted. ! yi œ 1 For a mixed strategy: 3 iœ 1 min y0 œ w . y1 y3 . To obtain an LP formulation. y3 0 w unrestricted. y1 y3 . y1 y2 Ñ Thus. w y1 y2 y1 . y1 y2 Ñ will equal to the smallest number that is simultaneously greater than or equal to y2 y3 . y1 y2 Ðcall this number wÑ. x2 . y2 . y3 Ñ that solves the LP. y1 y3 . C should choose Ðy1 . observe that for any choice of Ðy1 . y2 . y1 y2 Ñ as small as possible. C can ensure that C's expected losses will be at most w Ðwhatever R doesÑ. y3 Ñ max Ð y2 y3 . R will choose a strategy to ensure that R obtains an expected reward of max Ð y2 y3 . y3 and w respectively max min x1 x2 x3 v ___________________________________________________ y1 Ð 0Ñ y2 Ð 0Ñ y3 Ð 0Ñ 0 1 1 1 0 1 1 1 0 1 1 1 Ÿ 0 Ÿ 0 Ÿ 0 w ÐunrsÑ 1 1 1 0 œ 1 ___________________________________________________ 0 0 0 . w y1 y3 . Also y1 .GT (2003) 14 P SC y1 y3 y1 y2 Since R is assumed to know Ðy1 . y2 . w y2 y3 y1 y2 y3 œ 1 observe that w is a “ceiling" on C's expected losses Ðor on R's expected rewardÑ because by choosing a mixed strategy Ðy1 . THE RELATION BETWEEN THE LP'S OF R AND C It is easy to show that C's LP is the DUAL of the R's LP. x3 x3 v v v 0 0 0 1 Let the duals be y1 . y3 0. y2 . y2 . R's LP: max ST x0 œ v x2 x1 x1 x2 x1 x2 x1 .

. Suppose that you add a constant c to every element of A so that all coefficients become nonnegative: max ST x0 œ v´ v´ Ÿ Ða11 cÑx1 Ða21 cÑx2 .... xmÑ Thus.GT (2003) 15 We read R's LP across the above table and the dual of R's LP is obtained by reading down each column... (amn + c) xm (or v œ v´ c) . . For the Stone Paper Scissors game: R's LP Ê w œ 0 x1 œ " $ ... Ðamn cÑxm x1 x2 . y3 0 y3 w 0 y3 w 0 w 0 œ 1 y3 w unrestricted which is C's LP. Define v´ such that v´ œ v c Then the above max problem can be written as max x0 œ v´ v´ Ÿ (a11 + c) x1 . am1 xm c Ðx1 x2 ... x2 œ " $ .. It can be shown that the optimal strategies obtained via LP represent a stable equilibrium: neither player can improve by a unilateral change in strategy. v´ c Ÿ a11 x1 a21 x2 . xm 0 v´ ? Note: Ðx1 x2 . . y3 œ " $ The fact that v and w are unrestricted can be overcome. y2 œ " $ . the dual can be read down as min y0 œ w ST y2 y1 y1 y2 y2 y1 y1 . Ðam2 cÑxm ã v´ Ÿ Ða1n cÑx1 Ða2n cÑx2 . xm Ñ œ 1 v´ Ÿ a11 x1 a21 x2 . am1 xm The same holds for all the constraints.. x2 . C's LP Ê v œ 0 y1 œ " $ ... The common value v and w is the VALUE of the game to R. (am1 + c) xm ã v´ Ÿ (a1n + c) x1 . xm œ 1 x1 . Ðam1 cÑxm v´ Ÿ Ða12 cÑx1 Ða22 cÑx2 . y2 .. x3 œ " $.. Thus..... The DUAL THEOREM Ê v œ w ÐBoth LP's feasible and boundedÑ R's “floor" œ C's “ceiling" This is known as the minmax theorem. Recall that the dual constraint corresponding to v will be an œ constraint Ðas v is unrsÑ and the dual variable corresponding to the primal x1 x2 x3 œ 1 will be unrs...

. the person who confesses will go free while the person who does not confess will surely be convicted to a 20 year jail sentence.. The value v* of the original game: v* œ v*´ c and w* œ w*´ c. v´ 0. If both of you confess you will both be convicted and sent to prison for 5 years Ðbecause of lack of evidenceÑ. amn yn c Ð! yi Ñ i œ1 ! yi œ 1 n i yi 0 w´ 0. the strategies and rewards are: . Although both are guilty. 5. Thus after adding c we may assume that v´ and w´ 0 and ignore v and w unrs. the police chief tells each prisoner: “If only one of you confesses and testifies against the other ÐpartnerÑ.. Since after adding c the matrix will not have any negative elements (rewards) v*´ . TWO-PERSON NON CONSTANT-SUM GAMES Example 15: Prisoner's Dilemma Two prisoners who escaped and participated in a robbery have been recaptured and are awaiting trial for their new crime. the police chief is not sure whether he has enough evidence to convict them. Let v*´ . a1n yn c Ð! yi Ñ n i œ1 n min yo œ w´ ã w´ am1 y1 . What should the prisoners do? If the prisoners cannot communicate with each other. v we therefore have to do the following: add c œ | most negative element of matrix A | to all elements of A. To introduce 0 constraints for w. the solution of the primal LP does not change by adding c to every element of A and v´ œ v c. . Let v* and w* be the optimal values of the original game Ðc œ 0Ñ. In order to entice them to testify against each other.GT (2003) 16 ! xi œ 1 m i xi 0 i œ 1 . w*´ 0 hold. I shall convict you of a misdemeanour and you each will get 1 year in prison".. Thus... The dual solution also does not change and w´ œ w c. If neither confess.. m. The same can be shown for the dual w´ a11 y1 . w*´ be the optimal values when c is given as above.

RÑ not to be an equilibrium point requires T R (each player has temptation to double-crossÑ. however. If both double cross. If each prisoner follows his undominated “confess" strategy. If the players are cooperating Ðboth don't confessÑ each player can gain by double-crossing. PÑ is an equilibrium point. 1Ñ __________________________________________________ Å Å P 1's P2's reward reward Note that sum of rewards in each cell varies from 2 Ð œ 1 1Ñ to 20Ð œ 20 0Ñ Ê This is not a constant sum game. if each prisoner chooses the dominated “don't confess" strategy. RÑ In this game ÐP. This cannot occur in a constant sum game. TÑ ÐT. 20Ñ Don't Confess Ð 20. Example 16: Arms Race . each prisoner will get 5 years. Definition: As in a 2p0sg a choice of strategy by each player ÐprisonerÑ is an EQUILIBRIUM POINT if neither player can benefit by a unilateral change in strategy.GT (2003) 17 Prisoner 2 Prisoner 1 Confess Don't Confess __________________________________________________ Confess Ð 5. A game is reasonable if R P. For ÐR. 5Ñ Ð0. Does any strategy “dominate" the other? For each prisoner. they will be worse off than the cooperative strategy Ðdon't confessÑ. This requires P S. However this is not an equilibrium point because if we are at Ð 1. NC œ non-cooperative action C œ cooperative action P œ punishment for non cooperation S œ payoff to person who is double-crossed R œ reward for cooperating if both players cooperate T œ temptation to double-cross the opponent P2 P1 NC C _______________________________ NC C ÐP. each prisoner will get 1 year. 1Ñ. Thus. Ð 5. 5Ñ is an equilibrium point: either prisoner deviating from this decreases his reward to 20 Ðfrom 5Ñ. SÑ ÐR. “confess" dominates “don't confess" strategy. 0Ñ Ð 1. However each is better off at Ð 1. Thus. if each prisoner chooses his dominated strategy they are better off than if each chooses his undominted strategy. On the other hand. PÑ ÐS. a prisoner's dilemma requires: T R P S This example is of interest because it explains why two adversaries often fail to cooperate with each other. 1Ñ each prisoner can increase his reward from 1 to 0 by changing from Don't Confess to Confess Ðeach can benefit from double-crossing his opponentÑ.

Find the equilibrium point(s). . Each have two strategies: develop a new missile (DÑ or maintain the status-quo. 10Ñ Ðboth nations non-cooperativeÑ is an equilibrium point. Ð 10. the conquering nation earns a reward of 20 and the conquered looses 100 units. 10Ñ Ð10. This example shows how maintaining the balance of power may lead to an arms race. 0Ñ is not stable. 10Ñ. we see that in this situation each nation will gain from a double cross. the nation with the new missile will conquer the other nations. The reward matrix is based on the assumption that if only one nation develops a new missile. SU US D M ___________________________________________ D Ð 10. Assume cost of developing a missile 10 units.GT (2003) 18 US and SU are engaged in an arms race. Although Ð0. M : Cooperative. Ð0. 0Ñ leaves both nations better off than Ð 10. Thus. 0Ñ ___________________________________________ D : Non-cooperative. 10Ñ Ð0. 100Ñ M Ð 100. In this case.

3}) = 0 (any other coalition value is the maximum value a member of the coalition places on the land v({1. N-PERSON GAME THEORY In many situations. The latter follows because if all players are in S. 2. respectively. 4}) = 4 (if ± S ± œ 4). The characteristic function of this game is given by v({ }) = v({1}) = v({2}) = v({3}) = v({2. Example 19: The land development game Player 1 owns a piece of land and values it at £10000. 000. We therefore consider games with three or more players. 3. n} be the set of players in an n-person game which is specified by the game's characteristic function. 3}) = 0 v({1. 000. the coalition receives a reward of -b. The best that the members of any coalition can do is to dump all the rubbish on the property of owners not in S. 3}) = £30000 v({1. Consider any subset of player sets A & B: A B = g. Example 17: The drug game Dr Medicine (player 1) has invented a new drug. 2. If b bags of rubbish are dumped on a coalition of property owners. The chosen company will split a £1 million with Dr Medicine. There are no other prospective buyers. 2. Let N = {1. . Thus.. 2}) = v({1. Definition: Reward Vector Let Ô x1 × Öx Ù x= Ö 2Ù ã Õ xn Ø be a vector such that player i receives reward xi . v(S) can be determined by calculating the amount that members of S can get without help from players who are not in S. Solution concepts for n-person games are related to the reward that each player will receive. This is the reward vector.. For any n-person game the characteristic function satisfies superadditivity v(A B) v(A) + v(B). the characteristic function of the garbage game ( ± S ± is the number of players in S) is given by v({S}) = (4 ± S ± ) (if ± S ± 4) v({1. the characteristic function v of a game gives the amount v(S) that the members of S can be sure of receiving if they act together and form a coalition. Example 18: The rubbish dumping game Each property owner has one bag of rubbish and has to dump his/her bag on somebody's property. Thus. (any coalition that does not contain 1): v({ }) = v({2}) = v({3}) = v({2. If the players A B band together. 000. Players 2 and 3 are developers who can develop the land and increase its worth to £20000 and £30000. This would result in the coalition receiving an amount v(A) + v(B). Thus. 2}) = £20000 v({1. there are more than two competitors. 3}) = £30. they must dump their garbage on members of S. v(A B) must be at least as large as v(A) + v(B). 3}) = £1. 3}) = v({1.GT (2003) 19 6. one (but not the only) option is to let the players in A fend for themselves and players in B fend for themselves.. Definition: For each subset S of N. A reward vector x (xi is the ith element of x) is not a reasonable candidate for a solution unless it satisfies . Dr Medicine cannot manufacture this drug on his/her own but can sell the formula to company (player) 2 or company (player) 3.

To show y {1. y = [. v({1. . y gives players in {1. . Example 22: For the land development game in Example 19. yn ]T dominates x through a coalition S. Definition: If x satisfies both GR and IR. 3}) = . THEOREM: DETERMINATION OF THE CORE . 3} a total of . any reasonable reward vector must give all the players an amount that equals the amount that can be attained by the supercoalition consisting of all players. The following two examples illustrate domination.2}) = 0. ... £10000. £19000.2. 3}).2. . Next. xn ]T . 3}) = 1 Let x = [. £10000) (£5000. . £10100]T . note that both x and y are imputations. 3}) = 0. £11000) THE CORE OF AN n-PERSON GAME Definition: Given an imputation x = [x1 .e. Definition: The core of an n-person game is the set of all undominated imputations. y S x. £100.1. Since . x (£10000.. it is reasonable to assume that players 1 and 3 can band together and receive a total reward of . we need only observe that players 1 and 3 from y (£29900) does not exceed v({1. John von Neumann and Oskar Morgenstern argued that a reasonable solution concept for an n-person game was the set of all undominated imputations. Example 21: Consider a three person game with the following characteristic function: v({ }) = v({1}) = v({2}) = v({3}) = 0 v({1. £2000. because the players in S can object to the rewards given by x and enforce their objection by banding together and thereby receiving the rewards given by y (since the members of S can surely receive an amount equal to v(s)). player i must receive a reward at least as large as what he can get for himself (v({i})). i.80..2.2 does not exceed v({1. To show y {1. Thus.10]T . 2. x2 . If x were proposed as a solution. observe that with y.e.GT (2003) 20 v(N) = ! xi (Group Rationality: GR) n i=1 i. v({1. 3}) = 0.05]T . then both of the following must be true: Ä Each member of S prefers y to x Ä Since ! yi Ÿ v(S). The important point is that x cannot occur because players 1 and 3 would not allow it. Also. Any solution concept for n-person games chooses some subset of the set of imputations (possibly empty) as the solution to the n-person game. let x = [£19000. £100. and xi v({i}) (for each i − N) (Individual Rationality: IR) i.1 + . £5000) (£12000. 3} x. -£1000) (£11000. then x should not be considered as a possible solution to the game. y Sx is x an imputation? Yes No: x1 Ÿ v({1}). £10100]T .2. we call that x an imputation. so IR violated No: IR violated No: GR violated if ! yi Ÿ v(S) and for all i − S. £11000. player 1 would sell the land to player 3 and y (or some other imputation that dominates x) would result.90. yi xi i −S If y S x. y2 .2.. we say that the imputation y = [y1 .05. Example 20: Consider the payoff vectors of Example 19. players 1 and 3 will never allow the rewards given by x to occur. v({2.10.1 = .. 3} x. the members of S can attain the rewards given by y i −S Thus. players 1 and 3 receive more than they receive with x. .e. y = [£19800.

xn ]T is in the core of an n-person game if and only if for each subset (coalition) S of N we have ! xi v(S). Introduction to Game Theory. then all the above inequalities must hold simultaneously. 2. 2} x. £0). for example. x1 + x3 = £1000000. x3 satisfy the above conditions and x1 + x2 £1000000 . Example 23: The drug game (continued). x2 . x1 + x2 + x4 -1 . -1. we find that this contradicts the equality x1 + x2 + x3 + x4 = -4. x2 . x1 + x2 + x3 £1000000 If x = [x1 . Any x in the core must also satisfy x3 £0 and x1 Ÿ £30000 and any x satisfying x2 = £0 and x1 + x3 = £30000 in the core. Players 1 and 2 could. x = [x1 . which treats players 1. x2 . x3 ]T will be in the core iff x1 . if we choose an imputation that is not in the core. x1 + x2 + x3 £30000 Considering x1 + x3 £30000 and x1 + x2 + x3 = £30000 leads to x2 = £0 and x1 + x3 = £30000. Hence. x3 . Springer. y {1. x3 . -1. x3 satisfy the above conditions and x1 + x2 £20000 . To understand the reason for the empty core. £(30000 . x2 . x = [x1 . Morris. player 3 receives 30000-x1 and player 2 receives nothing. -1]T . we can show how it is dominated. x3 ]T is an imputation iff x1 £10000 . Example 24: The Garbage Game (Continued). x4 -3 . x3 ]T is in the core.. the core of the garbage game is empty. x3 0 .GT (2003) 21 An imputation x = [x1 . P. Then player 1 receives x1 . x3 -3 . x1 + x3 £30000 . This would give player 1 less than £1000000 and players 2 and 3 some money. i −S (for the proof see.. Example 25: The Land Development Game (Continued). x1 = £1000000 and x2 = x3 = £0. we find that 3(x1 + x2 + x3 + x4 ) -4. x1 + x2 + x3 = £30000 The theorem shows that x = [x1 . x2 . 2} x. x1 + x2 + x3 + x4 = -4 Applying the theorem to all three-player coalitions. 0]T . For this example. As x1 + x2 + x3 = £1000000. Thus. In a similar fashion any imputation could be dominated by another imputation. x2 0 . £0). x = [x1 . 1. x2 . The two player version of the game has a core consisting of (-1. By x1 + x2 £20000. consider the imputation x = [-2. x2 + x3 £0 . x2 .. £75000. -1) and for n 2. join together to ensure the imputation x = [-1. for example.5. £50000). we must also have x1 + x2 = £1000000 . x2 + x3 + x4 -1 Adding these four inequalities. the core contains an infinite number of points. Hence. x1 + x3 + x4 -1 . x4 ]T to be in the core. . then y {1. An alternative solution concept is the Shapley value. We see that this satisfies all the inequalities and thus the core of the game is the imputation x = (£1000000. £50000. for x = [x1 . x2 + x3 £0. We consider the core of the three games discussed so far. x3 ]T will be in the core iff x1 . the core is empty. Thus. In this example. x2 £0 . x2 -3 . 3 unfairly. we have x1 £20000. If we let y = (£925000. The core emphasises the importance of player 1. 1994) This theorem states that an imputation x is in the core (ie x is undominated) iff for every coalition S the total of the rewards for each player in S (according to x) is at least as large as v(S). THE SHAPLEY VALUE . Consider imputation x = (£900000. x3 £0 . -. no imputation can satisfy the conditions for the core.x1 )) will be in the core. x2 . x1 + x3 £1000000 . if £20000 Ÿ x1 Ÿ £30000 and any vector of the form (x1 . x1 + x2 + x3 = £1000000 The theorem shows that x = [x1 . Player 3 outbids player 2 and purchases the land from player 1 for price x1 (£20000 Ÿ x1 Ÿ £30000). x2 . x2 . x4 ]T is an imputation iff x1 -3 . £0.5. x3 ]T will be an imputation iff x1 0 . it is necessary to satisfy x1 + x2 + x3 -1 . Thus. £0.

and v({3}) = 15.. (2)(1) Players not in S{i} arrive = ± S ± !(n ± S ± 1)! Since there are a total of n! permutations of 1.. there is a unique reward vector x = [x1 . The probability that when player i arrives the players in the coalition S are present is pn (S). Kuhn and A.. . n i=1 Axiom 3 If v(S {i}) = v(S) holds for all coalitions S. the Shapley value vector for the game (v + v) The validity of Axiom 4 is often questioned: adding up rewards from two different games may be like adding up apples and oranges. Then the formulae imply that player i's reward should be the expected amount that player i adds to the coalition made up of the players who are present when player i arrives. Axiom 2 Group rationality. the players in the subset S will be present. n! = n (n 1)(n 2).. n that result in player i's arriving when players in the coalition S are present is given by ±S±(±S±1)(±S± 2) . Suppose the Shapley value of a three person game is x = [10. NJ. G. xn ]T satisfying axioms 1-4. 15. 15. then the Shapley value has xi = 0... Then. n has a 1/(n!) chance of being in which the players arrive. xn ]T satisfying the following axioms: Axiom 1 Relabelling of players interchanges the players' rewards.. That is. Shapley “Quota Solutions of n-Person Games" in: Contributions to the Theory of Games... player adds v(S {i}) v(S) to coalition S. To derive the formula for We now show that pn (S) given above is the probability that when player i arrives. (see: L. If player i adds no value to any coalition.GT (2003) 22 In the drug game.. n. Princeton University Press. 10]T . Observe that the number of permutations of 1. however. the probability that player i will arrive and see the players in S is . 1982). she finds players in the set S have already arrived.. if originally v({1}) = 10. x2 .. . x2 . Princeton. Suppose players 1. 2... we found that the core of the drug gave all the benefits or rewards to the game's most important player (the inventor of the drug). . n arrive in a random order. eds. Although the above formulae seem complex.... for any of the n! permutations of 1. Shapley proved: THEOREM (SHAPLEY VALUE) Given any n-person game with characteristic function v. 2. If we interchange the roles of player 1 and 3 (for example. For any characteristic function. .. For example. ^ is x + y. we would make v({1}) = 15 and v({3}) = 10) then the Shapley value for the new game would be x = [20.. (2)(1) S arrives (1) i arrives (n ±S± 1)(n ±S± 2).. ! xi = v(N). 2. Lloyd Shapley showed that there is a unique reward vector x = [x1 . . 1953 and Owen. 20]T . “Game Theory" Academic Press.. if n = 3. they have a simple interpretation. The reward to the ith player (xi ) is given by xi = ! pn (S) cv (S {i}) v(s) d . ^ Axiom 4 Let x be the Shapley value vector of game v and let y be the Shapley value vector for game v. H.(2)(1) (with 0x = 1). 2.. player i receives reward 0 from the Shapley value. An alternative concept for n-person games is the Shapley value which in general gives more equitable solutions than the core does.. If Axioms 1-4 are assumed to be valid.. If player i forms a coalition with the players who are present when he arrives. Florida. . pn (S) = ±S±! (n ±S± 1)! n! all S for which is not in S where ± S ± is the number of players in coalition S and for n 1. Tucker. then there is a 1/3! = 1/6 probability that the players will arrive in any one of the following sequences 1 1 2 2 3 1 3 2 3 2 3 3 3 1 2 1 2 1 Suppose that when player i arrives.

Moreover. add the results. 1 Since each of the six orderings are equally likely. Example 26 : The drug game (continued). 1. For each such coalition. and (n ± S ± 1)! permutations coming after. we obtain a new ordering in which xi is again in the kth position. In this example. . Thus. weight each one with probability 1/(n!) of that ordering occurring. With S being the set of players in this ordering. Thus. There are ± S ± ! permutations of the players coming before. Among the n! terms in the sum which defines xi there are many duplications.GT (2003) 23 ±S±! (n ±S± 1)! n! = pn (S) . 3. 3} Thus. 2. 3 2. the reward player 1 should receive. 2. Indeed. 6 Shapley value is essentially computed using the fact that player i should receive the expected amount that she adds to the coalition present when she arrives. For x1 . 3} {3} Since player 1 adds on the average (2/6)(0) + (1/6)(1000000) + 2/6(1000000) + (1/6)(1000000) = £ 4000000 6 the Shapley value concept recommends that player 1 receives a reward of £ 4000000 . we compute v(S {i}) v(S) and p3 (S): S {} {2} {2. we list all coalitions S in which player 1 is not a member. we find that the Shapley values are given as above. the definition of xi is this: calculate [v (S {i}) v(s) ] for each of the n! possible orderings of the players. we require the information: S {} {1} {3} {1. 3 1. there are ± S ± !(n ± S ± 1)! orderings. 6 Since. The Shapley value. given that any ordering has probability 1/(n!) and associated with [v (S {i}) v(s)]. To compute the Shapley 6 value for player 2. {i}. 2 3. 3}) = £1000000 to the players. the Shapley value must allocate a total of v({1. the term [v (S {i}) v(s)] occurs ± S ± !(n ± S ± 1)! times. the term [v (S {i}) v(s)] is the same. the Shapley value for player 2 is p3 (S) 2/6 1/6 1/6 2/6 v(S {2}) v(S) £0 £1000000 £0 £0 p3 (S) 2/6 1/6 2/6 1/6 v(S {1}) v(S) £0 £1000000 £1000000 £1000000 (1/6)(£1000000) = £ 1000000 . 3. this method yields the computation below ORDER OF ARRIVAL AMOUNT (£) ADDED BY PLAYER'S ARRIVAL Player 1 Player 2 Player 3 0 0 1000000 1000000 1000000 1000000 1000000 0 0 0 0 0 0 1000000 0 0 0 0 1. the Shapley value for player 3 is £1000000 x1 x2 = £ 1000000 . This explains the probability pn (S). 1. 2. 1 3. if we permute the part of the ordering coming before {i} and the part coming after it. 2 2. suppose that we have an ordering in which {i} occurs at position k. for both the original and the permitted orderings.

03.03 and each 747 pay £(13. 3}) = v({1. 2 3.1963) = 98% of the power in the Security Council resides with the permanent members. Littlechild and G. we assume that the three planes land in a random order and we determine how much cost (on the average) each plane adds to the cost incurred by the planes that are already present: COST ADDED BY PLAYER'S ARRIVAL(£) ORDER OF PROBABILITY Player 1 Player 2 Player 3 ARRIVAL OF ORDER 1. 3}) = -£400 To find the Shapley value (cost) to each player. 3. 2. For this characteristic function. each DC10 pay £(5. Thus.001865) = 1. For simplicity. 3}) = v({1. 2. the UN Security Council consists of five permanent members (who have the veto power over any resolution) and ten nonpermanent members. even if more than one plane of each type lands. the characteristic function for this game (cost Ê negative revenue) would be v({ }) = 0. 3 1. the airport will have a 400-yd runway.001865 and 5 ‚ (. Example 27: Suppose three types of planes use an airport. it must receive at least nine votes. 2 2. Suppose the cost in pounds of maintaining a runway for a year is equal to the length of the runway. Owen (1973). . A Piper Cub (player 1) requires a 100-yd runway. 2}) = v({2}) = -£150 v({3}) = v({2. 20. it has been shown that the Shapley value allocates runway operating cost as follows: all planes that use a portion of the runway should divide the cost of that portion of the runway (S. the Shapley value concept would recommend that each Piper Cub pay £100/(10 + 5 + 2)= £5.1963 and the Shapley value for each nonpermanent member is . suppose that each year only one plane lands at the airport. 1 3. Thus.1963) + 10 ‚ (.GT (2003) 24 The Shapley value can be used as a measure of the power of individual members of a political or business organisation. five DC10 and two 747 landings. the DC10's and 747's should pay the next 150 100=50 yd of runway and the 747's should pay the last 400 150 = 250 yd of runway. Thus. and a value 0 to all those that cannot defines a characteristic function. 370-372). 1. v({1. 2. How much of the £400 annual maintenance cost should be charged to each plane? We define a three-player game in which the value to a coalition is the cost associated with the runway length needed to service the largest plane in the coalition. If there were ten Piper Cub. 3. 1. the Shapley value indicates that 5 ‚ (. For example.88 + (150 100)/(5 + 2)) = £13. 1 Player 1 cost = Player 2 cost = Player 3 cost = 1/6 1/6 1/6 1/6 1/6 1/6 100 100 0 0 0 0 50 0 150 150 0 0 250 300 250 250 400 400 (1/6)(100 + 100) = £200/6 (1/6)(50 + 150 + 150) = £350/6 (1/6)(250 + 300 + 250 + 250 + 400 + 400) = £1850/6 In general. since 747's land at the airport. a DC-10 (player 2) requires a 150-yd runway and a 747 (player 3) requires a 400 yd runway. Vol.88. including the votes of all permanent members.03 + (400 150)/2)=£138. v({1}) = -£100. it can be shown that the Shapley value for each permanent member is . “A simple expression for the Shapley value in a special case" Management Science. 3 2. For a resolution to pass the Security Council. all planes should cover the cost of the first 100 yd of runway. Assigning a value 1 to all coalitions that can pass a resolution.

- MB0032 Operations Research
- Add maths assignment
- convexdualitychapter.pdf
- Handout_OPTIMIZATION.pdf
- OR-Unit1
- 1999_Guntman
- Introduction to Management Science a Modeling and Case Studies Approach With Spreadsheets 5th Edition Hillier
- Module 1
- LPP n Primal Dual
- Transportation Implementation
- VI-03 Minimum Variance Method_2007
- Pages From Slide_Optimization
- msgu-06
- Operation Research
- Tutorial to Excel Solver for the Solution of Linear Programming Models
- Z06_REND6289_10_IM_MOD6
- Neutrosophic Goal Programming applied to Bank
- Manuscript Template
- 22Chapter14 (1)
- Operation Research
- Optimal Spinning Reserve Capacity in Power Market.pdf
- Shifter
- EMR024 (5).pdf
- Abstract
- 158_2005_Article_BF01195945
- All12a
- Structural Sizing and Shape Optimisation of a Load Cell
- 9783319283906-c1
- Differential Evolution Algorithm With Self-Adaptive Population Resizing Mechanism
- Ch10

- Share Purchase Agreement
- Document
- Lea l1 Droit
- Mikhail Bakhtin
- The Religion of Cannan
- Droit l2s2
- soufi-mon-amour-elif-shafak.pdf
- emil-cioran-amurgul-gindurilor.pdf
- Mikhail Bakhtin
- the afganistan questions and the reset in US Russian relations.pdf
- Shahid_Rome-and-the-Arabs_Online.pdf
- De l'evolution de la conflictualité à celle des systèmes de force
- La Libre Circulation Des Marchandises Est l
- Alexandru_D._Xenopol_-_Cronologia_rațională_a_istoriei_universale.pdf
- Estime-de-Soi_3edition.pdf
- 46430487-Droit-des-obligations-Corrige-Commentaire-d-arret (1).pdf
- 189714190 Communication Interne
- The Interpreter
- Droit
- Business and Business Growth
- Envie Au Travail Bidaillet
- Două tipuri de
- A-Ber Sich Und Andere Sprechen
- Droit
- 111120423-Capitalismul
- TD MACRO Politiques Economiques L2 2012-2013
- 46430485-Droit-des-obligations-corrige-cas-pratique-l’objet (1).pdf
- 18684732 National Identity Popular Culture and Everyday Life
- Communication
- 8016227 a British Agenda for Europe RIAA

Sign up to vote on this title

UsefulNot usefulClose Dialog## Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

Loading