ABriefIntroductiontotheBasicsof Game Theory

Size: px
Start display at page:

Download "ABriefIntroductiontotheBasicsof Game Theory"

Transcription

1 ABiefIntoductiontotheBasicsof Game Theoy Roy Smead Notheasten Univesity Septembe 18, 2013 Game theoy is a set of mathematical tools that ae used to epesent and study social inteactions. This is a vey bief (and not compehensive) intoduction to some of the basic ideas of game theoy. In many ways, game theoy is an extension of decision theoy which is the study of ational choice when we ae faced with uncetainty. Fo this eason, some of the basic concepts fom decision theoy (e.g. pobability and utility) ae impotant fo undestanding the basics of game theoy. The fist section coves some cental ideas fom decision theoy and the discussion of game theoy begins in section two. 1 Decision Theoy What is the best decision to make when we ae faced with uncetainty about the outcomes? This is the cental question of decision theoy. A decision poblem typically consists of (i) a set of possible states of the wold, (ii) the pobability that each state will occu, (iii) a set of available actions to take and (iv) a value assigned to each act in each state of the wold. 1.1 Pobability A pobability of a state is just a numbe between 0 and 1 that epesents the chance that state will occu. If S is a possible event, P(S) isthepobabilitythat S will occu. In decision theoy, pobabilities ae usually intepeted as subjective degees of confidence that an event will occu. Pobability is govened by a small set of ules (called the Kolmogoov axioms): 1. All pobabilities ae in the [0, 1] inteval: 0 apple P(S) apple 1 2. The pobability of a necessay event N is equal to 1: P(N) =1 1

2 3. The pobability that one of two mutually exclusive events S and R occus is equal to the sum of thei pobabilities: P(S or) = P(S) + P(R) These thee axioms plus a definition fo conditional pobability P(S given R) = P(S and R)/P(R) aesu cient to define all of elementay pobability theoy. Hee ae a few facts of pobability that will be helpful fo ou puposes: S and R ae independent of one anothe if and only if P(S) =P(S given R). If S and R ae independent, P(S and R) =P(S) P(R), othewise P(S and R) =P(S) P(R given S). If S and R ae not mutually exclusive, then P(S o R) =P(S)+P(R) P(S and R). 1.2 Utility Taking a paticula action in a paticula state usually has some consequences. The desiability of those consequences is epesented by anothe numbe called utility. The utility of an outcome can be any eal numbe (positive o negative) and the highe numbe epesents moe desiability. Someone s utility is just a numeical epesentation of thei pefeences, whateve those may be. In othe wods, utility epesents what someone caes about and the degee to which they cae about it. u(x) epesentstheutilityofoutcomex. The utility of a paticula decision acoss all possible states can be epesented in a decision matix. Suppose thee ae two possible actions one might take (A and B) andtheepossibleelevantstatesofthewoldthatmightoccu(s, R, and T ). It is typical to assume that these states ae mutually exclusive and exhaustive, meaning one and only one will occu. We can epesent the decision poblem as follows: Actions States of the Wold S R T A B This means that the most desiable outcome is to choose act B in state S followed by act A in states T, R, ands espectively. Howeve, we can only choose ou actions and cannot choose which state of the wold will occu. So, if you want to 2

3 know what act is the best to do, you need to know the pobability that each state will occu. 1.3 Expected Utility We can combine the pobability that each state will occu with the utilities of choosing an action fo each state. This gives us the expected utility fo each action. Lets take the example above and suppose that the pobabilities fo each state ae as follows: P(S) =0.4, P(R) =0.4 andp(s) =0.2. Given the fo each act-state combination povided above we can calculate the expected utility of each action by combining the pobabilities with the utilities: EU(A) =P(S) u(a in S)+P(R) u(a in R)+P(T ) u(a in T ) = =1.8 EU(B) =P(S) u(b in S)+P(R) u(b in R)+P(T ) u(b in T ) = =2.0 Since utility epesents what one caes about (and the degee that they cae about it), one commonly accepted pinciple of ational decision making is to choose whicheve action geneates the highest expected utility. In this case, action B has highe expected utility than action A. The idea in this example can be extended and expessed vey geneally. If some action A has many possible consequences C 1,C 2,...,C n, then to figue out the expected utility fo that action, you simply conside the utility of each possible consequence multiplied by the pobability of that consequence. The expected utility is the sum of those poducts: EU(A) = X i u(c i )P(C i ). 1.4 Dominance In some special cases, one action may be bette than othe actions no matte what state of the wold we wind up in. When an one act is bette than anothe no matte what happens, this action is said to be dominant ove the othe. Thee ae two foms of dominance: weak and stong. An act stongly dominates anothe if and only if it leads to highe utility no matte what happens. An act weakly dominates anothe if and only if it does at least as good as the othe no matte what happens and does bette in at least one instance. Conside the following example: 3

4 Actions States of the Wold S R T A B C In this case, we can see that act A does bette than act C no matte if we end up in state S, R o T. So, act A stictly dominates act C. Also, act A weakly dominates act B since it does bette when S occus and equally well when R o T occus. Recognizing that one action dominates anothe can be helpful if we don t know the pobabilities of the states. If an act dominates anothe, it doesn t matte what the pobabilities of the states ae, we will do at least as well o bette by choosing the dominate action. In the example hee, it would be iational to choose any action othe than A since it dominates both othe actions. 2 Game Theoy A game is essentially a multi-peson decision poblem whee the outcome depends, not on some extenal state of the wold, but on the actions of all the playes. A game consists of a set of playes, a set of stategies (actions) fo each playe and apayo function fo each playe which specifies the utility that playe i gets fo evey possible combination of stategy choices by the playes. The most well known game is pobably the Pisone s Dilemma (shown below). Playe 2 Playe 1 Coopeate Defect Coopeate 2,2 0,3 Defect 3,0 1,1 Pisone s Dilemma In this game thee ae two playes {Playe 1, Playe 2}. Playe 1 s stategy set is {Coopeate, Defect} and Playe 2 s stategy set is {Coopeate, Defect}. The payo s fo evey combination of stategies is given in the game matix above. The left numbe in each cell is the payo to Playe 1, the ight numbe in each cell is the payo to Playe 2. Hee ae some othe impotant tems and ideas fo studying games: 4

5 When a game is expessed in a matix-fom as above it is called then Nomal Fom o the Stategic Fom of a game. An outcome of a game can be specified by a list of eveyones stategy choices: (Playe 1 s choice, Playe 2 s choice). Such a list is called a Stategy Pofile. A best esponse to othe playes stategies is a stategy that geneates the highest payo given what the othe playes ae doing. Most games we will conside will have just two playes, like the Pisone s Dilemma. And, in many cases, the playes playes will have the same stategies available to them and have symmetic payo functions. These games ae called two-playe symmetic games. It is possible to have moe than two playes and fo di eent playes to have di eent stategies available to them. We will conside these moe complicated games late on, but fist we will focus on two-playe symmetic games to intoduce the cental solution concept fo game theoy: the Nash equilibium. 3 Nash Equilibium If we examine the Pisone s Dilemma game, we will notice that Playe 1 does bette by playing Defect no matte what Playe 2 does (i.e. if Playe 2 coopeates, Playe 1 gets 3 by playing Defect and 2 by playing Coopeate and if Playe 2 defects, Playe 1 gets 1 by playing Defect and 0 by playing Coopeate ). And likewise, Playe 2 does bette by playing Defect no matte what Playe 1 does. In othe wods, the stategy Defect stictly dominates the stategy Coopeate fo both playes. Since choosing dominant actions ove non-dominant actions is consideed the ational choice, the ational solution to the Pisone s Dilemma is fo both playes to choose Defect. In othe wods the stategy pofile (Defect, Defect) is the unique ational solution to the Pisone s Dilemma. Playe 2 Playe 1 Right Left Right 1,1 0,0 Left 0,0 1,1 The Diving Game In many games, howeve, easoning fom dominance does not poduce a solution. Conside the Diving game (above). In this game each playe has two 5

6 stategies: Right and Left. But, neithe stategy dominates the othe since the best thing to do depends on what the othe playe does. If the Playe 2 dives on the Right, Playe one wants to dive on the Right. And, if Playe 2 dives on the Left, Playe 1 wants to dive on the Left. So, neithe stategy dominates the othe. The idea of a Nash equilibium will allow us to find the solutions to games whee the idea of dominance doesn t help. Playes ae at a Nash equilibium of agameeachplaye sstategyisabestesponsetothestategiesofalltheothe playes. This idea can be captued as follows: Nash Equilibium: A stategy pofile is a Nash equilibium if and only if no playe could do bette by choosing a di eent stategy given what the othe playes ae doing. In a two playe game, a stategy pofile (s, ) isapue stategy Nash equilibium i u 1 (s, ) u i (s 0,) fo all s 0 in Playe 1 s stategy set AND u 2 (s, ) u 2 (s, 0 )foall 0 in Playe 2 s stategy set. In the Pisone s Dilemma, thee is only one Nash equilibium: (Defect, Defect). In the Diving Game, thee ae at least two Nash equilibia: (Right, Right) and (Left, Left) in eithe of those states, neithe playe wants to change given what the othe playe is doing. Thee is a thid Nash equilibium in the Diving Game as well, but this thid equilibium is not a pue stategy equilibium and involves what ae called mixed stategies (see the next section). The idea of a Nash equilibium also woks in asymmetic games, when playes might have di eent stategies and di eent payo functions. Conside the game below whee the two playes have di eent stategies and the payo functions ae also not symmetic. Playe 2 Playe 1 a b c d 2,1 0,2 2,5 1,0 s 4,0 2,1 1,1 1,1 t 3,3 2,0 0,0 1,3 Example asymmetic game To find all the pue stategy Nash equilibia of this game, we simply have to go though evey possible combination of stategies and ask whethe eithe playe would want to switch given what the othe is doing. Doing an exhaustive seach on 6

7 the example game above shows us that thee ae fou pue stategy Nash equilibia in this game: (, c), (s, b), (s, d) and(t, d). Notice that thee of these equilibia actually involve (weakly) dominated stategies. Fo Playe 1, s weakly dominates t since it does bette against a and c and equally well against b and d. Likewise, fo Playe 2, c weakly dominates b. Despite this, howeve, these stategies can be pat of Nash equilibia of the game. Nash equilibia can involve stategies that ae weakly dominated, but not stategies that ae stongly dominated. So a Nash equilibium captues a solution to the game. But what does that mean? It tuns out that thee is not a clea consensus on what the solution is supposed to tell us. Does it tell us how we ought to play the game? Does it make apedictionabouthowpeoplewill play the game? O does it just tell us what idealized (and fictitious) ational beings would do if they played a game and if that is all it tells us, why is that useful to know? It tuns out that the concept of a Nash equilibium is useful no matte what you think the solution eally means. It tells us what we ought to do if we want to get the most we can out of agameifeveyoneelseistyingtodothesame(andweallknowthatweaeall tying to do this, and we all know that we all know, etc). But, the Nash equilibia of a game ae also the points that will goven the way that individuals will lean when playing a game. They also tell us something about how mindless agents (like bacteia) might evolve when playing a game. These ae topics that will be discussed late on. One inteesting fact abut all finite games games with a finite numbe of stategies and a finite numbe of playes is that all games have at least one Nash equilibium. This means that all games have at least one ational solution. If the games ae lage and complicated enough, it may be too di cult to find the solution even though we know thee is one. And in some games, the only Nash equilibia cannot be expessed in tems of a specific set of stategies and involve playes mixing between stategies of the game. 4 Mixed Stategies Conside the well known game Rock-Pape-Scissos (expessed in nomal fom below). We can wok though evey possible combination of stategies to pove what we all aleady know: no matte combination of stategies you choose, someone wants to switch to a di eent stategy. This means that thee is no pue stategy Nash equilibium in this game. Instead, the best thing to do in this game is to be unpedictable. We might ty to choose a stategy at andom. Simply playing in Rock- Pape-Scissos is called a pue stategy but when we ty to act unpedicatbly 7

8 Playe 2 Playe 1 p s 0,0-1,1 1,-1 p 1,-1 0,0-1,1 s -1,1 1,-1 0,0 Rock-Pape-Scissos sometimes using one stategy, sometimes anothe we ae using a mixed stategy. A mixed stategy is defined by giving a pobability to each of the pue stategies in the game. We use the symbol to epesent a mixed stategy and s to epesent pue stategies. A mixed stategy specifies the chance that a playe chooses each pue stategy. Fo example, the mixed stategy whee you choose, p and s with equal chance is: =(1/3, 1/3, 1/3). And the mixed stategy whee we play half the time and p and s one quate of the time each is: =(1/2, 1/4, 1/4). We can also have mixed stategies that ignoe some of the pue stategies. Fo example, the mixed stategy that only uses and s with equal chance is: =(1/2, 0, 1/2). When you know opponent is using a mixed stategy, you can conside the expected utility fo each action you might take. Suppose Playe 1 believes Playe 2willuseamixedstategy 2 =(1/2, 2/6, 1/6) in Rock-Pape-Scissos. If P 2 (s) is the pobability that Playe 2 uses stategy s, then Playe 1 can calculate how well she expects each of he stategies to do against Playe 2 using the idea of expected utility. Fo example, hee is Playe 1 s expected utility fom playing : EU(, 2) =P 2 ()u 1 (, )+P 2 (p)u 1 (, p)+p 2 (s)u 1 (, s) EU(, 2) =(1/2)(0) + (2/6)( 1) + (1/6)(1) = 1/6 Using the same method to calculate his expected utility fo he othe two stategies shows that playing p would give the best expected esult: EU(p, 2) =(1/2)(1) + (2/6)(0) + (1/6)( 1) = 2/6 EU(s, 2) =(1/2)( 1) + (2/6)(1) + (1/6)(0) = 1/6 So, if Playe 1 thinks that Playe 2 is going to use this stategy, she should just play p and she will win moe often than she loses. 8

9 Howeve, if Playe 2 can anticipate that Playe 1 might figue out how to exploit any tendencies towad one stategy ove anothe, Playe 2 might decide to just play all stategies with equal chance. And, if Playe 2 uses the mixed stategy =(1/3, 1/3, 1/3), Playe 1 will not be able to take advantage of any tendency. In othe wods, if Playe 2 uses the =(1/3, 1/3, 1/3) stategy, Playe 1 will be indi eent to all he stategies. Futhemoe, if Playe 1 easons the same way and is woied about Playe 2 potentially taking advantage of any tendencies, Playe 1 will also use the = (1/3, 1/3, 1/3) stategy. When both playes use this stategy in Rock-Pape- Scissos, neithe playe wants to do anything di eent, and it is a Nash equilibium of the game. So, Rock-Pape-Scissos has a Nash equilibium, but only when both playes ae using just the ight mixed stategies. To detemine the payo that two playes get when they ae both using mixed stategies, we use the expected utility as befoe, but must now combine both playes stategies. If Playe 1 has n pue stategies and plays a mixed stategy 1 = (x 1,x 2,...,x n )andplaye2hasmstategies and uses a mixed stategy 2 = (y 1,y 2,...,y m ), then Playe 1 s payo against Playe 2 can be calculated as follows: u( 1, 2) = nx mx u(s i, j )x i y j. i=1 j=1 s i epesents Playe 1 s ith stategy and j epesents Playe 2 s jth stategy. In othe wods, we conside evey possible between the two stategies and multiply the payo of that outcome with the chance that outcome happens. Then, we add up all those possibilities. 4.1 Finding a mixed stategy equilibium in 2 2 games Finding mixed stategy equilibia in games can be vey had, especially if the games have many playes o many stategies. But, fo small games thee ae some ticks we can use to find mixed stategy Nash equilibia. This section uses an example to demonstate how to find a mixed stategy Nash equilibium in games that have 2-playes with 2-stategies each (sometimes called 2 2games ). Notice that when we found the equilibium fo Rock-Pape-Scissos, it was when both playes wee using stategies that made the othe playe not cae what which stategy they played. This tuns out to be a geneal fact about the mixed stategies that make up Nash equilibia. So, if we can figue out which combination of stategies make playes not cae which pue stategy they play, that will tell us whee the equilibium is. This is easiest to see with an example, conside the game below. 9

10 P laye2 P laye1 c d a 0, 0 1, 3 b 2, 1 0, 0 Example game #2 This game has thee equilibia. Two ae pue stategy equilibia: (a, d) and (b, c). But, it also has a thid whee if both playes mix between thei stategies just ight, neithe will do bette by doing anything di eently. To find this mixed equilibium, we need to find the stategies that make each playe indi eent. Fist conside the game fom the pespective of P 1. If P 2 plays c with pobability y and d with pobability (1 y), we ask: what value of y makes P 1 indi eent between a and b? TheexpectedutilityfoeachofPlaye1 sstategiescanbedetemined as follows: EU 1 (a, y) =P(c)u 1 (a, c)+p(d)u 1 (a, d) =y(0) + (1 EU 1 (b, y) =P(c)u 1 (b, c)+p(d)u 1 (b, d) =y(2) + (1 y)(1) y)(0) Playe 1 is indi eent between his two stategies a and b wheneve the expected utility of his two stategies is the same: EU 1 (b, y) =EU 1 (a, y) o... y(2) + (1 y)(0) = y(0) + (1 y)(1). Solving fo y gives us y =1/3. This means that Playe 1 will be indi eent between he stategies when Playe 2 plays c with pobability 1/3 andd with pobability 2/3. Now we do the same thing, but fom the pespective of Playe 2. Suppose Playe 1 plays a with pobability x and plays b with pobability (1 x). We now ask: what value of x makes Playe 2 indi eent between he two stategies? EU 2 (c, x) =P(a)u 2 (a, c)+p(b)u 2 (b, c) =x(0) + (1 EU 2 (d, x) =P(a)u 2 (a, d)+p(b)u 2 (b, d) =x(3) + (1 x)(1) x)(0) Playe 2 is indi eent when EU 2 (d, x) =EV (c, x) owhenevex(3) + (1 x)(0) = x(0)+(1 x)(1). Solving fo x gives us x =1/4, which means that Playe 2 will be 10

11 indi eent between he two stategies wheneve Playe 1 plays a with pobability 1/4 andb with pobability 3/4. By combining these two calculations, we can get the mixed stategy equilibium. If Playe 1 uses the mixed stategy 1 =(1/4, 3/4) and Playe 2 uses the mixed stategy 2 =(1/3, 2/3), then both playes will be indi eent between thei two stategies. This means that neithe playe can do bette by switching thei stategy given what the othe is doing. So, the stategy pofile ( 1, 2) is a Nash equilibium. 5 The Equilibium Selection Poblem The game in the pevious section (Example game #2) has thee Nash equilibia: the two pue stategy equilibia (b, c) and(a, d) alongwiththemixedstategy equilibium descibed in the pevious section. Many games, and most of the games that ae eally inteesting, have multiple Nash equilibia. What does game theoy pedict (o ecommend, depending on how we intepet a solution to the game) in these cases? If we want to use game theoy make a specific pediction, o to ecommend a paticula stategy when playing a game, we need to choose which of the solutions is the eal solution. How we choose among the equilibia will, of couse, depend on what we ae tying to use the solutions to epesent. But, no matte what we ae tying to captue, we will need to select which of the equilibia ae the elevant ones. This is called the equilibium selection poblem in game theoy. Conside the Stag Hunt game (below). In this game thee ae thee equilibia: (c,c), (d,d) and a mixed equilibium whee both playes choose c 75% of the time and d the est of the time. So what would game theoy pedict people will do when playing this game when evey action that the playes could take fom pat of some equilibium o othe? Without a way to appoach the equilibium selection poblem, game theoy tells us nothing about what to pedict (o what to do) in this game. P laye2 P laye1 c d c 4, 4 0, 3 d 3, 0 3, 3 The Stag Hunt Thee ae many di eent aspects of the game and its vaious Nash equilibia that we might conside when facing this poblem. Fo example, the (c,c) equilib- 11

12 ium gives the best possible payo to eveyone. This equilibium is called paeto optimal and is the payo dominant equilibium. But an agument can be made that the othe equilibium, (d,d) is the moe natual solution because it would esult fom both playes avoiding isk. Suppose that both ae eally not sue what the othe will do and think thee is an equal chance that the othe playe will pay any of he stategies. Then, choosing d guaantees the playes a decent playo. The (d,d) equilibium is the isk dominant equilibium. It is these kind of consideations that can help appoach the equilibium selection poblem. Howeve, thee is not a unique solution to this poblem and the best way to appoach it will depend on what we ae tying to epesent using game theoy. As mentioned above, the equilibium selection poblem aises in many games. This is especially tue in games whee seveal moves ae made and the ode of the moves mattes. These games ae di cult to epesent in the way that we have been doing so fa (the nomal fom), but can be epesented in othe ways. 6 Extensive Fom Games Not all games ae best epesented in nomal fom. In some games, one playe moves fist and the othe playe can espond. In these games, the ode of actions makes a di eence and this is not easily epesented in the nomal fom matices. To captue games whee the sequence of actions is impotant, we use what is called an extenstive fom epesentation. Conside the chain-stoe game. In this game, one company (P1) must choose between enteing a new maket (e) o not (n). If they choose not to, then eveything emains as it is. If they choose to ente the maket, anothe company (P2) must choose whethe o not to challenge (c o d). If P2 decides not to challenge, the playes split the maket. If P2 decides to challenge, they must pay a heavy cost since both playes will be foced to sell vey cheaply o lose to the othe playe. This game is best epesented in the extensive fom (see Figue 1). Which captues the sequence of move in a tee-fom. The game can also be expessed in nomal fom (see below), but this fom does not captue the fact that the moves ae made in a cetain sequence. P 1 P 2 c d e 1, 1 2, 2 n 0, 4 0, 4 The Chain-Stoe Game (nomal fom) 12

13 0, 4 @ P 1, 1 2, 2 Figue 1: The chain-stoe game in extensive fom Examining the nomal fom epesentation of the chain-stoe game, we can see that thee ae two pue-stategy Nash equilibia: (e, d) and(n, c). The fist of these is whee P1 entes the maket and P2 gives up half of it. The second is whee P1 does not ente the maket and P2 would challenge if P1 wee to ente the maket. Notice that thee is something a little stange about the (n, c) equilibium if P1 eve actually chose e, it wouldn t make sense fo P2 to chose c, as they would be choosing -1 ove 2. It only makes sense fo P2 to choose c if P2 neve has to make a choice! This means that the (n, c) equilibiumismaintained by a theat that wouldn t be ational to cay out if it came down to it. This equilibium is not subgame pefect (see the next section). We can also epesent uncetainty within extensive fom games. And, we can use this uncetainty to epesent simultaneous moves using the extensive fom. Conside the Diving Game (fom above) whee =Rightandl =Left.Wecan suppose that Playe 1 does move fist, but that Playe 2 simply does not know what Playe 1 chose. The dotted line between the two choice points indicates that P2 does not know which choice point she is located. l 1, 1 P1 b HHH l H pppppppppppppppppppp P2 0, 0 0, 0 1, 1 Figue 2: The Diving Game in extensive fom One famous game that is best expessed in extensive fom is the Ultimatum Game. Suppose that thee ae two playe that must divide 10 units of some good: the Popose and the Responde. The Popose poposes a split to the Responde, who can accept o eject the split. If the Responde accepts, each gets thei shae of the poposed split. If the Responde ejects, each gets nothing. To simplify things, we can suppose that the Popose only has two possible poposals 13

14 to make: an unfai 9-1 split o a fai 5-5 split. This simplified game is epesented in Figue eject 5, 5 0, 0 9, 1 0, 0 Figue 3: A Simplified Ultimatum Game A stategy in an extensive fom game has to tell the playe what to do at evey single choice point that they might make. So, the Popose s stategy set is {fai, unfai}. The Responde s has fou possible stategies, which specify what she would say to each possible o e: {accept both, accept fai/eject unfai, eject fai/accept unfai, eject both} o {aa, a, a, } fo shot. Thee ae thee pue-stategy Nash equlibia of this game: (unfai, accept both), (unfai, eject fai/accept unfai), and (fai, accept fai/eject unfai). This is easiest to see if we epesent the game in nomal fom (below). Responde Popose aa a a fai 5,5 5,5 0,0 0,0 unfai 9,1 0,0 9,1 0,0 Simplified Ultimatum Game (nomal fom) 7 Subgame Pefection Two of the equilibia in the simplified Ultimatum Game have a athe stange featue if they ae to be consideed ational solutions. Both the (unfai, eject fai/accept unfai), and (fai, accept fai/eject unfai) equilibia involve ejections whee the Responde e ectively chooses nothing ove something. The only eason that these ae equilibia is just that the Responde neve actually has to make that choice given what the Popose is o eing. But, if the Popose wee to eve o e anything di eent, it wouldn t be in the Responde s inteest to eject the o e (something is always bette than nothing). 14

15 So, these equilibia ae maintained by the theat of a possible futue action that is iational. Some games, such as the simplified Ultimatum Game, contain what ae called subgames. A subgame is a game-within-a-game. The simplified Ultimatum Game (in Figue 3) has thee subgames: (i) the whole game, (ii) the fok in the tee on the left side beginning at the Responde s choice, and (iii) the fok in the tee on the ight side beginning at the Responde s choice. Using this idea, we can define a moe ational vesion of the Nash equilibium: the subgame pefect equilibium. Subgame Pefect Equilibium: A Nash equilibium of a game that also foms a Nash equilibium of all the subgames. In the case of the simplified Ultimatum Game, thee is only one subgame pefect equilibium: (unfai, accept both). Thinking of the two small subgames, it is iational fo the Responde to choose eject in eithe case. And, given that it is always ational fo the Responde to accept, it is ational fo the Popose to o e the unfai split. In the othe two Nash equilibia of the game, the Responde has a stategy that does not constitute a Nash equilibium of one o othe of the subgames. The idea of a subgame pefect equilibium is one way that we might go about solving the equilibium selection poblem in extensive fom games. This might be of special inteest if we want ou solutions to captue the ational outcomes of agame. 7.1 Backwad Induction One common method fo finding a subgame pefect equilibium is called backwad induction. This method begins by looking at the vey end of the game and asking: if this decision point wee eached, what would a ational playe do? Then imagining that the game ends with that ational decision, you move on to the pevious decision in the game and ask again: given that the last choice will be ational, what is the ational choice to make at the second-to-last choice points? And so on, to the beginning of the game. Pb1 P 2 R P 1 P 2 R P 1 P 2 R 6, 6 d D d D d D 2, 0 1, 3 4, 2 3, 5 6, 4 5, 7 Figue 4: The Centipede Game 15

16 Conside the Centipede Game (Figue 4). In this game, playes take tuns choosing whethe o not to take the majoity of a pool of esouces and end the game o to pass the choice to the othe playe, theeby inceasing the pool of esouces. The game continues fo a fixed numbe of ounds, if no one takes the majoity, the game ends with an equal split. Accoding to the method of backwad induction, lets begin at the end of the game. At the end of the game, Playe 2 should choose to take the majoity and end the game (play D). Given that playe two should do this, Playe 1 should take the majoity one ound soone and end the game. Given that Playe 1 should do this at the second-to-last point, Playe 2 should take the majoity at the the thid-to-last-point, and so on. The subgame pefect equilibium is fo both playes to take the majoity at evey possible choice point: (ddd, DDD). 16

Lesson 33: Horizontal & Vertical Circular Problems

Lesson 33: Horizontal & Vertical Circular Problems Lesson 33: Hoizontal & Vetical Cicula Poblems Thee ae a wide vaiety of questions that you do if you apply you knowledge of cicula motion coectly. The tough pat is figuing out how to set them up. You need

More information

10 Torque. Lab. What You Need To Know: Physics 211 Lab

10 Torque. Lab. What You Need To Know: Physics 211 Lab b Lab 10 Toque What You Need To Know: F (a) F F Angula Systems Evey lab up to this point has dealt with objects moving in the linea system. In othe wods, objects moving in a staight line. Now we ae going

More information

Depth-first search and strong connectivity in Coq

Depth-first search and strong connectivity in Coq 1 Depth-fist seach and stong connectivity in Coq Januay 9, 2015 2 The poblem Finding the stongly connected components of a diected gaph. Pedagogical value: The fist nontivial gaph algoithm. Pactical value:

More information

Design Engineering Challenge: The Big Dig Contest Platter Strategies: Ball Liberation

Design Engineering Challenge: The Big Dig Contest Platter Strategies: Ball Liberation Poblem Set 4: Unifom Cicula Motion Design Engineeing Challenge: The Big Dig.007 Contest Platte Stategies: Ball Libeation Oeall Notes: You ae not equied to pefom the actual analysis in this poblem set,

More information

Multi-Robot Forest Coverage

Multi-Robot Forest Coverage Multi-Robot Foest Coveage Xiaoming Zheng Sonal Jain Sven Koenig David Kempe Depatment of Compute Science Univesity of Southen Califonia Los Angeles, CA 90089-0781, USA {xiaominz, sonaljai, skoenig, dkempe}@usc.edu

More information

Noncrossing Trees and Noncrossing Graphs

Noncrossing Trees and Noncrossing Graphs Noncossing Tees and Noncossing Gaphs William Y. C. Chen and Shey H. F. Yan Cente fo Combinatoics, LPMC, Nanai Univesity, 300071 Tianjin, P.R. China chen@nanai.edu.cn, huifangyan@eyou.com Submitted: Sep

More information

Alternate stable states in coupled fishery-aquaculture systems. Melissa Orobko

Alternate stable states in coupled fishery-aquaculture systems. Melissa Orobko Altenate stable states in coupled fishey-aquacultue systems by Melissa Oobko A thesis submitted in confomity with the equiements fo the degee of Maste of Science Depatment of Ecology & Evolutionay Biology

More information

British Prime Minister Benjamin Disraeli once remarked that

British Prime Minister Benjamin Disraeli once remarked that GABREL COSTA, MCHAEL HUBER, & OHN SACCOMAN Cumulative Home Run Fequency and the Recent Home Run Explosion Bitish Pime Ministe Benjamin Disaeli once emaked that thee ae thee kinds of falsehoods: lies, damned

More information

Deception in Honeynets: A Game-Theoretic Analysis

Deception in Honeynets: A Game-Theoretic Analysis Poceedings of te 7 IEEE Woksop on Infomation Assuance United States Militay Academy, West Point, NY, - June 7 Deception in Honeynets: A Game-Teoetic Analysis Nandan Gag and Daniel Gosu Abstact Recently,

More information

Experiment #10 Bio-Physics Pre-lab Questions

Experiment #10 Bio-Physics Pre-lab Questions Expeient #10 Bio-Physics Pe-lab Questions ** Disclaie: This pe-lab is not to be copied, in whole o in pat, unless a pope efeence is ade as to the souce. (It is stongly ecoended that you use this docuent

More information

The Properties of. Model Rocket Body Tube Transitions

The Properties of. Model Rocket Body Tube Transitions The Popeties of Moel ocket Boy Tube Tansitions Date: Septembe 6, 000 Pepae By: J.. Bohm NA 7808 CA S680 ev: (June, 005) .0 Intouction When esigning moel ockets, esignes often choose to incopoate iffeent

More information

An Auction Based Mechanism for On-Demand Transport Services

An Auction Based Mechanism for On-Demand Transport Services An Auction Based Mechanism fo On-Demand Tanspot Sevices Malcolm Egan, Ni Oen and Michal Jakob Abstact Recently, a new class of tanspot sevices, exemplified by Ube and Lyft has emeged. Such sevices use

More information

Cyclostrophic Balance in Surface Gravity Waves: Essay on Coriolis Effects

Cyclostrophic Balance in Surface Gravity Waves: Essay on Coriolis Effects Jounal of Oceanogaphy, Vol. 53, pp. 311 to 315. 1997 Shot Contibution Cyclostophic Balance in Suface Gavity Waves: Essay on Coiolis Effects KERN E. KENYON 4632 Noth Lane Del Ma, CA 92014-4134, U.S.A. (Received

More information

OPTIMAL SCHEDULING MODELS FOR FERRY COMPANIES UNDER ALLIANCES

OPTIMAL SCHEDULING MODELS FOR FERRY COMPANIES UNDER ALLIANCES Jounal of Maine Science and Technology, Vol. 15, No. 1, pp. 53-66 (2007) 53 OPTIMAL SCHEDULING MODELS FOR FERRY COMPANIES UNDER ALLIANCES Shangyao Yan*, Chia-Hung Chen**, Hsin-Yen Chen*** and Tze-Chiang

More information

Session 6. Global Imbalances. Growth. Macroeconomics in the Global Economy. Saving and Investment: The World Economy

Session 6. Global Imbalances. Growth. Macroeconomics in the Global Economy. Saving and Investment: The World Economy Session 6. Global Imbalances. Gowth. v, and the Real Inteest Rate v Global Imbalances v Gowth v Intoduction to exchange ates and : The Wold Economy The eal inteest ate is the pice that equilibates saving

More information

Faster Nearest Neighbors: Voronoi Diagrams and k-d Trees

Faster Nearest Neighbors: Voronoi Diagrams and k-d Trees 154 25 Jonathan Richad Shewchuk Faste Neaest Neighbos: Voonoi Diagams and k-d Tees SPEEDING UP NEAREST NEIGHBOR CLASSIFIERS Can we pepocess taining pts to obtain sublinea quey time? 2 5 dimensions: Voonoi

More information

The Study About Stopping Distance of Vehicles

The Study About Stopping Distance of Vehicles Intenational Jounal of Infomation Engineeing and Applications 018; 1(1): 18- http://www.aascit.og/jounal/infomation The Study About Stopping Distance of Vehicles Zhao Chun-xue School of Mathematics and

More information

EcoMobility World Festival 2013 Suwon: an analysis of changes in citizens awareness and satisfaction

EcoMobility World Festival 2013 Suwon: an analysis of changes in citizens awareness and satisfaction IOSR Jounal Engineeing (IOSRJEN) ISSN (e): 2250-3021, ISSN (p): 2278-8719 Vol. 07, Issue 03(Mach 2017), V1 PP 40-48 www.iosjen.og EcoMobility Wold Festival 2013 Suwon: an analysis changes in citizens awaeness

More information

Experiment #10 Bio-Physics Pre-lab Comments, Thoughts and Suggestions

Experiment #10 Bio-Physics Pre-lab Comments, Thoughts and Suggestions Expeient #10 Bio-Physics Pe-lab Coents, Thoughts and Suggestions The pupose of this pape is to povide you with soe infoation which ay be useful fo solving the pe-lab questions and pefoing the lab. I will

More information

Rotor Design and Analysis of Stall-regulated Horizontal Axis Wind Turbine

Rotor Design and Analysis of Stall-regulated Horizontal Axis Wind Turbine Roto Design and Analysis of Stall-egulated Hoizontal Axis Wind Tubine Xinzi Tang Univesity of Cental Lancashie, Peston, UK XTang4@uclan.ac.uk Xiongwei Liu Univesity of Cental Lancashie, Peston, UK XLiu9@uclan.ac.uk

More information

Cheat-Proof Playout for Centralized and Distributed Online Games

Cheat-Proof Playout for Centralized and Distributed Online Games Cheat-Poof Playout fo Centalized and Distibuted Online Games Nathaniel E. Baughman Bian Neil Levine baughman@cs.umass.edu bian@cs.umass.edu Depatment of Compute Science Univesity of Massachusetts Amhest,

More information

CORESTA RECOMMENDED METHOD N 68

CORESTA RECOMMENDED METHOD N 68 COESTA ECOMMENDED METHOD N 68 DETEMINATION OF CABON MONOXIDE IN THE MAINSTEAM SMOKE OF CIGAS BY NON-DISPESIVE INFAED ANALYSIS (Januay 2010) 1. FIELD OF APPLICATION The method is applicable to the gas phase

More information

DETC A NEW MODEL FOR WIND FARM LAYOUT OPTIMIZATION WITH LANDOWNER DECISIONS

DETC A NEW MODEL FOR WIND FARM LAYOUT OPTIMIZATION WITH LANDOWNER DECISIONS Poceedings of the ASME Intenational Design Engineeing Technical Confeences& Computes and Infomation in Engineeing Confeence IDETC/CIE August 8-3,, Washington, DC, USA DETC-4777 A NEW MODEL FOR WIND FARM

More information

Multiple Vehicle Driving Control for Traffic Flow Efficiency

Multiple Vehicle Driving Control for Traffic Flow Efficiency Multiple Vehicle Diving Contol fo Taffic Flow Efficiency Seong-Woo Kim, Gi-Poong Gwon, Seung-Tak Choi, Seung-am Kang, Myoung-Ok Shin, In-Sub oo, Eun-Dong Lee, and Seung-Woo Seo Abstact The dynamics of

More information

A Deceleration Control Method of Automobile for Collision Avoidance based on Driver's Perceptual Risk

A Deceleration Control Method of Automobile for Collision Avoidance based on Driver's Perceptual Risk A Deceleation Contol Method of Automobile fo Collision Avoidance based on Dive's Peceptual Risk Takahio Wada, Shun ichi Doi, and Shoji Hiaoka Abstact To educe ea-end cash of automobiles, it is impotant

More information

Example. The information set is represented by the dashed line.

Example. The information set is represented by the dashed line. Ou ast equiibium concet The ast equiibium concet we study afte Nash eqm, Subgame Pefect Nash eqm, and Bayesian Nash eqm is Pefect Bayesian Equiibium. Pefect efes to the fact that the game wi be dynamic,

More information

An integrated supply chain design model with random disruptions consideration

An integrated supply chain design model with random disruptions consideration Afican Jounal of Business Management Vol. 4(1), pp. 393-401, 18 Septembe, 010 Available online at http://www.academicounals.og/ajbm ISSN 1993-833 010 Academic Jounals Full Length Reseach Pape An integated

More information

Rearranging trees for robust consensus

Rearranging trees for robust consensus Poceedings of the IEEE Conf. on Decision and Contol, Olando, FL, 011 Reaanging tees fo obust consensus Geoge Foest Young, Luca Scadovi and Naomi Ehich Leonad Abstact In this pape, we use the H nom associated

More information

UNIVERSITÀ DEGLI STUDI DI PADOVA. Dipartimento di Scienze Economiche Marco Fanno

UNIVERSITÀ DEGLI STUDI DI PADOVA. Dipartimento di Scienze Economiche Marco Fanno UNIVERSITÀ DEGLI STUDI DI PADOVA Dipatimento di Scienze Economiche Maco Fanno DUAL LICENSING IN OPEN SOURCE SOFTWARE MARKETS STEFANO COMINO Univesity of Udine FABIO M. MANENTI Univesity of Padova Januay

More information

Torque. Physics 2. Prepared by Vince Zaccone For Campus Learning Assistance Services at UCSB

Torque. Physics 2. Prepared by Vince Zaccone For Campus Learning Assistance Services at UCSB Toque Physics Toque Toque is what causes angula acceleation (just like a foce causes linea acceleation) Toque Toque is what causes angula acceleation (just like a foce causes linea acceleation) Fo a toque

More information

Multi-Robot Flooding Algorithm for the Exploration of Unknown Indoor Environments

Multi-Robot Flooding Algorithm for the Exploration of Unknown Indoor Environments 2010 IEEE Intenational Confeence on Robotics and Automation Anchoage Convention Distict May 3-8, 2010, Anchoage, Alaska, USA Multi-Robot Flooding Algoithm fo the Exploation of Unknown Indoo Envionments

More information

Wind and extremely long bridges a challenge for computer aided design

Wind and extremely long bridges a challenge for computer aided design Wind and extemely long bidges a challenge fo compute aided design oian JANJIC anaging iecto TV / entley Systems Gaz, Austia oian Janjic, bon 960, civil engineeing degee fom the Faculty of Civil Engineeing,

More information

Color Encodings: srgb and Beyond

Color Encodings: srgb and Beyond IS&T's 2002 PICS Confeence Colo Encodings: srgb and Beyond Kevin E. Spaulding Eastman Kodak Company, Rocheste, New Yok Jack Holm Hewlett-Packad Company, Palo Alto, Califonia Abstact Most digital imaging

More information

arxiv:cs/ v1 [cs.ds] 8 Dec 1998

arxiv:cs/ v1 [cs.ds] 8 Dec 1998 Minimum Cuts in Nea-Linea Time David R. Kage Febuay 1, 2008 axiv:cs/9812007v1 [cs.ds] 8 Dec 1998 Abstact We significantly impove known time bounds fo solving the minimum cut poblem on undiected gaphs.

More information

MODELLING THE INTERACTION EFFECTS OF THE HIGH-SPEED TRAIN TRACK BRIDGE SYSTEM USING ADINA

MODELLING THE INTERACTION EFFECTS OF THE HIGH-SPEED TRAIN TRACK BRIDGE SYSTEM USING ADINA MODELLING THE INTERACTION EFFECTS OF THE HIGH-SPEED TRAIN TRACK BRIDGE SYSTEM USING ADINA ABSTRACT Constança Rigueio Depatment of Civil Engineeing, Polytechnic Institute of Castelo Banco Potugal Calos

More information

I. FORMULATION. Here, p i is the pressure in the bubble, assumed spatially uniform,

I. FORMULATION. Here, p i is the pressure in the bubble, assumed spatially uniform, The natual fequency of oscillation of gas bubbles in tubes H. N. Og uz and A. Pospeetti Depatment of Mechanical Engineeing, The Johns Hopkins Univesity, Baltimoe, Mayland 21218 Received 28 July 1997; accepted

More information

Morrison Drive tel. Ottawa, ON, Canada K2H 8S fax. com

Morrison Drive tel. Ottawa, ON, Canada K2H 8S fax.   com acomaecom 302 1150 Moison Dive 613 820 8282 tel Ottawa, ON, Canada K2H 8S9 613 820 8338 fax www.aecom. com To Pat Seguin, P. Eng. Manage of Engineeing Page 1 CC Subject Valeie McGi, Ted Achuticz Pine Steet

More information

CS3350B Computer Architecture. Lecture 6.2: Instructional Level Parallelism: Hazards and Resolutions

CS3350B Computer Architecture. Lecture 6.2: Instructional Level Parallelism: Hazards and Resolutions CS3350B Compute Achitectue Winte 2015 Lectue 6.2: Instuctional Level Paallelism: Hazads and Resolutions Mac Moeno Maza www.csd.uwo.ca/couses/cs3350b [Adapted fom lectues on Compute Oganization and Design,

More information

PREDICTION OF THIRD PARTY DAMAGE FAILURE FREQUENCY FOR PIPELINES TRANSPORTING MIXTURES OF NATURAL GAS AND HYDROGEN Zhang, L. 1, Adey, R.A.

PREDICTION OF THIRD PARTY DAMAGE FAILURE FREQUENCY FOR PIPELINES TRANSPORTING MIXTURES OF NATURAL GAS AND HYDROGEN Zhang, L. 1, Adey, R.A. PREDICTION OF THIRD PARTY DAMAGE FAILURE FREQUENCY FOR PIPELINES TRANSPORTING MIXTURES OF NATURAL GAS AND HYDROGEN Zhang, L., Adey, R.A. 2 C M BEASY Ltd, Ashust Lodge, Southampton, SO4 7AA, UK, Lzhang@beasy.com

More information

A Force Platform Free Gait Analysis

A Force Platform Free Gait Analysis Poceedings A Foce Platfom Fee Gait Analysis Tokio Maeda 1,2, *, Tatsuo Ishizuka 3, Sakua Yamaji 4 and Yuji Ohgi 3 1 Keio Reseach Institute at SFC, 5322 Endo, Fujisawa, Kanagawa 252-0882, Japan 2 Koseki

More information

Trends in Cycling, Walking & Injury Rates in New Zealand

Trends in Cycling, Walking & Injury Rates in New Zealand Tends in Cycling, Walking & Injuy Rates in New Zealand Walking and cycling, minutes/week SUMMAR Estimates of walking and cycling fom 9 Household Tavel suveys in New Zealand, conducted etween 1989 and 2012,

More information

Matlab Simulink Implementation of Switched Reluctance Motor with Direct Torque Control Technique

Matlab Simulink Implementation of Switched Reluctance Motor with Direct Torque Control Technique Matlab Simulink Implementation of Switched Reluctance Moto with Diect Toque Contol Technique Vikamaajan Jambulingam Electical and Electonics Engineeing, VIT Univesity, India. Abstact - The switched eluctance

More information

Lecture Topics. Overview ECE 486/586. Computer Architecture. Lecture # 9. Processor Organization. Basic Processor Hardware Pipelining

Lecture Topics. Overview ECE 486/586. Computer Architecture. Lecture # 9. Processor Organization. Basic Processor Hardware Pipelining EE 486/586 ompute chitectue Lectue # 9 asic Pocesso Hadwae Pipelining Refeence: ppendix : Section.1 Lectue Topics Sping 2015 Potland State Univesity Oveview Pocesso Oganization typical computing task consists

More information

A tale of San Diego County s water. If you had to describe San Diego's weather, you probably would use

A tale of San Diego County s water. If you had to describe San Diego's weather, you probably would use keep San Diego beautiful Dought-toleant plants envionmentally-sensitve Bay-Delta The 1895 viewing Sweetwate Dam, Two women A tale of San Diego County s wate If you had to descibe San Diego's weathe, you

More information

Complexity of Data Tree Patterns over XML Documents

Complexity of Data Tree Patterns over XML Documents Complexity of Data Tee Pattens ove XML Documents Claie David LIAFA, Univesity Pais 7 and CNRS, Fance cdavid@liafa.jussieu.f Abstact. We conside Boolean combinations of data tee pattens as a specification

More information

Bicycle and Pedestrian Master Plan

Bicycle and Pedestrian Master Plan Falmouth, Maine Bicycle and Pedestian Maste Plan Januay, 2003 Pepaed by The Falmouth Tails Advisoy Committee Tom Williams, Chai Daeby Amble Elizabeth Ehenfeld John Locke William Lund William Robinson Robet

More information

Watford Half Marathon. Watford Half Marathon. Sunday February 4th Starting at 10.30am. Enjoy Your Run!!! Notice to all Entrants.

Watford Half Marathon. Watford Half Marathon. Sunday February 4th Starting at 10.30am. Enjoy Your Run!!! Notice to all Entrants. atfod Half Maathon 2018 Sunday Febuay 4th Stating at 1030am REF No 527 HEld UNdERUka RUlEs ORGANISED BY atfod Half Maathon Notice to all Entants elcome to the atfod Half Maathon, oganised by atfod Haies

More information

PlacesForBikes City Ratings Methodology. Overall City Rating

PlacesForBikes City Ratings Methodology. Overall City Rating 1 PlacesFoBikes City Ratings Methodology Oveall City Rating The PlacesFoBikes City Rating Scoe is based on five factos: Rideship, Safety, Netwok, Acceleation, and Reach. Each facto is scoed on a one to

More information

Incorporating Location, Routing and Inventory Decisions in Dual Sales Channel - A Hybrid Genetic Approach

Incorporating Location, Routing and Inventory Decisions in Dual Sales Channel - A Hybrid Genetic Approach Incopoating Location, Routing and Inventoy Decisions in Dual Sales Channel - A Hybid Genetic Appoach Chia-lin Hsieh 1, Shu-hsien Liao 2, Wei-chung Ho 2 1 Depatment of Statistics and Actuaial Science, Aletheia

More information

The Solution to the Bühlmann - Straub Model in the case of a Homogeneous Credibility Estimators

The Solution to the Bühlmann - Straub Model in the case of a Homogeneous Credibility Estimators 5 Economy Infomatics, -4/005 The Solution to the Bühlmann - Staub Model in the case of a Homogeneous Cedibility Estimatos Lect. Viginia ATANASIU Mathematics Depatment, Academy of Economic Studies Oiginal

More information

POSSIBLE AND REAL POWERFLOWS IN CONNECTED DIFFERENTIAL GEAR DRIVES WITH η 0 <i pq <1/η 0 INNER RATIO

POSSIBLE AND REAL POWERFLOWS IN CONNECTED DIFFERENTIAL GEAR DRIVES WITH η 0 <i pq <1/η 0 INNER RATIO 5th INTERNATIONAL MEETING OF TE CARPATIAN REGION SPECIALISTS IN TE FIEL OF GEARS POSSIBLE AN REAL POWERFLOWS IN CONNECTE IFFERENTIAL GEAR RIVES WIT η 0

More information

Watford Half Marathon. Watford Half Marathon. Sunday February 5th Starting at 10.30am. Enjoy Your Run!!! Notice to all Entrants.

Watford Half Marathon. Watford Half Marathon. Sunday February 5th Starting at 10.30am. Enjoy Your Run!!! Notice to all Entrants. atfod Half Maathon 2017 Sunday Febuay 5th Stating at 1030am REF No 527 HELD UNDER UKA RULES ORGANISED BY atfod Half Maathon Notice to all Entants elcome to the atfod Half Maathon, oganised by atfod Haies

More information

Accel. (m/s 2 ) Time (sec) Newton s 3 rd Law and Circular Motion. Group Problem 04

Accel. (m/s 2 ) Time (sec) Newton s 3 rd Law and Circular Motion. Group Problem 04 1) A 200 kg tuck acceleates eastwads on a hoizontal oad in esponse to a gadually inceasing fictional foce fom the gound. Thee is an unsecued 50 kg block sitting on the tuck bed line. Thee is fiction between

More information

Bubble clustering and trapping in large vortices. Part 1: Triggered bubbly jets investigated by phase-averaging

Bubble clustering and trapping in large vortices. Part 1: Triggered bubbly jets investigated by phase-averaging Intenational Jounal of Multiphase Flow 33 (2007) 1088 1110 www.elsevie.com/locate/ijmulflow Bubble clusteing and tapping in lage votices. Pat 1: Tiggeed bubbly jets investigated by phase-aveaging Rade

More information

A Collision Risk-Based Ship Domain Method Approach to Model the Virtual Force Field

A Collision Risk-Based Ship Domain Method Approach to Model the Virtual Force Field A Collision Risk-Based Ship Domain Method Appoach to Model the Vitual Foce Field Tengfei Wang a,b,c, Xinping Yan a,b *, Yang Wang a,b, and Qing Wu a,b,c a Intelligent Tanspotation System Reseach Cente,

More information

55CM ROUND CHARCOAL KETTLE BBQ

55CM ROUND CHARCOAL KETTLE BBQ IMPORTANT - Please ead the instuctions caefully befoe assembling 55CM ROUND CHARCOAL KETTLE BBQ Assembly A Instuctions PLEASE RETAIN THE INSTRUCTIONS FOR FUTURE USE. Safety and Cae Advice the Wait fo it

More information

Interior Rule of the Quebec Open 2017

Interior Rule of the Quebec Open 2017 Inteio Rule of the Quebec Open 2017 Aticle 1. The Oganize The Oganize is the club d'échecs de Sainte-Foy which will be epesented by M Chales Tenblay, chaiman of the oganizing committee Le tounoi se déoulea

More information

READING AREA TRANSPORTATION STUDY BICYCLE AND PEDESTRIAN TRANSPORTATION PLAN ADOPTED NOVEMBER 18, 2010

READING AREA TRANSPORTATION STUDY BICYCLE AND PEDESTRIAN TRANSPORTATION PLAN ADOPTED NOVEMBER 18, 2010 READING AREA TRANSPORTATION STUDY BICYCLE AND PEDESTRIAN TRANSPORTATION PLAN ADOPTED NOVEMBER 18, 2010 READING AREA TRANSPORTATION STUDY C/o Beks County Planning Commission 633 Cout St. FL 14 Reading,

More information

PREDICTION OF ELECTRICAL PRODUCTION FROM WIND ENERGY IN THE MOROCCAN SOUTH

PREDICTION OF ELECTRICAL PRODUCTION FROM WIND ENERGY IN THE MOROCCAN SOUTH OL., NO., FEBRUARY 7 ISSN 89-668 REDICTION OF ELECTRICAL RODUCTION FROM WIND ENERGY IN THE MOROCCAN SOUTH Elyaqouti Mustapha, Ezzaidi Asma, Bouhouch Lahoussine and Ihlal Ahmed ERTAIER, ESTA, Univesity

More information

EC-FRM: An Erasure Coding Framework to Speed up Reads for Erasure Coded Cloud Storage Systems

EC-FRM: An Erasure Coding Framework to Speed up Reads for Erasure Coded Cloud Storage Systems 215 44th Intenational Confeence on Paallel Pocessing EC-FRM: An Easue Coding Famewok to Speed up Reads fo Easue Coded Cloud Stoage Systems Yingxun Fu, Jiwu Shu *, and Zhiong Shen Tsinghua National Laboatoy

More information

Performance Characteristics of Parabolic Trough Solar Collector System for Hot Water Generation

Performance Characteristics of Parabolic Trough Solar Collector System for Hot Water Generation Intenational Enegy Jounal: Vol. 7, No. 2, June 2006 137 Pefomance Chaacteistics of Paabolic Tough Sola Collecto System fo Hot Wate Geneation www.sed.ait.ac.th/eic A. Valan Aasu and T. Sonakuma Faculty

More information

THE GREAT CARDBOARD BOAT RACE INTRODUCTION, BOAT BUILDING & RULES by United Way of Elkhart County

THE GREAT CARDBOARD BOAT RACE INTRODUCTION, BOAT BUILDING & RULES by United Way of Elkhart County THE GREAT CARDBOARD BOAT RACE INTRODUCTION, BOAT BUILDING & RULES by United Way of Elkhat County WHEN: Fiday August 26 2016 WHERE: LaSalle Bistol Pond Elkhat INTRODUCTION The Geat Cadboad Boat Race is

More information

Carnegie Mellon University Forbes Ave., Pittsburgh, PA command as a point on the road and pans the camera in

Carnegie Mellon University Forbes Ave., Pittsburgh, PA command as a point on the road and pans the camera in Panacea: An Active Senso Contolle fo the ALVINN Autonomous Diving System Rahul Sukthanka, Dean Pomeleau and Chales Thope Robotics Institute Canegie Mellon Univesity 5 Fobes Ave., Pittsbugh, PA 15213-3891

More information

0ur Ref:CL/Mech/ Cal /BID-01(11-12) Date: 29 July 2011

0ur Ref:CL/Mech/ Cal /BID-01(11-12) Date: 29 July 2011 0u Ref:CL/Mech/ Cal /BID-01(11-12) Date: 29 July 2011 SUBJECT: PROCUREMENT OF CALIBRATION SERVICES FOR THE EQUIPMENTS IN Cental Laboatoy (Mechanical) Dea Sis, Technical & Commecial s ae invited fo the

More information

Data Sheet. Linear bearings

Data Sheet. Linear bearings Data Pack G Issued Septembe 1999 1502325042 Data Sheet Linea beaings Instument ball beaings and shafts The RS ange of instument quality ball bushing beaings ae fo 1 8in, 3 16in and 1 4in shafts. Each beaing

More information

OUTDOOR FOOTBALL FIELD

OUTDOOR FOOTBALL FIELD www integalspo com USING AREAS: Mini football fields, pivate buildings, pivate schools and pivate entepeunes. SYSTEM DESCRIPTION: Composite steel constuction with synthetic tuf floo, oof and sides coveed

More information

THE IMPACTS OF CONGESTION ON COMMERCIAL VEHICLE TOUR CHARACTERISTICS AND COSTS

THE IMPACTS OF CONGESTION ON COMMERCIAL VEHICLE TOUR CHARACTERISTICS AND COSTS Figliozzi 1 THE IMPACTS OF CONGESTION ON COMMERCIAL VEHICLE TOUR CHARACTERISTICS AND COSTS Miguel Andes Figliozzi Depatment of Civil and Envionmental Engineeing Maseeh College of Engineeing and Compute

More information

Electrical Equipment of Machine Tools

Electrical Equipment of Machine Tools Electical Equipment of achine Tools Electification in machine tools today is vey extensive because it helps to simplify thei constuction, educe thei weight, and develop automatic contol. any advanced tends

More information

Fundamental Algorithms for System Modeling, Analysis, and Optimization

Fundamental Algorithms for System Modeling, Analysis, and Optimization Fundamental Algoithms fo System Modeling, Analysis, and Optimization Edwad A. Lee, Jaijeet Roychowdhuy, Sanjit A. Seshia UC Bekeley EECS 44/44 Fall With thanks to R. K. Bayton, K. Keutze, N. Shenoy, and

More information

Design and Simulation Model for Compensated and Optimized T-junctions in Microstrip Line

Design and Simulation Model for Compensated and Optimized T-junctions in Microstrip Line Intenational Jounal of Advanced Reseach in Compute Engineeing & Technology (IJARCET) Volume 3 Issue, Decembe 4 Design and Simulation Model fo Compensated and Optimized T-junctions in Micostip Line Alok

More information

Cavitation Bubble Dynamics in Non-Newtonian Fluids

Cavitation Bubble Dynamics in Non-Newtonian Fluids Cavitation Bubble Dynamics in Non-Newtonian Fluids Emil-Alexandu Bujan Depatment of Hydaulics, Univesity Politehnica Buchaest, 060042 Buchaest, omania Cavitation phenomena play impotant oles in many aeas

More information

THE performance disparity between processor speed and the

THE performance disparity between processor speed and the This aticle has been accepted fo publication in a futue issue of this jounal, but has not been fully edited. Content may change pio to final publication., VOL. X, NO. X, MONTH X Matix Stipe Cache-Based

More information

SHRiMP: Accurate Mapping of Short Color-space Reads

SHRiMP: Accurate Mapping of Short Color-space Reads SHRiMP: Accuate Mapping of Shot Colo-space Reads Stephen M. Rumble 1,2, Phil Lacoute 3,4, Adian V. Dalca 1, Mac Fiume 1, Aend Sidow 3,4, Michael Budno 1,5 * 1 Depatment of Compute Science, Univesity of

More information

Genetic Mapping Exercise - Extra Credit. Do not work together - each person to do their own work.

Genetic Mapping Exercise - Extra Credit. Do not work together - each person to do their own work. Geneic Mapping Execise - Exa Cedi Name Secion # Do no ok ogehe - each peson o do hei on ok. hen loci of diffeen allelic goups ae on he same chomosome, linkage occus bu is no absolue. Cossing-ove alays

More information

DECO THEORY - BUBBLE MODELS

DECO THEORY - BUBBLE MODELS DECO THEORY - BUBBLE MODELS This page descibes pinciples and theoies about bubble geneation and bubble gowth in the scuba dives body and about the effect of bubble fomation on decompession and decompession

More information

RESOLUTION No A RESOLUTION OF THE CITY OF SALISBURY, MARYLAND AUTHORIZING THE MAYOR TO ENTER INTO AN AGREEMENT BETWEEN THE CITY OF

RESOLUTION No A RESOLUTION OF THE CITY OF SALISBURY, MARYLAND AUTHORIZING THE MAYOR TO ENTER INTO AN AGREEMENT BETWEEN THE CITY OF li RESOLUTION No. 2749 A RESOLUTION OF THE CITY OF SALISBURY, MARYLAND AUTHORIZING THE MAYOR TO ENTER INTO AN AGREEMENT BETWEEN THE CITY OF SALISBURY, MARYLAND AND WICOMICO COUNTY, MARYLAND FOR THE DELIVERY

More information

f i r e - p a r t s. c o m

f i r e - p a r t s. c o m A M E R I C A N F L A M E A F AF-1000 Seies Manual Gas Contol Valve APPLICATION The AF-1000 Manual Gas Contol Valve is designed fo use in gas in heath poducts, wall mount funaces, and space heating appliances.

More information

Asteroid body-fixed hovering using nonideal solar sails

Asteroid body-fixed hovering using nonideal solar sails Reseach in Aston. Astophys. 4 Vol. X No. XX, http://www.aa-jounal.og http://www.iop.og/jounals/aa Reseach in Astonomy and Astophysics Asteoid body-fixed hoveing using nonideal sola sails Xiang-yuan Zeng,

More information

tr0 TREES Hanan Samet

tr0 TREES Hanan Samet t0 TREES Hanan Samet ompute Science epatment and ente fo utomation Reseach and Institute fo dvanced ompute Studies Univesity of Mayland ollege Pak, Mayland 074 e-mail: hjs@umiacs.umd.edu opyight 1997 Hanan

More information

A CONCEPTUAL WHEELED ROBOT FOR IN-PIPE INSPECTION Ioan Doroftei, Mihaita Horodinca, Emmanuel Mignon

A CONCEPTUAL WHEELED ROBOT FOR IN-PIPE INSPECTION Ioan Doroftei, Mihaita Horodinca, Emmanuel Mignon A CONCEPTUAL WHEELED ROBOT FOR IN-PIPE INSPECTION Ioan Dooftei, Mihaita Hoodinca, Emmanuel Mignon Univesité Libe de Buxelles Active Stuctues Laboatoy 50, Av. F.D.Roosevelt, B-1050, Bussels, Belgium. Email:

More information

"A Home for Gracious Living" (real estate brochure for Batten House)

A Home for Gracious Living (real estate brochure for Batten House) Byn Maw College Scholaship, Reseach, and Ceative Wok at Byn Maw College Achitectue, Gounds, and Histoy Facilities 1959 "A Home fo Gacious Living" (eal estate bochue fo Batten House) Let us know how access

More information

Phase Behavior Introduction to Phase Behavior F.E. Londono M.S. Thesis (2001)

Phase Behavior Introduction to Phase Behavior F.E. Londono M.S. Thesis (2001) Natual Gas Engineeing Phase Behavio Intoduction to Phase Behavio F.E. Londono M.S. hesis (001).. Blasingame, exas &M U. Depatment of Petoleum Engineeing exas &M Univesity College Station, X 77843-3116

More information

Numerical study of super-critical carbon dioxide flow in steppedstaggered

Numerical study of super-critical carbon dioxide flow in steppedstaggered The 6th Intenational Supecitical CO2 Powe Cycles Symposium Mach 27-29, 2018, Pittsbugh, Pennsylvania Numeical study of supe-citical cabon dioxide flow in steppedstaggeed labyinth seals Yuming Zhu a,b,

More information

PETTIPAUG YACHT CLUB PSA PETTIPAUG SAILING ACADEMY PETTIPAUG JUNIOR SAILING PROGRAM Parent & Sailor Handbook

PETTIPAUG YACHT CLUB PSA PETTIPAUG SAILING ACADEMY PETTIPAUG JUNIOR SAILING PROGRAM Parent & Sailor Handbook PETTIPAUG YACHT CLUB PSA PETTIPAUG SAILING ACADEMY PETTIPAUG JUNIOR SAILING PROGRAM 2017 Paent & Sailo Handbook PSA SCHEDULE Date Satuday June 17 th Thusday June 22 th Monday June 26 th - August 11 th

More information

Experimental and Numerical Studies on Fire Whirls

Experimental and Numerical Studies on Fire Whirls Expeimental and Numeical Studies on Fie Whils K. Matsuyama, N. Ishikawa 2, S. Tanaka 2, F. Tanaka, Y. Ohmiya 2, and Y. Hayashi 3 Cente fo Fie Science and Technology, Tokyo Univesity of Science, 264, Yamasaki,

More information

A Machine Vision based Gestural Interface for People with Upper Extremity Physical Impairments

A Machine Vision based Gestural Interface for People with Upper Extremity Physical Impairments IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS-PART A: SYSTEMS AND HUMANS 1 A Machine Vision based Gestual Inteface fo People with Uppe Extemity Physical Impaiments Haiong Jiang, Badley S. Duestock,

More information

West Central Scotland Cycling Feasibility Study Brief

West Central Scotland Cycling Feasibility Study Brief West Cental Scotland Cycling Feasibility Study Bief 14 Febuay, 2005 Intoduction Foesty Commission Scotland (FCS) in patneship with Scottish Entepise Foth Valley (and potentially Agyll and Island Entepise)

More information

Lecture 24. Wind Lidar (6) Direct Motion Detection Lidar

Lecture 24. Wind Lidar (6) Direct Motion Detection Lidar Lectue 24. Wind Lida (6) Diect Motion Detection Lida Diect Motion Detection Wind Lida Lida tacking of aeosol motions Lase time-of-flight velocimety Lase Dopple velocimety Compaison of wind lida techniques

More information

Tracking of High-speed, Non-smooth and Microscale-amplitude Wave Trajectories

Tracking of High-speed, Non-smooth and Microscale-amplitude Wave Trajectories Tacking of High-spee, Non-smooth an Micoscale-amplitue Wave Tajectoies Jiaech Kongthon Depatment of Mechatonics Engineeing, Assumption Univesity, Suvanabhumi Campus, Samuthpakan, Thailan Keywos: Abstact:

More information

Red-Black Trees Goodrich, Tamassia Red-Black Trees 1

Red-Black Trees Goodrich, Tamassia Red-Black Trees 1 Red-Black Tees 8 00 Goodich, Tamassia Red-Black Tees 1 Fom (,) to Red-Black Tees A ed-black tee is a epesentation of a (,) tee by means of a binay tee whose nodes ae coloed ed o black In compaison with

More information

CARDBOARD BOAT BUILDING 101

CARDBOARD BOAT BUILDING 101 CARDBOARD BOAT BUILDING 101 by United Way of Elkhat County What floats you Cadboad boat... United Way Kick Off August 28, 2015 CONSTRUCTION RULES The ENTIRE boat must be built of cadboad, duct tape, and

More information

ECE Introduction to Comm/SP, Fall 2011 Homework #5 Due: 11/04/11

ECE Introduction to Comm/SP, Fall 2011 Homework #5 Due: 11/04/11 2 O ECE 563 - Intoduction to Comm/SP, Fll 2011 omewok #5 Due: 11/04/11 1. Conside the following system to pocess continuous-time signls with discete-time pocessing. Smple: Feq = Convet to Sequence Rte

More information

Range Extension Control System for Electric Vehicles Based on Front and Rear Driving Force Distribution Considering Load Transfer

Range Extension Control System for Electric Vehicles Based on Front and Rear Driving Force Distribution Considering Load Transfer Range Extension Contol System fo Electic Vehicles Based on and Diving Foce Distibution Consideing Load Tansfe Sho Egami and Hioshi Fujimoto The Univesity of Tokyo 5--5, Kashiwanoha, Kashiwa, Chiba, 227-856

More information

tr0 TREES Hanan Samet

tr0 TREES Hanan Samet t0 TRS Hanan Samet ompute Science epatment and ente fo utomation Reseach and Institute fo dvanced ompute Studies Univesity of Mayland ollege Pak, Mayland 074 e-mail: hjs@umiacs.umd.edu opyight 1997 Hanan

More information

ANALYSIS AND TESTING OF AN INTEGRATED REFRIGERATION AND STORAGE SYSTEM FOR LIQUID HYDROGEN ZERO BOIL-OFF, LIQUEFACTION, AND DENSIFICATION

ANALYSIS AND TESTING OF AN INTEGRATED REFRIGERATION AND STORAGE SYSTEM FOR LIQUID HYDROGEN ZERO BOIL-OFF, LIQUEFACTION, AND DENSIFICATION ANALYSIS AND TESTING OF AN INTEGRATED REFRIGERATION AND STORAGE SYSTEM FOR LIQUID HYDROGEN ZERO BOIL-OFF, LIQUEFACTION, AND DENSIFICATION By WILLIAM USILTON NOTARDONATO A DISSERTATION PRESENTED TO THE

More information

Fault tolerant oxygen control of a diesel engine air system

Fault tolerant oxygen control of a diesel engine air system Fault toleant oxygen contol of a diesel engine ai system Raine Nitsche, Matthias Bitze, Mahmoud El Khaldi, Géad Bloch To cite this vesion: Raine Nitsche, Matthias Bitze, Mahmoud El Khaldi, Géad Bloch.

More information

CLASS: XI: MATHEMATICS

CLASS: XI: MATHEMATICS LASS: XI: MATHEMATIS ERMUTATIONS AND OMBINATIONS RATIE QUESTIONS ON FATORIAL AND FUNDAMENTAL RINILES OF OUNTING ove the followig fo N, 1. (2 )! 2.!.[1.3.5...(2 1)] FORMULA USED Factoial otatio:! o 1.!

More information

1 of 6 5/12/2015 8:02 PM

1 of 6 5/12/2015 8:02 PM 1 of 6 5/12/2015 8:02 PM TUESDAY MAY 12, 2015 10:01 PM oebea ate~12-5 Delmava Shoebids v. Augusta @ Athu W. Pedue Stadium /delmava-shoebidsv-augusta /?instance_id=994255) 11:00 pm Pub Paty Tivia @ The

More information

STUDY OF IRREGULAR WAVE-CURRENT-MUD INTERACTION

STUDY OF IRREGULAR WAVE-CURRENT-MUD INTERACTION STUDY OF IRREGULAR WAVE-CURRENT-MUD INTERACTION Mohsen Soltanpou, Fazin Samsami, Tomoya Shibayama 3 and Sho Yamao The dissipation of egula and iegula waves on a muddy bed with the existence of following

More information

Multiple Adjunction in Feature-Based Tree-Adjoining Grammar

Multiple Adjunction in Feature-Based Tree-Adjoining Grammar Multiple Adjunction in Featue-Based Tee-Adjoining Gamma Claie Gadent, Shashi Naayan To cite this vesion: Claie Gadent, Shashi Naayan. Multiple Adjunction in Featue-Based Tee-Adjoining Gamma. Computational

More information