Научная статья на тему 'Существование равновесия по Бержу'

Существование равновесия по Бержу Текст научной статьи по специальности «Математика»

CC BY
60
12
i Надоели баннеры? Вы всегда можете отключить рекламу.
Ключевые слова
БЕСКОАЛИЦИОННАЯ ИГРА / NONCOOPERATIVE GAME / ФУНКЦИЯ ВЫИГРЫША / PAYOFF FUNCTION / PAYOFF / NASH AND BERGE EQUILIBRIUM / ГЕРМЕЙЕРОВСКАЯ СВЕРТКА / GERMEIER CONVOLUTION / СМЕШАННЫЕ СТРАТЕГИИ / MIXED STRATEGIES / ВЫИГРЫШ / РАВНОВЕСИЕ ПО НЭШУ И БЕРЖУ

Аннотация научной статьи по математике, автор научной работы — Жуковский В.И., Макаркина Т.В., Бельских Ю.А.

В статье рассматривается способ построения равновесной по Бержу ситуации, сводящейся к нахождению минимаксной стратегии в специальной гермейеровской свертке, эффективно строящейся по исходной математической модели бескоалиционной игры.Кроме того, доказано существование равновесной по Бержу ситуации в смешанных стратегиях, если множества стратегий суть компакты, а функции выигрыша непрерывны на ситуациях.

i Надоели баннеры? Вы всегда можете отключить рекламу.
iНе можете найти то, что вам нужно? Попробуйте сервис подбора литературы.
i Надоели баннеры? Вы всегда можете отключить рекламу.

The Existence of Berge Equilibrium

The sufficient conditions of the existence of Berge equilibrium situation in noncooperative game of many persons in normal form are established. On the basis of these conditions the existence of Berge equilibrium situation in mixed strategies (by compact sets of strategies of players and continuity of their payoff functions) is proved. Let us consider the history of the appearance of the Berge equilibrium notion. In 1949 the 21-years-old PhD student of Prinston University, John F. Nash (jun.), formalized the notion of “good” solution in noncooperative games (later called “Nash equilibrium”). It has got the broad spectrum of applications in economics, sociology, military sciences. And now after more than 50 years, in any journal of system analysis, game theory, mathematical programming we find the papers devoted to Nash equilibrium (NE). In 1994 John Nash won the Nobel Prize in economics in a common effort with John Harsanyi and R. Selten “for fundamental analysis of equilibria in noncooperative game theory”. Actually 20-years-old Nash developed the foundation of the scientific method that played the great role in the development of world economy. However “in the sun there are spots” (proverb). And the main of them is “the eqoistic character” of Nash equilibrium concept. It appears in the fact that every player tries to increase only his own payoff, i.e. follows “politica dei campanile”, without considering interests of other participants of the conflict. One of the methods to remove this negative is to use the approach (by formalization of “good” solution of the game), which differs from “dictated” Nash equilibrium. Such approach was proposed in 1994 at the scientific seminar (leader V.I.Zhukovskiy) at discussing the book of C.Berge «Theorie generale des jeux a n personnes games» (this book was published in Paris in 1957 and in 1961 it was translated into Russian [1]). Concretely the criticism of NE was caused by non-existence NE at strongly concave in strategy at least one player his payoff function (but the decision making is necessary!). The sense of the new approach lies in change of condition of solution stability not to deviation of the player whom belongs “payoff function” but to deviation of all players except the one who is “the owner” of this payoff function. We shall note three circumstances. Firstly, we called the proposed new concept “BE”. The term “BE” arouse as the result of reviewing Claude Berge’s book. Secondly, in 1994 K.S.Vaisman (then the post-graduate student of V.Zhukovskiy) was engaged in construction of initial foundations of mathematical BE theory. In 1995 K. Vaisman defended his thesis “Berge equilibrium” (BE) in Leningrad University. (K.Vaisman died in 1998 at the age of 35 years). His sudden death suspended further development of the Berge equilibrium in Russia, but the notion of a Berge equilibrium was “exported from Russia” by Algerian scholars of V.Zhukovskiy M. Radjef and M. Larbani. This notion caused the broad interest of our foreign colleagues. The acquaintance with their publications showed that “par le temps qui ceurt” the most papers of this direction devoted to the properties of Berge equilibrium, singularities, modifications of this notion, relations with Nash equilibrium. It is supposed that in originated theory of Berge equilibrium the stage of formation of strict mathematical theory becomes nearer. Probably an intensive accumulation of facts will be replaced by the stage of evolutionary internal development. At this stage one should traditionally answer two fundamental questions: 1. Does the Berge equilibrium exist? 2. How one should find this equilibrium? The present article is just devoted to answers of these both questions. Thirdly, the authors were motivated by the IX Moscow Festival of Science that partially was held in a new building of MSU Fundamental library on October 10, 2014. Apart from lectures of Nobel laureates chemists Kurt Wuthrich (USA, California), Jean-Marie Lehn (France), biochemist Sir Richard Roberts (USA), RAS academician M.Ya.Marov (“The Chelyabinsk meteor”), L.M. Zelenyi (“Exoplanets: Searching for a second Earth”), Doctors of Sciences A.V. Markov (“Why a human has large brain”), Yury I. Aleksandrov (“Neurons, humans and cultures”), the program included the lecture of RAS academician, director of RAS Institute of Philosophy A.A.Guseinov “The Golden Rule of ethics”. Being inspired by lecture the first author of this article addressed the following question to the speaker, “Are you interested in a mathematical theory of the Golden rule?” The answer was confirmative. Now, at our strong belief, the concepts of Berge equilibrium most completely meet main requirement of the Golden Rule of Ethics, “Behave to others as you would like them to behave to you”. Thus, the article offered to the reader, first, suggests the method of construction of Berge equilibrium situation for finding minimax strategy in specific Germeier convolution, effectively constructed in assumed mathematical model of existence strategies if sets of strategies are compact and payoff function is continuous according to situations.

Текст научной работы на тему «Существование равновесия по Бержу»

УДК: 519.833 MSC2010: 91A10, 91В52

СУЩЕСТВОВАНИЕ РАВНОВЕСИЯ ПО БЕРЖУ © В. И. Жуковский

Московский государственный университет им. М. В. Ломоносова

факультет вычислительной математики и кибернетики Ленинские горы, Москва, ГСП-1, 119991, Российская Федерация e-mail: zhkvlad@yandex.ru

© Т. В. Макаркина

Государственный гуманитарно-технологический университет физико-математический факультет ул. Зеленая, д. 22., г. Орехово-Зуево, Московс. область, 142600, Российская Федерация

e-mail: tatmak147@yandex.ru

© Ю. А. Бельских

Государственный гуманитарно-технологический университет физико-математический факультет ул. Зеленая, д. 22., г. Орехово-Зуево, Московс. область, 142600, Российская Федерация

e-mail: fizmat@ggtu.ru

The Existence of Berge Equilibrium.

Zhukovskiy V. I., Makarkina T. V., Belskih Y. A.

Abstract.

The sufficient conditions of the existence of Berge equilibrium situation in noncooperative game of many persons in normal form are established. On the basis of these conditions the existence of Berge equilibrium situation in mixed strategies (by compact sets of strategies of players and continuity of their payoff functions) is proved. Let us consider the history of the appearance of the Berge equilibrium notion.

In 1949 the 21-years-old PhD student of Prinston University, John F. Nash (jun.), formalized the notion of "good" solution in noncooperative games (later called "Nash equilibrium"). It has got the broad spectrum of applications in economics, sociology, military sciences. And now after more than 50 years, in any journal of system analysis, game theory, mathematical programming we find the papers devoted to Nash equilibrium (NE). In 1994 John Nash won the Nobel Prize in economics in a common effort with John Harsanyi and R. Selten "for fundamental analysis of equilibria in noncooperative game theory". Actually 20-years-old Nash developed the foundation of the scientific method that played the great role in the development of world economy.

However "in the sun there are spots" (proverb). And the main of them is "the eqoistic character" of Nash equilibrium concept. It appears in the fact that every player tries to increase only his own payoff, i. e. follows "politica dei campanile", without considering interests

of other participants of the conflict. One of the methods to remove this negative is to use the approach (by formalization of "good" solution of the game), which differs from "dictated" Nash equilibrium. Such approach was proposed in 1994 at the scientific seminar (leader V.I.Zhukovskiy) at discussing the book of C.Berge «Theorie generale des jeux a n personnes games» (this book was published in Paris in 1957 and in 1961 it was translated into Russian [1]). Concretely the criticism of NE was caused by non-existence NE at strongly concave in strategy at least one player his payoff function (but the decision making is necessary!). The sense of the new approach lies in change of condition of solution stability not to deviation of the player whom belongs "payoff function" but to deviation of all players except the one who is "the owner" of this payoff function. We shall note three circumstances.

Firstly, we called the proposed new concept "BE". The term "BE" arouse as the result of reviewing Claude Berge's book. Secondly, in 1994 K.S.Vaisman (then the post-graduate student of V.Zhukovskiy) was engaged in construction of initial foundations of mathematical BE theory. In 1995 K. Vaisman defended his thesis "Berge equilibrium" (BE) in Leningrad University. (K.Vaisman died in 1998 at the age of 35 years). His sudden death suspended further development of the Berge equilibrium in Russia, but the notion of a Berge equilibrium was "exported from Russia" by Algerian scholars of V.Zhukovskiy M. Radjef and M. Larbani.

This notion caused the broad interest of our foreign colleagues. The acquaintance with their publications showed that "par le temps qui ceurt" the most papers of this direction devoted to the properties of Berge equilibrium, singularities, modifications of this notion, relations with Nash equilibrium. It is supposed that in originated theory of Berge equilibrium the stage of formation of strict mathematical theory becomes nearer. Probably an intensive accumulation of facts will be replaced by the stage of evolutionary internal development. At this stage one should traditionally answer two fundamental questions:

1. Does the Berge equilibrium exist?

2. How one should find this equilibrium?

The present article is just devoted to answers of these both questions. Thirdly, the authors were motivated by the IX Moscow Festival of Science that partially was held in a new building of MSU Fundamental library on October 10, 2014. Apart from lectures of Nobel laureates chemists Kurt Wuthrich (USA, California), Jean-Marie Lehn (France), biochemist Sir Richard Roberts (USA), RAS academician M.Ya.Marov ("The Chelyabinsk meteor"), L.M. Zelenyi ("Exoplanets: Searching for a second Earth"), Doctors of Sciences A.V. Markov ("Why a human has large brain"), Yury I. Aleksandrov ("Neurons, humans and cultures"), the program included the lecture of RAS academician, director of RAS Institute of Philosophy A.A.Guseinov "The Golden Rule of ethics". Being inspired by lecture the first author of this article addressed the following question to the speaker, "Are you interested in a mathematical theory of the Golden rule?" The answer was confirmative. Now, at our strong belief, the concepts of Berge equilibrium most completely meet main requirement of the Golden Rule of Ethics, "Behave to others as you would like them to behave to you".

Thus, the article offered to the reader, first, suggests the method of construction of Berge equilibrium situation for finding minimax strategy in specific Germeier convolution, effectively constructed in assumed mathematical model of existence strategies if sets of strategies are compact and payoff function is continuous according to situations.

Keywords : noncooperative game, payoff function, payoff, Nash and Berge equilibrium, Germeier convolution, mixed strategies.

Введение

Упорядоченная тройка

Г= (N, (Xi}i€N, {/¿(x)}^>,

образует нормальную форму бескоалиционной игры N лиц. В игре Г множество N = {1, 2,... , N} — порядковые номеров игроков, стратегии i-го игрока

Х G X* С , ситуация x = (xb x2,... , xN) G X, n = n*. Игроки, не имея возмож-

îgn

ности объединяться в коалиции, выбирают свои стратегии; на множестве X ситуаций определены функции выигрыша /¿(x) : X ^ R каждого игрока. Цель участия i-го игрока (на «содержательном уровне») — выбор своей стратегии так, чтобы его выигрыш (значение его функции выигрыша) стало возможно большим.

Далее используем общепринятое (в играх вида Г) обозначения: (x||zi) = (xi, . . . ,Xi-i,Zi,Xi+i, . . . ,Xn) для Zi G Xi.

Определение 1. Ситуация xe = (xi,... , x2) G X называется равновесной по Нэшу в игре Г (РН), если

max /¿(xe||xi) = /¿(xe), i G N. (1)

XiEXi

Определение 2. Ситуация xf = (xf,... , xN) G X называется равновесной по Бер-жу в игре Г, если

max /¿(x||xf) = /¿(xf), (i G N),

XN\{i}€XN\{i}

здесь

xn\{i} = (xi,... ,xi-i,xi+i,... ,xn) и Xn\{î} = Д Xj. (2)

j€n\{i}

Для игры трех лиц (N = {1, 2, 3}) требования (1) и (2) соответственно означают (3) и (4), именно

max /i(xi,x2,x3) = /i(xe),

xi€Xi

max /2(xi,x2,x3) = /"2(xe), (3)

X2GX2 4 y

max /3(xi,x2,x3) = /3(xe),

„ хз€Хз

max /1 (xf ,х2,хз) = /i(xB),

Х2,Х3

< max /2 (xi,xf ,хз) = /2(xf), (4)

Х1,хз 4 y

max /3 (xi,x2,xf) = /з(хв).

„ Х1,Х2

Перейдем к интерпретации равновесия по Бержу для семьи из трех человек: мужа (игрок I), жены (игрок II) и сына (игрок III). Например, для мужа — муж, забывая о своих интересах (первое равенство из (4)), направляет все свои усилия, чтобы помочь жене и сыну достичь наибольших выигрышей (второе и третье равенство из (4)). Те (жена и сын), в свою очередь, помогают мужу достичь максимально возможного успеха (первое равенство из (4)). Аналогичное положение для игрока 2 (жены) и игрока 3 (сына). Такой альтруизм при разрешении конфликтов типичен для христианства, ислама, иудаизма, конфуцианства. Такой подход к принятию решения заведомо исключает вооруженные столкновения, войны. Мир стал бы много лучше, если бы при уравновешивании конфликтов применялось бы равновесие по Бержу (а не по Нэшу!).

Достаточные условия

Для игры Г будем использовать новые переменные z Е X (i Е N) и z = (z1,... , zN) Е X, а также (x||zj) = (x1,... , xi-1, Zj, xi+1,... , xN) Е X. Составим с помощью этих обозначений N скалярных функций:

<^(x, z) = /j(x||Zj) - /j(Zj) (i Е N) (5)

и гермейеровскую свертку ^¿(x,z), именно,

w(x, z) = max Wj(x,z). (6)

i&i

Ограничимся случаем N = {1, 2, 3}, т. е. игрой Г трех лиц. Теперь рассмотрим вспомогательную антагонистическую игру

Г" = <{/,//}, {X, Z = X},p(x,z)>,

в которой игрок I за счет подходящего выбора своей стратегии x Е X стремится максимально увеличить платежную функцию <^(x,z), а игрок II с помощью выбора z Е Z = X - максимально уменьшить <^(x, z).

Бесспорным решением Г" является седловая точка (x0,zf) Е X х X, т.е.

max <^(x, zf) = ^(x0, zf) = min ^(x0, z), (7)

x€X z€X

причем x° является максиминной, а zB — минимаксной стратегией в Га, т. е.

min max ((x, z) = max ((x, zB), max min ((x, z) = min ((x°, z). zeXie^ xe^ i€X ze^ zex

Утверждение 1. Если в игре Га существует седловая точка (x°, zB), то минимаксная стратегия x° является равновесной по Бержу ситуацией в игре Г.

Доказательство. Из (7) следует справедливость цепочки неравенств:

((x, zB) < ((x°, zB) < ((x°, z) Vx, z G X. (8)

Из (5) и (6) при z = x° имеем ((x°, x°) = 0. Отсюда, согласно (8) (по транзитивности) получаем ((x°, zB) < 0 и поэтому ((x, zB) < 0 Vx G X. Тогда с учетом (6) и (5)

/i(x||zB) - /¿(zB) < 0 Vx G X (i G N),

т. е.

/¿(x||zf) < /i(zB) Vx G X (i G N).

Замечание 1. Из приведенного утверждения следует ясный и наглядный способ построения ситуации равновесия по Бержу исходной игры Г:

1) по функциям выигрыша /¿(x) из Г построить с помощью (5) скалярные функции (¿(x,z), а затем с помощью (6) выписать гермейеровскую свертку ((x,z);

2) найти седловую точку (x°,zB) функции ((x,z) из (6).

Тогда минимаксная стратегия zB является искомой ситуацией равновесия по Бер-жу игры Г.

Пожалуй, наибольшая сложность здесь возникает при построении седловой точки негладкой функции ((x,z). Здесь уже следует привлечь к построению седловой точки негладкий анализ, разрабатываемый в России недавно скончавшемся ленинградским профессором Владимиром Федоровичем Демьяновым.

Существование в смешанных стратегиях

Нужно обладать большим оптимизмом, чтобы надеяться найти для игры Г равновесную по Бержу ситуацию в чистых стратегиях х^ С X г Е N для трех и более игроков. Поэтому следуя подходу Эмиля Бореля [8], Джона фон Неймана [9], Джона Нэша [4], [5] и их последователей, установим существование ситуации равновесия по Бержу в смешанных стратегиях. Причем будем следовать подходу к решению аналогичной задачи, предложенной первым автором в [6].

Напомним, что здесь и далее через comp обозначаем множество всех компактов (замкнутых и ограниченных подмножеств из евклидова п^-мерного пространства ), непрерывность на X скалярной функции /¿(ж) обозначаем /¿(-) G C(X). Не оговаривая особо, для элементов игры Г = (N, {X¿}¿eN, {/¿(x)}¿eN) предполагаем выполнение следующих требований:

X¿ G comp Rn, /¿О G C(X), i G N = {1, 2, 3}. (9)

Перейдем к понятию смешанного расширения игры Г, включающее смешанные стратегии, ситуации, математическое ожидание функции выигрыша. Будем предполагать, что для игры Г выполнены ограничения (9), тогда /¿(ж) непрерывна на

компакте X, где X = П X¿. На каждом компакте X¿ С (i G N) построим бо-

¿en

релевскую a-алгебру B(X¿) — множество подмножеств X¿ таких, что X¿ G B(X¿), причем B(X¿) замкнута относительно операций дополнения и объединения счетного числа множеств из B(X¿), кроме того, B(X¿) является минимальной a-алгеброй, которая содержит все замкнутые подмножества компакта X¿. Согласно математической теории игр, смешанную стратегию i-го игрока v¿(-) будем отождествлять с вероятностной мерой на компакте X¿. Вероятностная мера есть неотрицательная скалярная функция v¿(-), определенная на борелевской a-алгебре B(X¿) подмножеств компакта X¿ С и удовлетворяющая двум условиям:

(1) v¿ ^U = U для любой последовательности {Q^^i попарно не

пересекающихся элементов из B(X¿) (свойство счетной аддитивности функции v¿(^));

(2) v¿(X¿) = 1 (свойство нормированности) и поэтому v¿(Q(i)) < 1 для всех Q« G B(X¿).

Обозначим через {v¿} множество смешанных стратегий i-го игрока (i G N). Построим ситуацию в смешанных стратегиях в виде меры-произведения

v (dx) = v1(dx1)v2(dx2)v3(dx3),

множество которых обозначим через {v}, а также математическое ожидание

ЛИ = / fi[x]V(dx).

X

Получаем смешанное расширение игры Г, которое обозначим через

Г = (N = {1, 2, 3}, {v¿}¿en, {/¿[V] = J /¿[x]v(dx)}^). (10)

X

iНе можете найти то, что вам нужно? Попробуйте сервис подбора литературы.

Будем использовать также меры-произведения

(v||^¿) = Vi(dxi)... V¿-1 (dx¿-i)^¿(dz¿)v¿+i(dx¿+i),..., vw(dxw),

где вероятностные меры ^¿(dzj) G {^}, а новые чистые стратегии zj G Xj (i G N). Аналогично определению 2 введем

Определение 3. Ситуацию в смешанных стратегиях vB(•) G {v} назовем равновесной по Бержу в смешанном расширении (10) (или равновесной по Бержу ситуацией в смешанных стратегиях для игры Г), если

max /¿(v|KB) = /¿(vB) (i G N). v(-)eM

Наконец, приведем утверждение, доказанное в [8].

Утверждение 2. Если в игре Г множества Xj G comp и /¿[-] G C(X) (i G N), то для функции

( = max (r (x, z)

r—1,2,3

имеет место неравенство

max / (r(x, z)^(dz)v(dx) < / max (r(x,z)^(dz)v(dx)

r—1,2,3 / / r—1,2,3

Ixl X x!

при любых G {v}, v(•) G {v}; здесь, напомним, скалярные функции фг(x,z) определены в (5) и (6) (этот результат аналогичен свойству: максимум суммы не больше суммы максимумов).

Перейдем к доказательству центрального результата настоящей статьи: установим существование равновесной по Бержу ситуации в смешанных стратегиях в игре Г при выполнении условий (9).

Теорема 1. Если в игре Г множества Xj G comp Rni и /¿[^ G C(X) (i G N), то в этой игре существует равновесная по Бержу ситуация в смешанных стратегиях.

Доказательство. Как и в утверждении 1 рассмотрим вспомогательную антагонистическую игру

Га = <{/,//}, {X, Z = X},((x,z)>.

В игре Га множество X стратегий x первого I (максимизирующего ((x, z)) игрока совпадает с тем же X. Одним из решение Га является седловая точка (x°, zB) G X х X, именно, напомним: для нее при всех x G X и каждом z G X имеет место цепочка неравенств

((x,zB) < ((x°,zB) < ((x°,z). Теперь игре Га поставим в соответствие ее смешанное расширение

Гa = <{/,//}, {v}, {^},0(v,^)>.

где {V} — множество смешанных стратегий V(•) первого I, а = {V} — множество смешанных стратегий второго игрока II, функция выигрыша первого I (математическое ожидание)

= J <^(х,г^(^ж)^(^г). (11)

X хХ

Отметим, что в силу [7] и (6), (5), (9) функция <^(х, г) из (6) непрерывна на X. Решением игры Га (смешанного расширения Га) также будет седловая точка (V0, ), определяемая двумя последовательными неравенствами

) < ) < (12)

при любых V(•) е {V}, е {V}.

Эту пару (V°, ) иногда называют решением игры Га в смешанных стратегиях. В 1952 г. Ирвинг Гликсберг установил [8] теорему существования равновесной по Нэшу ситуации бескоалиционной игры N > 2 лиц в смешанных стратегиях, откуда (для частного случая — антагонистической игры Га) следует утверждение: пусть в игре Га множество X С Кга суть непустой компакт, а функция выигрыша первого (I) игрока <^(х, г) непрерывна на X х X. Тогда для игры Га существует решение (Vе, ), определенное в (12), то есть существует седловая точка в смешанных стратегиях. С учетом (11) неравенства (12) примут вид

max (x,z)v(dx)^B(dz) < / max <^(x, z)v°(dx)^B(dz) <

r—1,2,3 / r—1,2,3

X xX XxX

< / max (x, z)v°(dx)^(dz)

Xx X

при всех v(■) G {v}, G {v}. Положив в

<^(v°,^) = / max (x, z)v°(dx)^(dz)

/ r—1,2,3

Xx X

меру ^j(dzj) = v°(dxj) (i G N) (и тогда ^(dz) = v°(dx)), получаем, с учетом (12), что <^(v°, v°) = 0. Аналогично приходим к <^(v°, v°) = 0 и тогда из (12) имеем

^(v0,^ ) = 0. (13)

Согласно ) = 0 и неравенства в (13) (по транзитивности) приходим к

<^(v, ) = max (x,z)v(dx)^B(dz) < 0 Vv(■) G {v}.

I r—1,2,3 XX

Применяя затем утверждение 2, получаем

0 > J шах^^г-(^г) > шах^ У (^ж)^в(^г).

ХхХ ХхХ

Поэтому для всех г = 1,... , N будет

У (¿г) < 0 (■) е {V}.

ХхХ

Отсюда при г = 1, 2, 3 и согласно (5), а также нормированности V(■), приходим, например, при г = 1

0 > <£i(x,z)v(dx)^B(dz) = / max{/i(zi,x2, x3) — /i(z)}v(dx)^B(dz) > / / ien

^ (dz)= J tr

XxX Xx!

.BfJ~\ I t I . B t

> J /1(z1,x2,x3)v(dx)^ (dz) — J /1(z)^ (dz) у v (dx) =

Xx X X X

= /i^B ,V2,V3) — /i(^B ).

Аналогично устанавливается справедливость еще трех неравенств при r = 2, 3

0 > /2(vi,^B,V3) — /2(^B)

0 > /3(vi,V2,^B) — /3^B).

Откуда, в силу определения 3, следует равновесие по Бержу (•) G {v} в игре (10). □

Заключение

В работе получены следующие результаты:

Установлены достаточные условия существования равновесной по Бержу ситуации в бескоалиционной игре трех лиц в нормальной форме. На основе этих условий доказано существование ситуации равновесия по Бержу в смешанных стратегиях (при компактных множествах стратегий игроков и непрерывности их функций выигрыша).

Описок литературы

1. Berge, C. Théorie générale des jeux a n personnes / Berge, C. — Paris: Gauthier-Villar, 1957. — 114 c.

Берж, C. (1961) Общая теория игр нескольких лиц. М.: Физматгиз.

2. Borel, E. La théorie du jeu et les equations inténgrales a noyau symétrique // Comptes Rendus de l'Académie des Sciences. - 1921. - Vol. 173. - C. 1304-1308.

3. Von Neumann, J. Theorie der Gesellschaftspiele // Math. Ann. - 1928. - Vol. 100. - C. 295-320.

4. Nash, J. Non-cooperative games // Math. Ann. - 1951. - Vol. 54. - C. 286-295.

5. Nash, J. Equilibrium point in N-person games // Proc. Nat. Academ. Sci. USA. - 1950. - Vol. 36. -C. 48-49.

6. Zhukovskiy, V. I. and Kudryavtsev, K. N. Mathematical Foundations of the Golden Rule. I. Static Setting // Automation and Remote Control. Pleiades Publishing. Ltd. - 2017. - Vol. 78, № 10. -C.1920-1940.

7. Морозов, В. В., Сухарев, А. Г., Федоров, В. В. Исследование операций в задачах и упражнениях. - М.: Наука, 1986. - 285 c.

MOROZOV, V. and SUHAREV, A. and FEDOROV, V. (1986) Operations research in problems and exercises. Moscow: Nauka.

8. Glicksberg, I. L. A further generalization of the Kakutani fixed point theorem, with application to Nash equilibrium points // Proc. Amer. Math. Soc. - 1952. - Vol. 3, № 1. - C. 170 -174.

9. Zhukovskiy, V. I. and Kudryavtsev, K. N. Pareto-Equilibrium Strategy Strategies Profile: sufficient Conditions and Existence in mixed Strategies // Automation and Remote Control. Pleiades Publishing. Ltd. - 2016. - Vol. 77, № 8. - C. 1500 - 1510.

i Надоели баннеры? Вы всегда можете отключить рекламу.