李培中,何中全
(西華師范大學(xué)數(shù)學(xué)與信息學(xué)院,四川南充 637009)
凹映射Nash均衡的存在定理
李培中,何中全
(西華師范大學(xué)數(shù)學(xué)與信息學(xué)院,四川南充 637009)
以通常的Nash均衡為特例,在此基礎(chǔ)上,引入集值映射的Nash均衡概念,從而得到在凹映射和緊值條件下集值映射Nash均衡的存在定理.
Nash均衡;集值映射;凹映射;上半連續(xù)映射
博弈論屬應(yīng)用數(shù)學(xué)的一個(gè)分支,主要研究公式化了的激勵結(jié)構(gòu)間的相互作用.在1950 – 1951年期間,John Nash利用不動點(diǎn)定理證明了均衡點(diǎn)的存在,給出了n人非合作博弈的均衡存在定理[1-2].此后,博弈論不斷發(fā)展,得到廣泛應(yīng)用.
圍繞著Nash均衡點(diǎn)的存在性問題,很多學(xué)者進(jìn)行了大量的研究.例如,文獻(xiàn)[3-5]討論了局中人是有限且支付函數(shù)是單值函數(shù)的Nash均衡的存在性;文獻(xiàn)[6-9]討論了局中人是有限且支付映射為泛函的對策的Nash均衡的存在性;文獻(xiàn)[10]討論了局中人為任意多且支付映射為集值映射的對策系統(tǒng)的Loose Nash均衡的存在性;文獻(xiàn)[11]討論了支付映射為泛函的對策的Nash均衡點(diǎn)集的穩(wěn)定性.本文將在凹映射條件下,給出集值映射Nash均衡存在定理.
設(shè)X和Y是兩個(gè)拓?fù)淇臻g,2Y表示Y的非空子集的全體,K(Y)表示Y的所有非空緊子集,K(Y)和2Y均賦有Vietoris拓?fù)鋄8].
定義1 1)稱集值映射T:X→2Y在x∈X處是上半連續(xù)的,如果對Y中任意開集O且T(x)?O ,存在x的開鄰域N(x)?X ,使對任意x′∈N(x),有T(x′)?O;稱T在X上為上半連續(xù)的,如果T在任意x∈X處為上半連續(xù)的.
2)稱T在x∈X處是下半連續(xù)的,如果對Y中任意開集O且O∩Tx≠φ,存在x的開鄰域N(x)?X ,使對任意x′∈N(x),有T(x′)∩O≠φ;稱T在X上為下半連續(xù)的,如果T在任意x∈X處為下半連續(xù)的.
3)稱T在X上是連續(xù)的,如果T既是下半連續(xù)又是上半連續(xù)的.
定義2 設(shè)Y是實(shí)數(shù)域上的拓?fù)渚€性空間,稱C?Y為錐,如果對任意c∈C和任意非負(fù)數(shù)t∈[0,+∞),有tc∈C;稱C為凸集,如果對任意c,d∈C,集合{tc+(1?t)d:0≤t≤1}在C中.設(shè)C是Y中的一個(gè)凸錐,稱C是尖錐,如果C∩(?C)={0}.
對集合A,B?Y,數(shù)α∈R,記A+B={a+b :a∈A,b∈B},αA={αa:a ∈A}.
引理1[12]設(shè)X和Y是兩個(gè)拓?fù)淇臻g,F(xiàn):X→2Y,若X為緊集,F(xiàn)上半連續(xù)且為緊值,則F(X)=∪x∈XF(x)為緊的.
引理2[11]設(shè)X為拓?fù)淇臻g,Y為正則拓?fù)淇臻g,F(xiàn):X→2Y,若F在x0∈X處上半連續(xù)且為閉值,則對任意xn→x0,對任意yn∈F(xn)且yn→y0,有y0∈F(x0).
引理3[13]設(shè){Aα}α∈D是K(X)中的一個(gè)網(wǎng),且Aα→A ∈K(X)(在Vietoris拓?fù)鋄8]意義下),則對任意網(wǎng){xα:α∈D},其中對每個(gè)α∈D,xα∈Aα{xα:α∈D}有聚點(diǎn)x*∈A.
本文將討論下面的集值映射Nash均衡問題.
設(shè)指標(biāo)集I是局中人的集合,對每個(gè)i∈I,Xi是局中人i的策略集,Ei是拓?fù)湎蛄靠臻g,記
則x*就是通常的Nash均衡點(diǎn).
下面引入集值映射的凹映射定義.
定義3 設(shè)T:X →2Y是一個(gè)集值映射,我們稱T是凹映射,如果?y1,y2∈X ,λ∈(0,1),T(x,λy1+(1?λ)y2)?λT (x,y1)+(1?λ)T(x,y2).
另一方面,由F在x0的上半連續(xù)性可知,存在x0的一個(gè)鄰域N(x),使得對所有的x∈N(x0),有F(x)?T.
但xn→x0意味著存在一個(gè)指標(biāo)i*,使得當(dāng)i≥i*時(shí),xi∈N(x0),隨之,F(xiàn)(xi)?T .這樣的話,當(dāng)i≥max{i0,i*}時(shí),得到y(tǒng)i∈F(xi)?T .這樣就出現(xiàn)一個(gè)矛盾.
由引理4可得集值映射的Nash平衡點(diǎn)的存在定理.
i)fi上半連續(xù)且具有非空緊值;
ii)對任意ui∈X ,fi(?,ui)是緊值的上半連續(xù)映射;
iii)對任意x∈X,Ci(x)是閉凸尖錐且intCi(x)≠φ;
iv)對任意x∈X,存在yi∈fi(x),使(y?fi(x))??intCi(x);
v)對任意x∈X,fi(i,?)是凹映射;
證明:對每個(gè)i∈I,作映射Ki:Xi→2X如下:
下面要證明Ki(ui)是閉集.
[1] Nash J. Equilibrium point in N-person games [J]. Proceedings of the National Academy of Sciences of the United States of America, 1950, 36(1): 48-49.
[2] Nash J. Non-cooperative games [J]. Mathematische Annalen, 1951, 54: 286-295.
[3] Nikaido H, Isoda K. Note on non-cooperative convex games [J]. Pacific Journal of Mathematics, 1995, 5: 807-815.
[4] Nishimura K, Friedman J. Existence of Nash equilibrium in n-person games without quasi-concavity [J]. International Economic Review, 1981, 22: 637-648.
[5] Kim W K, Kum S. On the existence of Nash equilibria in N-person games with C-concavity [J]. Computers and Mathematics with Applications, 2002, 44: 1219-1228.
[6] Aubin J P. Mathematical methods of game and economic theory [M]. Amsterdam: North Holland, 1982: 145- 156.
[7] Aubin J P. Optima and equilibria an introduction to nonlinear analysis [M]. New York: Springer-Verlag, 1993: 234-245.
[8] Tan K K, Yu J, Yuan X Z. Existence theorems of Nash equilibria for aoacoopsrative N-person games [J]. International Journal of Game Theory, 1995, 24: 217-222.
[9] Guillerme J. Nash equilibrium for set-valued maps [J]. Journal of Mathematical Analysis and Applications, 1994, 187:705-715.
[10] Yu J. Essential equilibria of n-person noncooperative games [J]. Journal of Mathematical Economics, 1999, 31: 361-375.
[11] Klein E, Thomopson A C. Theory of correspondences [M]. New York: John Wiley and Sons, 1984: 134-145.
[12] Aubin J P, Ekeland I. Applied nonlinear analysis [M]. New York: John Wiley and Sons, 1984: 218-226.
[13] Beer G. On a generic optimization theory of Kenderov [J]. Nonlinear Analysis: Theory, Methods and Applications, 1988, 12: 674-655.
Existence Theorems of Nash Equilibrium for Concave Mappings
LI Peizhong, HE Zhongquan
(College of Mathematics and Information, China West Normal University, Nanchong, China 637009)
The concept of Nash Equilibrium for set- valued mappings was introduced by taking common Nash Equilibrium as a specific example. Then, the existence theorems of Nash Equilibrium for set-valued mappings under concave mappings and compact-valued assumptions were obtained.
Nash Equilibrium; Set-valued Mapping; Concave Mapping; Upper Semicontinuous Mapping
O175
A
1674-3563(2010)02-0001-05
10.3875/j.issn.1674-3563.2010.02.001 本文的PDF文件可以從xuebao.wzu.edu.cn獲得
(編輯:王一芳)
2009-06-09
李培中(1982- ),男,貴州遵義人,碩士研究生,研究方向:非線性分析