IV 5 : 2013 11 25
1. :. ( ) etc. etc.
1. :
: :
( ) etc. :
1. 2. 2
: 1. 2. : (agents): i N {1, 2,, n}: n( 1) x X: (alternatives) : X
( ) θ i Θ i : i : θ (θ 1, θ 2,, θ n ): (state). Θ Θ 1 Θ 2 Θ n θ: ψ( ): Θ ψ(θ): θ : v i : X Θ i R R v i (x, θ i, m i ) (1) u i (x, θ i ): θ i x t i R: i. t i > 0: t i 0:
( ) y (x, t 1, t 2,, t n ) Y : i v i (y, θ i ) f : Θ Y : (social choice function) θ y : f(θ) = ( x(θ), t 1 (θ), t 2 (θ),, t n (θ) ). (2) x(θ) X: (t 1 (θ), t 2 (θ),, t n (θ)): f( )
( ) Γ ( M 1, M 2,, M n, g( ) ) : M i : i i m i M i Γ i M M 1 M 2 M n. g : M Y : m i θ i g( )
( ) s i : Θ i M i ; Γ i s(θ) (s 1 (θ 1 ), s 2 (θ 2 ),, s n (θ n )) : Γ i 1 s ( ) Γ : (3)
( ) Γ f( ) 2 θ Θ Γ s ( ) (implements) Γ BNE BNE Γ f( )
µ µ µ Å
M g( ) f( ) 3 Γ D = (Θ 1, Θ 2,, Θ n, f( )) F ( ) i N θ Θ 1. 2.
(Bayesian incentive compatibility) 4 i N θ i Θ i s i (θ i) = θ i Γ D : (4) BIC BNE Γ D f( ) BIC
(revelation principle) BIC f( ) BIC BIC f( ) 1 (Myerson, 1979) f( ) BNE f( ) BIC
( ) BIC f( ) BNE BNE : : f( ) 1 : BIC f( )
1. (full implementation) F ( ) E(θ, M, g): θ F (θ) = g(e(θ, M, g)) : Maskin (1977, 1999). : Dasgupta, Hammond and Maskin (1979). : BNE Palfrey and Srivastava (1987, 1993) Jackson (1991).
( ) 1. ( ) ( ) : SPE Moore and Repullo (1988). Abreu and Sen (1990). (virtual implementation) F ( ) Matsushima (1988), Abreu and Sen (1991). Jackson (1991). (robust implementation) Bergemann and Morris (2005, 2007).
( ) 2. ( ) f( ) OK! f( ) 5, f( ) : (5)
2007 Hurwicz (1960): Maskin (1977, 1999): Myerson (1981):
M g Γ = (M, g) f( ) M i = Θ i, g = f BIC: