文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > Finite Time--Horizon Risk Sensitive Control and the Robust Limit under a Quadratic Growth A
侵权投诉

Finite Time--Horizon Risk Sensitive Control and the Robust Limit under a Quadratic Growth A

The finite time--horizon risk sensitive limit problem for continuous, nonlinear systems is considered. Previous results are extended to cover more typical examples. In particular, the cost may grow quadratically, and the diffusion coefficient may depend on

Also, let

: Ws= fFt{progressively measurable w such that kwkL ( s;T]: Wsm= fw 2 Ws: jwt(!)j m 8t 2 s; T]; ! 2 g: 2

)< 1g

(19) (20)

Let the dynamics of the modi ed game be ";b";b";b dyt= f b (yt; ut)+ (yt )wt] dt+";b ys= x; r

" (y";b ) dB t 2 t

(21) )

and de ne the payo by ";b;R Jg (s; x; u; w)= E

(Z

T

s

";b`R (yt; ut)?

2 R";b 2 jwtj dt+ (yT ):

2

(22)

An admissible strategy for the player with control u is a mapping: Ws ! Us such that if, almost surely, wr= wr for a.e. r 2 s; t], then, almost surely, w]r= w]r for a.e. r 2 s; t], for any t 2 s; T] e e b s . De ne the set of admissible strategies 19]. The set of all admissible strategies will be denoted by for the player with control w analogously, and denote them by b s . Then (se

e 19]), the upper and lower values are ";b;R V 0;";b;R;m(s; x)= inf supm Jg (s; x; w]; w) 2b w2W s s

(23) (24)

V 0;";b;R;m(s; x)= sup uinf Jg";b;R(s; x; u; u]): 2bs 2Us

We further restrict these strategies to observable admissible strategies as follows. An observable admissible strategy for the player with control u is a mapping: Ws ! Us such that if, wr= wr for e a.e. (r; !) 2 s; t] A, then, w]r= w]r for a.e. (r; !) 2 s; t] A, for any t 2 s; T] and A 2 Ft. e The set of all observable admissible strategies will be denoted by s . De ne the set of observable admissible strategies for the player with control w analogously, and denote them by s . Conceptually, the observability restriction requires that if one alters w only in some portion of the sample space, then the corresponding control played by the strategy,, should not change in a di erent portion of the sample space. Now, note that if wr= wr a.e. on s; t] B for B 2 FT, then wr= wr a.e. on s; t] A where e e: A= inf fC 2 Ft: B Cg. Therefore, for any 2, w]r= w]r a.e. in s; t] A s; t] B. e Consequently, one may replace Ft by FT in the de nition of observable admissible strategy. De ne the upper and lower values as V";b;R;m (s; x)= inf supm Jg";b;R(s; x; w]; w) 2 s w2Ws

(25) (26)

V";b;R;m (s; x)= sup uinf Jg";b;R(s; x; u; u]): 2 s 2Us

(Obviously, one has V 0;";b;R;m V";b;R;m and V";b;R;m V 0;";b;R;m.) Then, by 19], V 0;";b;R;m= V 0;";b;R;m is the unique, continuous viscosity solution to Isaacs equation (27). Further, by trivial modi cation of the proofs in 19] (see appendix), V";b;R;m= V";b;R;m is the unique, continuous viscosity solution to Isaacs equation, and consequently, all the values are equal.

上一页第8页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email