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

0= Vs+ 2" 2

max ai;j (x)Vxi xj+ min jwj m f b (x; u)+ (x)w] rV+`R (x; u)? 2 jwj2 u2U i;j=1 V (T; x)= R (x):

n X

2

(27)

Lemma 3.1 V";b;R;m= V";b;R;m is the unique, C 1;2 solution to (27). Proof.>From above, we know it is the unique, continuous viscosity solution. Additionally, 19] tells us that it is bounded, Lipschitz in x and Holder continuous in s. (We will be interested in speci c bounds independent of"; b; R; m further below, but for simplicity we do not look for these here.) Under our assumptions, it is not di cult to show that it is also Lipschitz in s, and we do so now. Fix s; s+ 2 0; T],> 0. We will bound jV";b;R;m (s; x)? V";b;R;m(s+; x)j. Fix any x 2<n, u 2 Us m and w 2 Ws . Let the corresponding solution of (21) be denoted by y0;";b . Recall the probability space is (; fFtg; P). De ne a new Brownian motion B by Bt= Bt? on a new probability space with ltration Ft0= Ft? . Let ut= ut? and wt= wt? . Let the solution of (21) corresponding to this new Brownian motion, controls u; w and initial condition x at time s+ be denoted by y;";b . Then;";b yt;";b= yt0? 8 t 2; T+]: Consequently, one has

Jg";b;R(s+; x; u; w )= E Therefore,

"Z

T?

s

0`R (yt0;";b; ut)? 2 kwtk dt+ R (yT;";b):?

2

#

Jg";b;R(s+; x; u; w

)? Jg";b;R (s; x; u; w) "

=E which by (A4:v) 2

R (y0;";b )? T?

Z

T

?

T?

2 0 0`R (yt;";b; ut)? 2 kwtk dt? R (yT;";b )

#

0 0 (R+ 2 m2 )+ MR EjyT;";b? yT;";b j?"Z# T 2 0;";b; u )+ (y0;";b )w j dt (R+ 2 m2 )+ MR E jf b (yt t t t T?

>From this (and an analogous bound on the other side using approximately optimal w), one easily nds that V";b;R;m is Lipschitz in s. Also, by the above Lipschitz properties, it is seen that 2:?(s; x)= min jwj m f b (x; u)+ (x)w] rV+`R (x; u)? 2 jwj2 max u2U is bounded in L1 . Then, by 17], V";b;R;m is, in fact, a C 1;2 solution of (27). 9

which by de nition of f b and (A2:iii) 2 (R+ 2 m2 )+ MR (b+ Mm):

2

上一页第9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