查询词典 stochastic dynamic programming

This paper is mainly the dynamic input-output model that the time lag is one, which is base on the above models. After studying, we consider stochastic factor step by step in it, namely when consumption coefficient matrix is stochastic (when investment matrix is stochastic, it is almost same. So we dont research it), and they are both stochastic, then we research the stable increase solution. We utilize the means of the modern stochastic analysis and Markov process, that the stochastic dynamic input-output model don not exist the stable solution is proved. Namely, economic system must is adjusted constantly. The probability that the collapse time of the economic system is o is one.

本文对在上述基础上构造的一类时滞为1的动态投入产出模型，进行了深入研究，将随机因素逐步考虑进去，即对投入产出消耗系数矩阵为随机的情况(投资系数矩阵为随机的情况与投入产出消耗系数矩阵为随机的情况大致相同，这里就不再证明)，以及二者同时为随机矩阵时所得到的动态投入产出模型的稳定增长解问题，利用现代概率分析及马氏过程的工具，证明了不存在随机动态投入产出模型的稳定增长解；即投入产出模型反映的经济系统必须经常进行调整，其崩溃时间为无穷大的概率为零。

Firstly, an empirical approximation model of stochastic programming is obtained by replacing the probability measure of original program with empirical probability measure. Sequentially, the constrained stochastic programming is transformed into an equivalent unconstrained stochastic programming. Finally, using the epi-convergence theory, the almost everywhere upper semiconvergence of optimal solution set of empirical approximations for stochastic programming is obtained.

首先通过经验概率测度替代初始规划的概率测度得到随机规划的经验逼近模型，然后将带有约束的随机规划问题转化成与其等价的无约束的随机规划问题，最后利用上图收敛性理论，给出了随机规划经验逼近最优解集的几乎处处上半收敛性。

Quasi Hamiltonian system; nonlinear stochastic optimal control; robustness; robust control; parametric uncertainty; uncertain disturbance; Bouc-Wen hysteretic system; Preisach hysteretic system; minimax optimal control; stochastic stabilization; stochastic averaging method; stochastic dynamical programming principle; stochastic differential game; maximal Lyapunov exponent

国家自然科学基金；拟Hamilton系统；非线性随机最优控制；鲁棒性；鲁棒控制；参数不确定性；不确定扰动； Bouc-Wen滞迟系统； Preisach滞迟系统；极小极大最优控制；随机稳定化；随机平均法；随机动态规划原理；随机微分对策；最大Lyapunov指数

By employing the local Lipschitz condition and Picard sequence, the local existence-uniqueness of solutions of stochastic functional differential equations of Ito-type is firstly obtained. Furthermore, a continuation theorem for stochastic functional differential equations of Ito-type is given by using stochastic analysis technique and the quasi-boundedness condition. Finally, by establishing some delay differential inequalities and using properties of H_m-functions, a stochastic version of Wintner theorem and the global existence-uniqueness of solutions of stochastic functional differential equations of Ito-type are given. The results generalize the earlier publications.

首先，利用局部Lipschitz条件和Picard序列，获得了伊藤随机泛函微分方程解的局部存在唯一性；其次，利用随机分析技巧和拟有界条件，建立了伊藤随机泛函微分方程解的延拓定理；最后，通过建立一些时滞微分不等式和利用H_m-函数的特性，得到了Wintner定理的随机版本和伊藤随机泛函微分方程解的全局存在唯一性，推广了已有的一些结果。

Based on the general dynamic programming, in the grey dynamic programming model combined with the grey theory, the runoff is described by the interval grey numbers. The model is combined by the advantages of determinant/stochastic dynamic programming methods, it will be a new idea for the reservoir operation.

利用区间灰数描述径流，在普通动态规划的基础上融合了灰理论的灰色动态规划模型，综合了确定型动态规划方法和随机型动态规划方法的优点，为水库调度的动态规划应用研究提供了新的思路。

This is the Massachusetts Institute of Technology on Dynamic Programming and Stochastic Control in courseware, want to learn dynamic programming and stochastic control to all of us help.

详细说明：这是美国麻省理工学院的有关动态规划与随机控制方面的课件，希望对学习动态规划和随机控制方面对大家有所帮助。

This book discusses linear and nonlinear programming, dynamic programming, integer programming and stochastic programming.

本书讨论了线性和非线性规划，动态规划，整数规划和随机规划。

In this paper, based on the conjugate function and the dual programming of convex programming, with the dual cone, to conic programming, the dual programming is fully discussed, the expression of the dual programming is educed, the main dualities are proved, and the dualities of the familiar conic programming are studied.

以共轭函数和凸规划的对偶规划为基础，利用对偶锥的概念，全面讨论了一般锥规划的对偶问题，严格推导出锥规划对偶规划的表示形式，给出了锥规划的主要对偶性质，并用这些结果研究了常见锥规划的对偶性。

According to the statistic analysis of Prof. Amari on static recognition with neural networks, and introducing the concept of vector position stochastic variable of vector stochastic sequence, the stochastic variable is processed by DRNN. The relationship between input and output variance is analyzed. The stochastic analysis of dynamic identification is given, so as to explain the DRNN characteristics in recognition process.

第二，根据Amari教授对神经网络静态识别时的统计分析，引入矢量随机序列的矢量位置随机变量的概念，以该随机变量经过DRNN处理后，分析其输入与输出方差之间的关系，试给出了DRNN做动态识别时的统计分析，以从理论上说明DRNN在识别过程中的特性。

This course offers an introduction to primary convex analysis, linear programming, non-linear programming, dynamic programming, multi-objective programming, multi-level programming, calculus of variations, and optimal control, which are the optimization methods commonly used in economics and finance.

课程内容简介：本课程介绍凸分析基础、线性规划、非线性规划、动态规划、多目标规划、多层规划、变分法、最优控制，这些都是西方经济学、金融学中常用的最优化方法。

"Bird Man's Diary, Butting on the Love Island", Changsha, China

2008年"鸟人日记，顶在爱情岛"，中国长沙

As long as they were in the service of a family of wizards, their magic would remain dormant.

只要他们一天为家庭服务，他们的法力就潜藏不动。