Studies on a robust structure of optimal fuzzy decision processes

最优模糊决策过程的鲁棒结构研究

• 批准号: 21540112
• 负责人: NAKAGAMI Jun-ichi
• 金额: $ 1.66万
• 依托单位: Chiba University
• 依托单位国家: 日本
• 项目类别: Grant-in-Aid for Scientific Research (C)
• 财政年份: 2009
• 资助国家: 日本
• 起止时间: 2009 至 2011
• 项目状态: 已结题
• 来源: https://kaken.nii.ac.jp/grant/KAKENHI-PROJECT-21540112/
• 关键词: 多段決定過程; マルコフ決定過程; 不完全情報; ファジー最適化; 頑健性; 凸解析; 区間解析; 双対定理

Our research objectives are initiated with a further reconsideration of the practical examples which deal with a machine maintenance problem in our two papers (2003,2006).The 1st objective is to construct a new fuzzy duality theorem for our fuzzy valued Markov decision processes in the same way as ordinary scalar valued ones between the solutions of an optimal equation and of a linear programming problem. This result is summarized and prepared for a contribution to suitable journals. The 2nd objective is to discuss an inverse problem of the optimality by the detailed calculations of a machine maintenance problem in our two papers. It means that if some particular optimal policy would be also optimal in a certain measure of fluctuations. This result is derived in the individual problem separately, we are now trying to show this results in the general situations.

我们的研究目标是从进一步重新考虑我们的两篇论文（2003，2006）中处理机器维护问题的实际例子开始的。第一个目标是为我们的模糊值马尔可夫决策过程构建一个新的模糊对偶定理。与最优方程和线性规划问题的解之间的普通标量值相同。对该结果进行了总结并准备向合适的期刊投稿。第二个目标是通过我们两篇论文中机器维护问题的详细计算来讨论最优性的逆问题。这意味着某些特定的最优政策在一定程度的波动下也是最优的。这个结果是在个别问题中单独得出的，我们现在尝试在一般情况下展示这个结果。

项目成果

Weighted quasi-arithmetic means and conditional expectations in Lecture Notes in Artificial Intelligence 6408 'Modeling Decisions for Artificial Intelligence ; MDAI2010'

人工智能讲义 6408“人工智能决策建模”中的加权准算术平均值和条件期望；

• 发表时间: 2010
• 作者: Y. Yoshida; V. Torra; Y. Narukawa M. Daumas
• 通讯作者: Y. Narukawa M. Daumas

An interval Bayesian Method for Uncertain MDPs

不确定MDP的区间贝叶斯方法

• 发表时间: 2009
• 作者: T. Iki; M. Horiguchi; M. Kurano M. Yasuda
• 通讯作者: M. Kurano M. Yasuda

An Optimal Value-at-Risk Portfolio under Uncertainty

不确定性下的最优风险价值投资组合

• 发表时间: 2009
• 作者: 吉田祐治

The expected total cost criterion for Markov decision processes under constraints: a convex analytic approach

约束下马尔可夫决策过程的预期总成本准则：凸分析方法

• 发表时间: 2012
• 影响因子: 1.2
• 作者: F. Dufour; M. Horiguchi A. B. Piunovskiy
• 通讯作者: M. Horiguchi A. B. Piunovskiy

An Average Value-at-Risk Portfolio Model under Uncertainty: A perception-Based Approach by Fuzzy Random Variables

不确定性下的平均风险价值投资组合模型：基于感知的模糊随机变量方法

• 发表时间: 2011
• 影响因子: 0.7
• 作者: Y. Yoshida