【作者】 李衍杰； 殷保群； 奚宏生； 周亚平；

【Author】 Li Yanjie Yin Baoqun Xi Hongsheng Zhou Yaping (Department of Automation, University of Science and Technology of China, Hefei, 230027) ( Department of Management Science, University of Science and Technology of China, Hefei, 250027)

【机构】 中国科技大学自动化系； 中国科技大学管理科学系；

【摘要】 本文基于离散和连续Markov过程的性能势理论,研究了一类半Markov过程(SMP)的灵敏度分析和平均代价准则下性能优化问题。将半Markov过程转化为与之性能相同的离散和连续Markov 过程,利用两者的性能势理论, 导出了用势能表示的SMP灵敏度公式,同时在无限水平平均准则下, 给出了系统最优平稳策略所满足的最优性方程。更多还原

【Abstract】 Based on performance potential theory, the sensitivity analysis and performance optimization under the average cost criterion of SMP will be studied in this paper. After the SMP is transformed into the embedded Markov chain and Continuous Time Markov Process (CTMP), the potentials of the both are used to obtain the sensitivity analysis and performance optimization of SMP. The formulas of the sensitivity and optimality equation can be expressed by the potentials of the both.更多还原