In this paper, by considering the stochastic proces s of the busy period and the idle period, and introducing the unfinished work as a supplementary variable, a new vector Markov process was presented to study th e M...In this paper, by considering the stochastic proces s of the busy period and the idle period, and introducing the unfinished work as a supplementary variable, a new vector Markov process was presented to study th e M/G/1 queue again. Through establishing and solving the density evolution equa tions, the busy-period distribution, and the stationary distributions of waitin g time and queue length were obtained. In addition, the stability condition of th is queue system was given by means of an imbedded renewal process.展开更多
Customer is the source of business income,a stable customer base is the guarantee of enterprise survival and development of enterprises by using Markov decision process,decision-makers in the new decision point in tim...Customer is the source of business income,a stable customer base is the guarantee of enterprise survival and development of enterprises by using Markov decision process,decision-makers in the new decision point in time,to the latest state of observation system and adopt an original decision,decided in a well-posed option set an action sequence,and then choose to create value and total revenue in this sequence the most significant behavior,obtain the best marketing strategy,formulate relevant enterprise actual customer,dynamic programming model.展开更多
匹配场统计反演海底声参数的根本目的是求解未知参数的后验概率分布(PPD)。针对现有各种求解参数PPD的数值方法如穷举搜索、Markov Chain Monte Carlo采样、最近邻域插值近似算法普遍存在计算速度慢、时间长、难以满足实际应用的问题,...匹配场统计反演海底声参数的根本目的是求解未知参数的后验概率分布(PPD)。针对现有各种求解参数PPD的数值方法如穷举搜索、Markov Chain Monte Carlo采样、最近邻域插值近似算法普遍存在计算速度慢、时间长、难以满足实际应用的问题,本文提出了一种基于支持向量机的快速求解参数PPD的新算法。该算法利用了支持向量机强大的小样本学习能力,通过训练学习拟合未知海底声参数和后验概率之间存在的函数关系,从而在求解参数PPD时简化了利用声场传播模型计算后验概率的复杂过程,减少了计算时间。数值仿真算例和海上实验数据的处理结果验证了该算法在低维匹配场统计反演海底声参数问题中的有效性。展开更多
基金Project supported by the National Natural Science Foundation of China(Grant No.70171059)
文摘In this paper, by considering the stochastic proces s of the busy period and the idle period, and introducing the unfinished work as a supplementary variable, a new vector Markov process was presented to study th e M/G/1 queue again. Through establishing and solving the density evolution equa tions, the busy-period distribution, and the stationary distributions of waitin g time and queue length were obtained. In addition, the stability condition of th is queue system was given by means of an imbedded renewal process.
基金Special Research Project of Shaanxi Provincial Department of Education(19JK0631)School-level Project Funded by Xi’an Peihua University(PHKT19026).
文摘Customer is the source of business income,a stable customer base is the guarantee of enterprise survival and development of enterprises by using Markov decision process,decision-makers in the new decision point in time,to the latest state of observation system and adopt an original decision,decided in a well-posed option set an action sequence,and then choose to create value and total revenue in this sequence the most significant behavior,obtain the best marketing strategy,formulate relevant enterprise actual customer,dynamic programming model.
文摘匹配场统计反演海底声参数的根本目的是求解未知参数的后验概率分布(PPD)。针对现有各种求解参数PPD的数值方法如穷举搜索、Markov Chain Monte Carlo采样、最近邻域插值近似算法普遍存在计算速度慢、时间长、难以满足实际应用的问题,本文提出了一种基于支持向量机的快速求解参数PPD的新算法。该算法利用了支持向量机强大的小样本学习能力,通过训练学习拟合未知海底声参数和后验概率之间存在的函数关系,从而在求解参数PPD时简化了利用声场传播模型计算后验概率的复杂过程,减少了计算时间。数值仿真算例和海上实验数据的处理结果验证了该算法在低维匹配场统计反演海底声参数问题中的有效性。
