摘要
在经济系统中,决策越来越呈现出在线性特征,传统优化方法在解决这类在线问题时,通常假设未来输入是一随机变量从而寻求概率意义上的最优决策。近年来,在优化领域兴起了一种新的研究方法———在线算法与竞争分析,为解决这类在线问题提供了新的视角,但传统的竞争分析方法有意规避概率分布假设。对于在线租赁决策问题,由于其输入结构简单且具有良好的统计性质,似乎忽略这些有用的信息而只运用标准的竞争比方法分析显然具有不足之处。在本文中,我们将其输入结构的概率分布引入纯竞争分析方法中,从而建立了具有概率情形的最优在线租赁模型,并得到了最优竞争策略及其竞争比。
In economic system, the decision making is representing the onhne characteristics. Dealing with online decision problem,the traditional Bayesian analysis often depends on the probabilitic assumptions so that it gives out the optimal result in the probability sense. However, there is coming a focus in the algorithmic fields in the recent years,which is the online algorithms and competitive analysis. It gives a new angle of view to deal with online problem,but this method always intentionally avoids probabilistic distribution. For the online leasing problem,beacuse its input structure has a simple and good statistical property, it would be a waste to ignore this knowledge which is precisely what the traditional competitive ratio does. In this paper, we introduce the information distribution of future input into the competitive analysis so as to build online leasing model with probability distribution, and obtain their optimal competitive strategies and competitive ratios.
出处
《中国管理科学》
CSSCI
2005年第5期33-38,共6页
Chinese Journal of Management Science
基金
国家自然科学基金资助项目(10371094)
国家自然科学基金资助项目(70471035)
关键词
在线算法
金融租赁
竞争分析
竞争比
online algorithm
financial leasing
competitive analysis
competitive ratio
