摘要
提出一种在非平稳需求以及区间预测需求情况下的基于置信规则推理的库存控制方法.该方法不依赖于需求的分布模型,区间预测需求利用能够处理多种定性和定量不确定性信息的ER(证据推理)框架进行表达,领域专家知识可以用来构建和初始化置信规则库,历史需求信息可以用来训练置信规则库,以得到更加可信的推理.给出了一个汽车4S店库存-销售实例,证实了该方法的可行性及其相对于传统方法的优越性.
A belief-rule-based (BRB) inference method, which was independent of customer deman distribution, was proposed for inventory control under nonstatonary demand and interval foreeastin demand. Interval forecasting demand was represented in the ER (evidential reasoning) framewor which can deal with various kinds of qualitative and quantitative uncertain information. Domain exper knowledge was used to construct and initialize a belief rule base. Historical demand information can b used to train the belief rule base to get more reliable inference. Compared with traditional methods the feasibility of the BRB inventory control method and its features were illustrated through an aut, 4S inventory-sales example.
出处
《华中科技大学学报(自然科学版)》
EI
CAS
CSCD
北大核心
2011年第7期76-79,共4页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
国家自然科学基金资助项目(60674085
70572033
70971046
60736026)
国家科技部国际科技交流项目(20072607)
英国工程与物理科学研究委员会资助项目(EP/F024606/1)
关键词
专家系统
库存控制
置信规则库
证据推理
不确定性
expert system
inventory control
belief-rule-based
evidential reasoning
uncertainty