Fuzzy netal network (FNN) is a new tool for extraction of fuzzy control rules from experimental data, butno such rule can be extracted directly from small samples. This paper presents a new approach to fuzzy rules and...Fuzzy netal network (FNN) is a new tool for extraction of fuzzy control rules from experimental data, butno such rule can be extracted directly from small samples. This paper presents a new approach to fuzzy rules andmembership function for small samples i. e. clustering by the Hebb differential competition rule and extending eachitem of sample information to the control point in its factor space while BP algorithm is applied to the study of factornetwork weights in it. This approach has ben successfully applied to the simulation of rainfall prediction.展开更多
基于扩展规则的方法是一种定理证明方法.在IER(improved extension rule)扩展规则算法的基础上,提出了IMOM(improved maximum occurrences on clauses of maximumsize)和IBOHM(improved BOHM)启发式策略,并将两种启发式策略用于IER算法...基于扩展规则的方法是一种定理证明方法.在IER(improved extension rule)扩展规则算法的基础上,提出了IMOM(improved maximum occurrences on clauses of maximumsize)和IBOHM(improved BOHM)启发式策略,并将两种启发式策略用于IER算法中,有指导性地选择限定搜索空间的子句,设计并实现了算法IMOMH_IER和IBOHMH_IER.实验结果表明,由于这两种启发式策略能够选择较为合适的搜索空间,可以尽快地判定出原问题是否可满足,故其速度平均能够达到原有算法DR(directional resolution)和IER的10~200倍.展开更多
给出一种采用多层次优化技术的XACML(extensible access control markup language)策略评估引擎实现方案MLOBEE(multi-level optimization based evaluation engine).策略判定评估前,对原始策略库实施规则精化,缩减策略规模并调整规则顺...给出一种采用多层次优化技术的XACML(extensible access control markup language)策略评估引擎实现方案MLOBEE(multi-level optimization based evaluation engine).策略判定评估前,对原始策略库实施规则精化,缩减策略规模并调整规则顺序;判定评估过程中,在引擎内部采用多种缓存机制,分别建立判定结果缓存、属性缓存和策略缓存,有效降低判定引擎和其他功能部件的通信损耗.通过两阶段索引实现的策略缓存,可显著降低匹配运算量并提高策略匹配准确率.仿真实验验证了MLOBEE所采用的多层次优化技术的有效性,其整体评估性能明显优于大多数同类系统.展开更多
文摘Fuzzy netal network (FNN) is a new tool for extraction of fuzzy control rules from experimental data, butno such rule can be extracted directly from small samples. This paper presents a new approach to fuzzy rules andmembership function for small samples i. e. clustering by the Hebb differential competition rule and extending eachitem of sample information to the control point in its factor space while BP algorithm is applied to the study of factornetwork weights in it. This approach has ben successfully applied to the simulation of rainfall prediction.
文摘基于扩展规则的方法是一种定理证明方法.在IER(improved extension rule)扩展规则算法的基础上,提出了IMOM(improved maximum occurrences on clauses of maximumsize)和IBOHM(improved BOHM)启发式策略,并将两种启发式策略用于IER算法中,有指导性地选择限定搜索空间的子句,设计并实现了算法IMOMH_IER和IBOHMH_IER.实验结果表明,由于这两种启发式策略能够选择较为合适的搜索空间,可以尽快地判定出原问题是否可满足,故其速度平均能够达到原有算法DR(directional resolution)和IER的10~200倍.
文摘给出一种采用多层次优化技术的XACML(extensible access control markup language)策略评估引擎实现方案MLOBEE(multi-level optimization based evaluation engine).策略判定评估前,对原始策略库实施规则精化,缩减策略规模并调整规则顺序;判定评估过程中,在引擎内部采用多种缓存机制,分别建立判定结果缓存、属性缓存和策略缓存,有效降低判定引擎和其他功能部件的通信损耗.通过两阶段索引实现的策略缓存,可显著降低匹配运算量并提高策略匹配准确率.仿真实验验证了MLOBEE所采用的多层次优化技术的有效性,其整体评估性能明显优于大多数同类系统.