完备内积空间中有界闭凸集的非扩张自映射有不动点。概率度量空间的非扩张自映射也有类似的不动点定理。完备变内积空间中,变内积引导出一个普通内积,作为内积空间和概率度量空间,上述结论显然是成立的。但变内积空间关于概率范数的非...完备内积空间中有界闭凸集的非扩张自映射有不动点。概率度量空间的非扩张自映射也有类似的不动点定理。完备变内积空间中,变内积引导出一个普通内积,作为内积空间和概率度量空间,上述结论显然是成立的。但变内积空间关于概率范数的非扩张自映射是否也有不动点呢?在本文中,我们证明了,若 B 是完备变内积空间 H 的紧凸集,T 是从 B 到 B的非扩张映射,I-T 是单调映射,则 T 有不动点,而且 T 的不动点组成的集合是凸集。展开更多
In this paper we introduce a routing algorithm for commuication networks with multiple QoS metrics. This algorithm can be used for QoS-based route computaion for ATM PNNI and Internet protocols such as QoS OSPF. Acomm...In this paper we introduce a routing algorithm for commuication networks with multiple QoS metrics. This algorithm can be used for QoS-based route computaion for ATM PNNI and Internet protocols such as QoS OSPF. Acommunication network containing links with multiple QoS metrics is modeled as a graph with multiple weights associated with its edges. This method takes a metric(e. g. cost)as the optmization target and another metric(e. g. delay)as a constraint. This algorithm is guaranteed to find a nearly optimal path satisfying the given comstraint if such a path exists. The algorithm is very efficient and its computational complexity is O(n2).展开更多
文摘完备内积空间中有界闭凸集的非扩张自映射有不动点。概率度量空间的非扩张自映射也有类似的不动点定理。完备变内积空间中,变内积引导出一个普通内积,作为内积空间和概率度量空间,上述结论显然是成立的。但变内积空间关于概率范数的非扩张自映射是否也有不动点呢?在本文中,我们证明了,若 B 是完备变内积空间 H 的紧凸集,T 是从 B 到 B的非扩张映射,I-T 是单调映射,则 T 有不动点,而且 T 的不动点组成的集合是凸集。
文摘In this paper we introduce a routing algorithm for commuication networks with multiple QoS metrics. This algorithm can be used for QoS-based route computaion for ATM PNNI and Internet protocols such as QoS OSPF. Acommunication network containing links with multiple QoS metrics is modeled as a graph with multiple weights associated with its edges. This method takes a metric(e. g. cost)as the optmization target and another metric(e. g. delay)as a constraint. This algorithm is guaranteed to find a nearly optimal path satisfying the given comstraint if such a path exists. The algorithm is very efficient and its computational complexity is O(n2).
