Search

A new different method has been developed for solving one-dimensional parabolic differential equations in this paper. Here, the standard backward Euler difference scheme was improved by the clustering algorithm. The method is illustrated by numerical examples. It not only reduces computing time remarkably, but also shows better performance in data storage and at the same time is more stable than the standard backward Euler difference scheme.

While applying attribute based access control in dynamic open multi-domain environments, an important critical problem is how to make these domains know each other about the semantics of attributes defined by them. Towards this problem, a novel method of attribute matching is proposed in this paper. This matching is based on other domains’ knowledges of the specific concept relationship. The knowledge involves five different concept relationships. Compared with the numerical similarity between two concepts, these concept relationships are more applicable in access control. In addition, the proposed method does not need the overall information of ontologies to which the matched concepts belong. For this, examples are given to illustrate applications of the method.

Concept alignment, namely how to make different domains know each other about the semantics of concepts defined by them, is one of the key issues of applying attribute based access control in dynamic open multi-domain environment. In this paper, a novel approach is proposed for building up a concept alignment framework based on other domains’ knowledge about specific concept relationship. The knowledge comprises five different definite concept relationships which are represented by five logical formulae respectively. Compared with the numerical similarity and those related approaches, the knowledge is more suitable for depicting concept relationship, while the proposed approach is more adapted in access control. Finally, a trust degree model is introduced into the framework in order to improve the quality of concept alignment results, and how to use and manage the trust degree is discussed.

Traditional trust models are based on reputation which is only a numerical value. Therefore, it is not fit for fine-grained access control which is needed in many online applications. Towards this problem, a novel trust model based on temporal historical data for access control is proposed. It could provide flexible method for users to customize specific access control rules. A policy language is presented using temporal logic and its syntax and semantics are defined formally. At the same time, a verification algorithm which is used to decide whether a policy is satisfied is introduced. We use e-commerce applications to demonstrate advantages of fine-grained access control and enhanced flexibilities of the model while comparing with the traditional models.

In the traditional history-based access control model, the policy is designed regarding only one subject. This policy is vulnerable, because if multi-subjects who have common interests collaborate, the access control may be bypassed. In addition, designing policy for multi-subjects' access is indispensable for many online applications. Towards this problem, a temporal access control based on multi-subjects is proposed in this paper. The main idea is to make decision of access according to active multi-subjects' history executions. A policy language is introduced and its syntax and semantics are defined formally. Finally, a policy verification algorithm is given and an example is used to illustrate it.