Research And Development Of Relation-based Decision Data Browse Mode

Data support is an important part of Modern C2systems, as real-time situation data and related data play a key role in the decision making process. Led by the fast development of data collection and transmission techniques, available data increases at a high speed. In traditional tree-unfolding data browse mode, while seeking data, commanders spend much time on browsing other data that has little relation with his cared problem. On checking one data, commander usually associates about some other data that have kinds of relations with it, and may wonder their values to help him solve problem. However, traditional mode does not support this associative think pattern. In this thesis, a relation-based data browse mode is proposed as a complement to the traditional one, which provides data relations as a hint that helps user seeking data.Based on D2R technique, methods are proposed on atomatic relation extraction from relational database, and construction and publication of linked data. Applying JSP(Java Server Pages) as Web page development tool, and MyEclipse as development environment, a user interface that support relation-based data browse mode is designed and develped. On this basis, to overcome the latter's defficiency on navigation convergence, a data query system that integrate tree-unfolding and relation-based data browse functions is developed. The two query methods are combined together and works in synchronization, in order to realize complementary advantages, and improve system usability and efficiency. Furthermore, a detailed process on data browse simulating typical military scenario is designed based on military database, in order to show the use mode and effect of relation based data browse mode.To verify the advantage of proposed mode, it is compared with traditional mode on various performance such as navigation mode, problem relativity, support to associative think pattern, user skill requirement, operation efficiency, and navigation convergence, with test data support from a data query experiment using the developed simulation system. It is proved through comparison that relation-based data browse mode is a good complement to traditional tree-unfolding mode, rather than substitute. This mode has better problem relativity, provides better support to associative think pattern, and requires lower user skill requirement, while does not guarantee navigation convergence. Therefore, it is suggested to be combined with traditional mode in practical application, in order to achieve better performance.