Abstract:B+ -trees are considered standard organization for indexes in database systems. The concurrent control mechanism on B+ -trees has a great effect on the performance of DBMS supporting multi-users. A variant of B+ -tree called e-B+ -tree(elastic B+ -tree) is presented in this paper. Then, the safe nodes and operations on e--B+ trees are defined and the moment of reconstruction of e-B+tree is regulated. All these measures not only reduce the overhead for maintaining the e--B+ tree and the time spent on locking, but also decrease the frequency of splitting and catenating operation. Therefore, the degree of concurrency of operations on e-B+tree and efficiency of the database system can increase by a big margin.

Abstract:The expansion usage of computers in publishing, advertising, packing fieldshas given rise to great requirement for more Chinese fonts. In this paper, the authorspresent a new method for combining the styles of various Chinese fonts and generating newfonts. The description method for Chinese character may be used for simulation of handwritten Chinese characters.

Abstract:In the real world, Al systems arc constantly and adversely influenced bynoisy data. This is also true of EBL (explanation-based learning). This paper discusseshow to cope with noisy data in explanation - based learning and proposes a NR -EBL(noise resistant explanation - based learning) algorithm. Unlike existing algorithms,NR-EBI. can learn macro rules and find the problem distribution when there is noise intraining examples. Also unlike similar work, NR-EBI. reveals the dependency of classifying examples correctly upon the regularities of noise and attempts to detect noise regularities from a set of training examples. With the help of knowledge of noise regularities,NR -EBI. can have a higher rate of correct recognition than traditional algorithms and previous work. NR EBI. is the combination of explanation - based learning and statisticalpattern recognition. Traditional algorithms are only special cases of NR-EBL when thereis no noise in training examples.

Abstract:The LP (logic of paradox) is a paraconsistent logic. One of the motivationsbehind paraconsistent logic, namely LP, is that it should not allow that everything followsfrom a single contradiction. However it has one important drawback: that some classicalinferences would be invalid in LP. The LP.(logic of minimal paradox) can overcome thisdrawback, such that paraconsistent logic would be equivalent to classical logic when therewas not direct effect of a contradiction. Originally, LP and LP. were only defined in semantic. Although some proof theories for LP were introduced, it has left open how to obtain a satisfactory proof theory for LP.. This paper propose the sound and completestableaux with respect to the semantics of LP and LP., respectively.

Abstract:Based on algebra and recursive function theory, a key concept of algebraicrecursive predicates is defined in this paper. Based on mathematical induction, a backwardreduction method and its corresponding algorithm reduction are given for proving the universal truth of algebraic recursive predicates. Because the method is reduction based, theefficiency of theorem-proving is improved greatly.

Abstract:In this paper, the correspondence among initial(terminal) semantics of gluedtheories and factor theories in institutions is clarified under certain intuitive conditions.

Abstract:To support the effective reasoning in the circumstances of incomplete andinsufficient facts, the author presented an inductive machine learning method, and designed a rule vector projection algorithm. With the method implemented by the algorithmthe primitive knowledge are processed by inductive approach, created a network knowledge base full of new classification concepts and paths for reasoning. Based on the knowledge base the quantity of facts needed in a diagnostic decision making by a machine inference system can sharply be reduced, thus a high performance can still be achieved by thesystem, despite of short information.

Abstract:This paper proposes a new object clustering method--hybrid method for effectively supporting both query of objects within a class and components of a complex objectsin object-oriented databases. The method has better performance in both average and theworst case than usual clustering methods.

Abstract:In order to avoid the disadvantage of BP algorithm, which is to trap into thelocal minimum, this paper provides Hybrid Algorithm consisted of BP with new combination activation function and statistics; then this new algorithm is applied to weather forecast and loan and the results are compared with Hybrid Algorithm based on backpropagation(standard and momentum-term) and statistics. Simulation results show that the accuracy and learning rate of prediction are both satisfactory. This program is written inTURBO-PASCAL 6. 0 and run on IBM PC/386 and 486.