Research On Totally-ordered And Partially-ordered Temporal Database

With the development of database and information technology, information system faces many new application and requirements. There are more urgent requirements for temporal information management. Temporal information management has become crusial technology in many new database and information system. Research of temporal database has become an important and vigourous study field in database and information system.As with traditional relational database, in the study of temporal database, one of the important problems is logic design of temporal database. Temporal database with multiple granularities has some advanced properties because it can support multiple time granularities. It's an important task for temporal database researchers to design effective normalization method for temporal database with multiple granularities. The former design methods for temporal database with multiple granularities are too complicated to be used for effective temporal database design, because the operations of the temporal types involved in the algorithm are very complicated. This paper dedicates to investigate more effective and more practical design method for temporal database with multiple granularities.Temporal type sets involved in many real world applications can constitute one totally ordered relation, and totally ordered temporal schemes with totally ordered temporal type sets have fine properties. Study of totally ordered temporal scheme design is feasible. This paper first investigates temporal data dependencies and relevant temporal normal form decompositions for totally-ordered temporal database, and on this basis further investigates partially-ordered temporal database normalization.In the study of membership problem of totally-ordered temporal database, the basic concepts in totally-ordered temporal scheme, the inference rules for totally-ordered temporal functional dependencies (TFDs), the attribute's clousure algorithm for totally-ordered temporal scheme and the membership algorithm for totally-ordered temporal database are proposed. This paper also solves the TFDs cover problem in totally-ordered temporal database.In the study of totally-ordered temporal Boyce-Codd normal form's decomposition of totally-ordered temporal scheme, the paper proposes the concepts of totally-ordered temporal Boyce-Codd normal form, etc., the algorithm of temporal candidate keys set and the algorithm of decomposition of totally-ordered temporal Boyce-Codd normal form which can be totally-ordered losslessly joint.In the study of totally-ordered temporal simple normal form's decomposition of totally-ordered temporal scheme, the paper proposes the concepts of totally-ordered temporal simple normal form, etc., the algorithm of totally-ordered temporal simple candidate keys set and the algorithm of decomposition of totally-ordered temporal simple normal form which can be totally-ordered losslessly joint and preserve the temporal functional dependencies. Its normalization level is higher than totally-ordered temporal third normal form.In the study of multi-valued dependencies in strong-totally-ordered temporal scheme, the paper proposes the concepts of regular temporal multi-valued dependency, etc., the inference rules for totally-ordered regular temporal multivalued dependencies, the relevant theroms about regular temporal multi-valued dependencies and the lossless decomposition of strong-totally-ordered temporal scheme. The paper also gives the membership algorithm of mixed dependencies set in strong-totally-ordered temporal scheme and the cover algorithm of totally-ordered temporal mixed dependencies set.In the study of TFDs in non-strict-partially-ordered temporal schemes, the paper introduces the concepts of non-strict-partially-ordered temporal types set, non-strict-partially-ordered temporal schemes, etc. The paper also proposes the inference rules for non-strict-partially-ordered TFDs and the closure algorithm for attributes in non-strict-partially-ordered temporal scheme.In the study of Boyce-Codd normal form's decomposition of non-strict-partially-ordered temporal scheme, the paper proposes the relavant theorems for non-strict-partially-ordered temporal Boyce-Codd normal form's (NSPO_TBCNF) decomposition, gives the NSPO_TBCNF's deciding algorithm and its decomposition algorithm.The time complexities of the membership algorithm of totally-ordered temporal scheme, the TFD's cover algorithm, the totally-ordered lossless decomposition algorithm of totally-ordered temporal Boyce-Codd normal form, the totally-ordered lossless decomposition algorithm of totally-ordered temporal simple normal form and the membership algorithm of mixed dependencies set in strong-totally-ordered temporal scheme are all polynomial. They are more practical and effective. They can effectively eliminate the temporal redundancies in totally-oredered temporal schemes and maintain the consistency of temporal data in totally-oredered temporal schemes. Without new temporal types created, the non-strict-partially-ordered lossless NSPO_TBCNF decomposition algorithm can maximumly eliminate the temporal redundancies in non-strict-partially-ordered temporal schemes with constraint of TFDs. Its complexity is also polynomial. The paper sets a solid foundation for totally-ordered and partially-ordered temporal databases' normalization.