الفهرس 
introductionOnly 14 pages are availabe for public view
 |  | from | 16 |  |  |
| | | from | 16 | | |
Abstract
Multidimensional indexing is one of the relatively new and important fields in Databases. It is concerned with the indexing of multi-attributed records. where queries, can be applied using some or all of the attributes. Multidimensional index structures are designed to index Illulti-attributcd records without favoring any of the attributes over the others. unless it is explicitly stated to do so.
Indexing multi-attributed records is referred to by the term of 11I1/1/ idilllensionlll indexing because each record is considered to be a point in a multidimensional space with a number of dimensions that is equal to the number of attributes. The values of the point coordinates along cach dimension are equivalent to the values of the corresponding attributes,
Multi-attributed records are formed by the extraction of object features. This formation of the records relies either on an expert system to extract these features or on manual extraction. Once the features are extracted. it is the responsibility of the multidimensional index structure to store these features and speed up their indexing.
In this thesis. a new index structure for multidimensional spaces has been developed. This index structure is an efficient structure for indexing multidimensional points. One of its maior advantages is that it is .parallel by nature. Moreover. it does not lose its efficiency if it is serially processed or processed using a few number of processors. It can take advantage of as many processors as the dimensionality of the space.
The proposed structure is compared to the Ifj’hrid free. which is known for its superiority among different index structures. Experimental results show that the parallel processing of the proposed structure dramatically reduces the number of disk accesses involved in the search operation. Even in its serial case. the proposed structure outper/(Jrms the Hybrid tree for large database sizes and for large search ranges. The proposed structure is easily implemented and tuned to get the best performance.