Data Structures for Sorting, Searching, Hashing, Computational Geometry and Nonprocedural Databases

点击翻译按钮获取中文摘要

Data structures to accommodate fast retrieval are perhaps the oldest and certainly one of the most fundamental topics in computer science. This project addresses several open problems of both pragmatic and theoretical interest. The first group of open questions concerns worst-case algorithms that outperform the information theoretical lower bound. Recently an O(log N/log log N) worst-case time algorithm for dynamic tree operations and an accompanying O(N log N/log log N) worst-case time sorting algorithm have been designed. Both algorithms run on essentially all digital computers and on all possible data inputs. The new approach to sorting and searching rests on utilizing the power of a digital computer's arithmetic and bitwise logical operations. Numerous open questions arise about what upper and lower bounds for sorting, searching, balance tree operations, and hashing can be demonstrated in this modified mathematical model. A second research area focuses on computational geometry. Many geometric problems, such as Voronoi diagram construction, have been thought to require at least N log N time because of their reduction to sorting procedures, but all such theoretical lower bounds need to be reconsidered in light of the recent results for sorting and searching. There are also several open questions raised by earlier research into range query theory. One particular question concerns developing lower bounds that take into account the presence of subtraction, and a second concerns the time-space tradeoff for especially reporting queries. Range query algorithms for batch queries, especially procedures for executing relational calculus and SETL-like expressions, will also be studied.