Framework for expressing prioritized constraints using infinitesimal logic

• Main Author: Agarwal, Ruchi
• Other Authors: Wadge, W. W.
• Language: English; en
• Published: 2009
• Subjects: querying; databases; UVic Subject Index::Sciences and Engineering::Applied Sciences::Computer science
• Online Access: http://hdl.handle.net/1828/1833

In this thesis, we propose an extension to the multiple-valued infinitesimal logic frame-work to provide a simple representation for prioritized constraints. We introduce two unary operators, µ and w, to infinitesimal logic in order to define preferential constraints and backup constraints, respectively. The new framework naturally allows us to define a hierarchy of priorities among constraints. Also, we present a lazy algorithm for evaluating the multiple-valued prioritized constraint expressions of our representation. Our algorithm, which is similar to the alpha-beta pruning technique for minimax game tree evaluation, is based on a recursive depth-first traversal of the parse tree for the expression and works by evaluating operands within an increasingly narrower range of interest. Our implementation of this representation for querying a movies database demonstrates the expressive power and flexibility of our framework.