Prince, Rawle C.S.
(2011)
Aspects of the theory of containers within automated theorem proving.
PhD thesis, University of Nottingham.
Abstract
This thesis explores applications of the theory of containers within automated theorem proving.
Container theory provides a foundational analysis of data types as containers, specified by a type $S$ of shapes and a function P assigning to each shape its set of positions for data.More importantly, a representation theorem guarantees that polymorphic functions between container data types are given by container morphisms, which are characterised by mappings between shapes and positions.
Container theory is interesting, in this context, for the following reasons. A mechanism for representing and reasoning with ellipsis (the dots in x_1, x_2, ... , x_n) in lists, existing in the literature, has proved to be very useful for formalisations involving abstractions. Success with this mechanism came by means of a metalevel representation through which many functions that normally require recursive definitions can be given explicit ones. As a result, not only can induction and generalisation be eliminated from proofs but, by means of an associated portrayal system, the resulting proofs are also intuitive and much closer to informal mathematical proofs.
This ellipsis mechanism, however, is not based on any formal theory, making it rather exiguous in comparison with rival techniques. There also remains questions about its scope and applications. Our aim is to improve this ellipsis mechanism. In this connection, we hypothesize that the theory of containers provides a formal underpinning for such representations. In order to test our hypothesis, we identify limitations of the ellipsis mechanism and show how they can be addressed within the theory of containers. We subsequently develop a new reasoning system based on containers, which does not suffer from these limitations.
This judicious containerbased system endorses representations of polymorphic rewrite rules using arithmetic, which naturally lends itself to applications of arithmetic decision procedures. We exploit this facet to develop a new technique for deciding properties of lists. Our technique is developed within a quasicontainer setting: shape maps are given as piecewiselinear functions, while a new representation is derived for reindexing functions that obviates the need for dependent types, which are fundamental in a judicious container approach. We show that this new setting enables us to represent and reason about a large class of properties.
Item Type: 
Thesis (University of Nottingham only)
(PhD)

Supervisors: 
Altenkirch, T. 
Subjects: 
Q Science > QA Mathematics > QA 75 Electronic computers. Computer science 
Faculties/Schools: 
UK Campuses > Faculty of Science > School of Computer Science 
Item ID: 
11793 
Depositing User: 
EP, Services

Date Deposited: 
03 Nov 2011 13:51 
Last Modified: 
13 Sep 2016 17:46 
URI: 
http://eprints.nottingham.ac.uk/id/eprint/11793 
Actions (Archive Staff Only)

Edit View 