Home : Research Staff : Dr Stephen Fitzpatrick : Publications

Previous Projects

Previous Projects at Kestrel Institute

Evolutionary Design of Complex Systems

September 1998-September 1999
Role: Principal investigator/project leader
Funding body: DARPA

The objectives of the EDCS project are: to develop a practical design record formalism for precisely capturing the high-level specification and design of software systems and refinement into low-level executable code; to investigate the use of design records in incrementally modifying executable code to reflect changes in software specifications and/or designs; and to develop automated tool support for code modification based on design records.

Real-time Asset Tracking and Continuous Rescheduling

October 1997-October 1998
Principal investigator: Dr Douglas Smith
Associated groups: SAVI Technology, Mountain View, California
Role: Computer scientist
Funding body: DARPA

The goal of this on-going work is to demonstrate the effectiveness of a generic architecture for real-time asset tracking and dynamic scheduling/rescheduling. The technology for real-time tracking is being developed and deployed by Savi Technology. Kestrel Institute is developing a software architecture for the rapid development of fast (re)scheduling software that will use the real-time tracking information to continually optimize schedules as they are executed. My technical contributions during my involvement with this project were:

  • the development of formal domain theories to model the deployment environment, including the assets being scheduled;
  • the development of software architectures to represent generic scheduling and rescheduling algorithms;
  • the development of formal program transformations to instantiate the generic algorithms for specific deployment domains and to optimize resulting code;
  • the development of an interface specification between Kestrel's software and the SAVI tracking software and deployment database;
  • testing and evaluating the scheduling software and evaluating the schedules produced.

Automated Synthesis of Planners and Schedulers

February 1997-October 1997
Principal investigators: Dr Douglas Smith, Prof. Subbarao Kambhampati
Role: Computer scientist
Funding body: DARPA

This project focused on applying synthesis technology to the routine production of high-performance, specialized planning and scheduling tools. My technical contributions to this project were:

  • the assessment of the execution performance of extant synthesized implementations and the identification of optimization opportunities;
  • the development of formal program transformations to exploit the identified optimizations;
  • the extension of extant domain theories to incorporate a more complex and realistic model for key assets.

Previous Projects Elsewhere

Web Dissemination of Local News Broadcasts

Genesis Project Limited, Northern Ireland
November 1996-January 1997
Role: System designer and implementer

A software system was developed that captured the audio component of a specified television channel (UTV, a commercial television channel covering Northern Ireland) at specified times (corresponding to the local news bulletins). The audio captures were contemporaneously available as live audio streams accessible via the World Wide Web and one capture per day was archived to permit subsequent access.

Transformational Refinement of Algorithms for Scientific Applications

Department of Computer Science, The Queen's University of Belfast, Northern Ireland
September 1993-July 1995
Project leaders: Prof. Maurice Clint & Dr Peter Kilpatrick
Role: Researcher
Funding body: SERC (UK Science and Engineering Research Council)

The aim of the research was to investigate the extent to which the construction of efficient implementations of numerical algorithms for the solution of problems on novel architectures could be facilitated by the use of automatic program transformation. In particular, the research involved:

  • studying implementations of a number of different algorithms that exploit the specialized facilities of particular hardware with a view to discovering the nature of the refinement process from machine independent functional specification to imperative implementation;
  • developing transformational derivations which refine functional specifications to efficient implementations;
  • studying these transformational derivations with the aim of identifying transformations those which have wide applicability.

Ph.D. Thesis: The Specification of Array-Based Algorithms and the Automated Derivation of Parallel Implementations through Program Transformation

Department of Computer Science, The Queen's University of Belfast, Northern Ireland
October 1990-September 1994
Supervisor: Dr Terence Harmer

It is generally a difficult task to construct efficient implementations of numerical mathematical algorithms for execution on high-performance computer systems. The difficulty arises from the need to express an implementation in a form that reflects the nature of the computer system, rather than a form that reflects the computations performed by the algorithm. This thesis develops the method of program transformation to derive automatically efficient implementations of algorithms from high-level, machine-independent specifications. The primary system considered is the AMT DAP array processor, but sequential and vector systems are also considered. The transformational method is further extended to automatically tailor implementations to use sparse programming techniques.

Visiting Scholarship

Mathematics and Computer Science Division, Argonne National Laboratory, U.S.A.
October 1991-May 1992
Supervisor: Dr James Boyle
Associated group: Functional Programming Research Group, Amoco Production Research Laboratory, Tulsa

The research was concerned with the specification of algorithms for the solution of systems of linear equations and their implementation through automated program derivation. Implementations were derived for a sequential processor and an array processor.

M.Sc. Project: An Environment for Transformational Programming

M.Sc. Project
Department of Computer Science, The Queen's University of Belfast, Northern Ireland
May 1990-September 1990
Supervisor: Dr Terence Harmer

The project resulted in the construction of a software tool that provides an environment for the TAMPR transformation system on the NeXT workstation. This environment manages the development of transformations, derivations (collections of transformations) and test programs, and the application of derivations to test programs.

 

Back to top

 

 - Home - About Kestrel - Kestrel Staff - Kestrel Research - Research Partners -
- Career Opportunities - Available Software - Contact Kestrel - Sitemap -