The 2002 U.K. Workshop on Computational Intelligence

(UKCI'02)

2-4 September 2002, Birmingham, U.K.


Introduction

Call For Papers

Submission Details

Poster Guidelines

Important Dates

Workshop Programme

Registration Details

Committees

Location / Directions

Birmingham Attractions

Contact Details

Pictures

UKCI-01


School of
Computer Science

PROGRAMME

Monday 2 September

10.00 — 11.00 Registration and Tea/Coffee

11.00 — 11.10 Welcome address

11.10 — 12.00 Keynote address 1

Using Evolutionary Algorithms To Create Algorithms

Peter Ross

Abstract: In a wide range of commercially important areas, evolutionary algorithms (EAs) have been strikingly successful -- when applied by academics. Among commercial users the take-up has been more variable; there are many unusual success stories but there are also good reasons why somebody might still not wish to jump into the unfamiliar waters of EAs. This talk is about a general line of development that might address some of the existing concerns. The general idea is, rather than using an EA to find a solution to a specific problem, use it to find an algorithm that can solve a family of problems well.

12.00 — 1.00 Session 1 — Evolutionary Computation

Maximum Cardinality Matching by Evolutionary Algorithms

Jun He & Xin Yao

Genetic Algorithms for Discrete Sequence Prediction

Anikó Ekárt

1.00 — 2.00 Lunch

2.00 — 3.30 Session 2 — Genetic Programming & Evolvable Hardware

What Can Automatic Programming Learn from Theoretical Computer Science?

Colin G. Johnson

Studying the Emergence of Multicellularity with Cartesian Genetic Programming in Artificial Life

Joseph A. Rothermich & Julian F. Miller

Measuring Fitness of Digital Circuits in Evolvable Hardware

Mark Lucas

3.30 —4.00 Tea/Coffee

4.00 — 4.30 Invited Talk

Data Mining, Natural Computation and Intelligence Amplification

Tom Khabaza

4.30 — 5.30 Session 3 — Evolutionary Applications

Evolutionary Approach for Vehicle Routing Problem with Time Windows and Facility Location Allocation Problem

Kamal Gupta & Xin Yao

A Hybrid Approach to Feature Selection and Generation Using an Evolutionary Algorithm

Oliver Ritthoff, Ralf Klinkenberg, Simon Fischer & Ingo Mierwa

5.30 — 6.00 Poster Session Adverts

Poster presenters get four minutes each to introduce themselves and advertise their poster.

6.00 — 11.15 Pre-dinner drinks, Dinner, Post-dinner drinks


Tuesday 3 September

9.00 — 11.00 Session 4 — Learning and Optimization

An Instance-based Learning Approach based on Grey Relational Structure

Chi-Chun Huang & Hahn-Ming Lee

Towards Comprehensive Clustering of Mixed Scale Data with K-Means

Boris Mirkin

The Stability of General Discounted Reinforcement Learning with Linear Function Approximation

Stuart I. Reynolds

No Free Lunch, Program Induction and Combinatorial Problems

John Woodward & James Neil

11.00 — 11.30 Tea/Coffee

11.30 — 1.00 Session 5 — Neural Networks

Maximum Likelihood Topology Preserving Algorithms

Emilio Corchado & Colin Fyfe

Artificial Speciation of Neural Network Ensembles

Vineet Khare & Xin Yao

To Modularize or Not To Modularize?

John A. Bullinaria

1.00 — 2.00 Lunch

2.00 — 3.00 Keynote address 2

Intelligent Systems for On-line Monitoring and Control of Patient Unconsciousness in Operating Theatres

Derek Linkens

Abstract: The monitoring and control of unconsciousness under surgical conditions in operating theatre is a major challenge to both anaesthetists and machines. An intelligent system combining a number of advanced computational paradigms has been developed for the assessment of depth of anaesthesia which utilises auditory evoked brain potentials, heart rate and blood pressure measurements. Using wavelets analysis the features within the auditory evoked signals are extracted and then fed to a learning neuro- fuzzy system which in turn classifies the depth of anaesthesia. In addition, the heart rate and blood pressure signals are used as a second measure based on a rule-based fuzzy logic system. The two measures are then fused to give a final indication of anaesthetic depth. This is then fed back to a Target Controlled Infusion (TCI) system for regulating the infusion of the drug propofol for the maintenance of anaesthetic state. The fuzzy logic rule-base is constructed from pharmacological knowledge elicited from experts. In addition, fuzzy-based modelling has been utilised in producing a patient model which includes the effects of analgesic agents on patients drug sensitivity and surgical stimulation of varying intensities. The architecture has been validated via extensive simulation and in clinical trials. The system has also been developed as a computer- based training device.

3.00 — 4.00 Session 6 — Data Analysis

Label Prototypes for Data Analysis

Jonathan Lawry

Intelligent Sensor Fusion in Uncertain Taxonomical Hierarchies

Jonathan M. Rossiter, Toshiharu Mukai & Paul R. Goddard

4.00 — 4.30 Tea/Coffee

4.30 — 6.00 Session 7 — Fuzzy Systems

Modelling Capabilities of Fuzzy Relational Models

Yue Wu & Arthur Dexter

Aiding Fuzzy Rule Induction with Fuzzy Rough Attribute Reduction

Richard Jensen & Qiang Shen

Fuzzy Modelling for Student Academic Performance Evaluation

Khairul A. Rasmani & Qiang Shen

6.00 — 11.15 Pre-dinner drinks, Conference Banquet, Post-dinner drinks


Wednesday 4 September

9.00 — 10.00 Keynote address 3

Applications of "Intelligent" Methods in Logistics

Hans-Jurgen Zimmermann

Abstract: Transportation Logistics has been considered as an important area of applications of Operations Research methods for decades and its importance will not decrease but rather increase in the future. The general structure is generally conceived as that of combinatorial optimisation and models such as the Travelling Salesman, Routing, Scheduling etc. are standard models that have been around for the last half century. What has been neglected very often in the OR Literature are non-standard topologies, the consideration of uncertainty and fast changing dynamic environments. Further, no distinction has been made between planning models and control applications. Since the beginning of the 90's methods of Computational Intelligence have increasingly been used to complement classical OR approaches to tackle the above mentioned aspects in real applications. The presentation will give a survey of such applications in traffic management, fleet management and in-house logistics and describe in more detail approaches used for optimising the control of container terminals.

10.00 — 11.00 Session 8 — Agents and Swarms

A Multi-Objective Algorithm based upon Particle Swarm Optimisation, an Efficient Data Structure and Turbulence

Jonathan E. Fieldsend & Sameer Singh

A Framework for Comparing Agent Architectures

Aaron Sloman & Matthias Scheutz

11.00 — 11.30 Tea/Coffee

11.30 — 1.00 Invited talks from Industrial Representatives

Paul Marrow

BT Laboratories, Ipswich

Dave Cliff

HP Labs, Bristol

1.00 — 2.30 Poster Session with Buffet Lunch

Automatic VHDL-AMS Code Generation from UML Diagrams for Analogue System Modelling

C.T. Carr, T.M. McGinnity & L.J. McDaid

On the Implementation of Rough Set Attribute Reduction

Alexios Chouchoulas, Joe Halliwell & Qiang Shen

Intelligent Chain Code Quantisation for Multiple Classifier Based Shape Recognition

S. Hoque, K. Sirlantzis & M.C. Fairhurst

Intelligent Shopping Negotiation Agents that can Adapt User Preferences

R. Huang, T. Yamazaki & H. Ouchiyama

Game-Playing by Machine Learning

Suraj Kumar

Towards a Multi-Algorithm Vehicle Routing Problem Solver

Finlay Smith, Ann Tighe & Gerard Lyons

Estimation of Distribution Algorithm Based on Mixtures: Preliminary Experimental Results

Qingfu Zhang, Jianyong Sun, Edward Tsang & John Ford

2.30 - ? Afternoon free for meetings/discussions/sightseeing/early departures



Reserve talk to be fitted into a gap somewhere

Statistics-based Adaptive Non-Uniform Crossover for Genetic Algorithms

Shengxiang Yang


Last modified Thursday, 30-Jan-2003 12:55:10 GMT . Please report errors to John Bullinaria.