Skip to main content

Engineering

Photo of Prof Edward Keedwell

Prof Edward Keedwell

Professor of Artificial Intelligence

 E.C.Keedwell@exeter.ac.uk

 (Streatham) 4014

 01392 724014


Overview

Professor Keedwell is Professor of Artificial Intelligence. He joined the Computer Science discipline in 2006 and was appointed as a lecturer in 2009. He has research interests in optimisation (e.g. genetic algorithms, swarm intelligence, hyperheuristics) machine learning and AI-based simulation and their application to problems in bioinformatics and engineering yielding over 160 journal and conference publications. He leads a research group focusing on applied artificial intelligence and has been involved with successful funding applications totalling over £3.75 million from the EPSRC, Innovate UK, EU and industry.  Particular areas of current interest are the optimisation of transportation systems, the development of sequence-based hyperheuristics and human-in-the-loop optimisation methods for applications in engineering.  Current projects include:

  • Exploiting Quantum Computing for Large-Scale Transport Models (Innovate UK with CityScience)
  • Advanced Metaheuristics and Hyperheuristics for Telecommunications Scheduling Problems (with BT)
  • EPSRC iCASE: Ethics and Artificial Intelligence: Privacy in Private Spaces (with Dyson)

Previous projects include:

Recent activity

  • Chair: Evolution Artificielle 2022 - 31st October-2nd November 2022, Exeter
  • Speaker:  BCS Real Artificial Intelligence 2022, September 2022, London
  • Speaker:  Somerset Innovation Exchange, September 2022, Taunton
  • Keynote speaker & Panellist:  BCS AI-2021, December 2021, Online (Cambridge)
  • Keynote speaker: ROADEF 2021, April 2021, Online (Mulhouse)
  • Contributor:  Alan Turing Institute Presents AI-UK 2021, March 2021, Online (London)
  • BT Thought Leadership Seminar, July 2020, Online (Ipswich)
  • Keynote speaker: Evolution Artificielle 2019, October 2019, Mulhouse, France
  • Keynote speaker:  MCDM Workshop, GECCO 2019, Prague, Czech Republic

Membership of professional bodies

  • Member of the Association of Computing Machinery (ACM)
  • Fellow of the Higher Education Academy.
  • Member of the EPSRC Peer Review College
  • Member of the IEEE Task Force on Hyperheuristics
  • Fellow of the Alan Turing Institute (2021-2023)
  • Member of the AISB Committee (2010-2020)

Administrative responsibilities

  • Co-Lead, IDSAI Trustworthy AI Theme (2021-)
  • Academic Lead, AI Group (2020-)
  • Director of Research, Computer Science (2015-2021)
  • Chair of the EPSRC Strategy Group (2018-2019)
  • Admissions Tutor, Computer Science (2013-2017)

Qualifications

BSc Cognitive Science, University of Exeter (1998)

PhD Computer Science, University of Exeter (2003)

Back to top


Publications

Copyright Notice: Any articles made available for download are for personal use only. Any other use requires prior permission of the author and the copyright holder.

| 2023 | 2022 | 2021 | 2020 | 2019 | 2018 | 2017 | 2016 | 2015 | 2014 | 2013 | 2012 | 2011 | 2010 | 2009 | 2008 | 2007 | 2006 | 2005 | 2004 | 2003 | 2002 | 2000 | 1999 |

2023

2022

2021

2020

2019

  • Bunce E, Keedwell E. (2019) Optimisation of a Checkers Player Using Neural and Metaheuristic Approaches, Evolution Artificielle 2019, Mulhouse, France, 29th - 30th Oct 2019, https://www.springer.com/gb/book/9783030457143, DOI:10.1007/978-3-030-45715-0_5.
  • Johns MB, Mahmoud H, Keedwell E, Savic D. (2019) A Diameter Probability Distribution Genetic Algorithm for Least-cost Water Distribution Network Design, 17th International Computing & Control for the Water Industry Conference, Exeter, United Kingdom, 1st - 4th Sep 2019.
  • Mahmoud HA, Johns M, Keedwell E, Savic D, Mahmoud H. (2019) Generalising human heuristics in augmented evolutionary water distribution network design optimisation, 17th International Computing & Control for the Water Industry Conference, University Of Exeter, 1st - 4th Sep 2019.
  • Younis MC, Keedwell E. (2019) Semantic segmentation on small datasets of satellite images using convolutional neural networks, Journal of Applied Remote Sensing, volume 13, no. 04, pages 1-1, DOI:10.1117/1.jrs.13.046510. [PDF]
  • Johns M, Keedwell E, Savic D. (2019) Knowledge Based Multi Objective Genetic Algorithms for the Design of Water Distribution Networks, Journal of Hydroinformatics, DOI:10.2166/hydro.2019.106.
  • Ross N, Johns M, Keedwell EC, Savic D. (2019) Human-Evolutionary Problem Solving through Gamification of a Bin-Packing Problem, Genetic and Evolutionary Computing Conference 2019, Prague, Czech Republic, 13th - 17th Jul 2019, Genetic and Evolutionary Computing Conference Companion, DOI:10.1145/3319619.3326871.
  • Fraser D, Keedwell EC, Michell S, Sheridan R. (2019) EMOCS: Evolutionary Multi-objective Optimisation for Clinical Scorecard Generation, Genetic and Evolutionary Computation Conference, GECCO 2019, Prague, Czech Republic, 13th - 17th Jul 2019.
  • Johns M, Mahmoud H, Walker D, Ross N, Keedwell EC, Savic D. (2019) Augmented Evolutionary Intelligence: Combining Human and Evolutionary Design for Water Distribution Network Optimisation, Genetic and Evolutionary Computing Conference, GECCO 2019, Prague, Czech Republic, 13th - 17th Jul 2019.
  • Yates WB, Keedwell EC. (2019) An analysis of heuristic subsequences for offline hyper-heuristic learning, Journal of Heuristics, volume 25, no. 3, pages 399-430, DOI:10.1007/s10732-018-09404-7. [PDF]

2018

  • Rakhshani H, Idoumghar L, Lepagnot J, Brevilliers M, Keedwell E. (2018) Automatic hyperparameter selection in Autodock, DOI:10.48550/arxiv.1812.02618.
  • Meyers G. (2018) Data-Driven Approaches for Near Real-Time Forecasting of Discolouration Events in Pipe Networks.
  • Rakhshani H, Idoumghar L, Lepagnot J, Brevilliers M, Keedwell E. (2018) Automatic hyperparameter selection in Autodock, PROCEEDINGS 2018 IEEE INTERNATIONAL CONFERENCE ON BIOINFORMATICS AND BIOMEDICINE (BIBM), pages 734-738. [PDF]
  • Fayeez ATI, Keedwell E, Collett M. (2018) Investigating behavioural diversity via gaussian heterogeneous ant colony optimization for combinatorial optimization problems, ACM International Conference Proceeding Series, pages 46-50, DOI:10.1145/3292448.3292459.
  • Younis MC, Keedwell E, Savic D. (2018) An Investigation of Pixel-Based and Object-Based Image Classification in Remote Sensing, ICOASE 2018 - International Conference on Advanced Science and Engineering, pages 449-454, DOI:10.1109/ICOASE.2018.8548845.
  • Wilson DG, Rodrigues S, Segura C, Loshchilov I, Huer F, Buenl GL, Kheiri A, Keedwell E, Ocampo-Pineda M, Özcan E. (2018) Summary of evolutionary computation for wind farm layout optimization, GECCO 2018 Companion - Proceedings of the 2018 Genetic and Evolutionary Computation Conference Companion, pages 31-32, DOI:10.1145/3205651.3208208.
  • Rakhshani H, Idoumghar L, Lepagnot J, Brévilliers M, Keedwell E. (2018) A novel population initialization method based on support vector machine, IEEE International Conference on Systems, Man, and Cybernetics (SMC2018), Miyazaki, Japan, 7th - 10th Oct 2018.
  • Johns M, Walker DJ, Keedwell E, Savic D. (2018) Interactive Visualisation of Water Distribution Network Optimisation, International Conference on Hydroinformatics (HIC 2018), 1st - 6th Jul 2018.
  • Walker DJ, Johns M, Keedwell E, Savic D. (2018) Generating Heuristics to Mimic Experts in Water Distribution Network Optimisation, Computing and Control in the Water Industry (CCWI 2018), 23rd - 25th Jul 2018.
  • Walker D, Johns M, Keedwell E, Savic D. (2018) Towards Interactive Evolution: A Distributed Optimiser for Multi- objective Water Distribution Network Design, 13th International Hydroinformatics Conference (HIC 2018).
  • Wilson D, Rodrigues S, Segura C, Loshchilov I, Hutter F, Lopez G, Kheiri A, Keedwell EC, Ocampo-Pineda M, Ozcan E. (2018) Evolutionary computation for wind farm layout optimization, Renewable Energy.

2017

  • Rosin T, Romano M, Woodward K, Keedwell E, Kapelan Z. (2017) Prediction of CSO chamber water levels using rainfall forecasts, CCWI 2017 - 15th International Conference on Computing and Control for the Water Industry.
  • Meyers G, Kapelan Z, Keedwell E. (2017) Data-driven approach to short-term forecasting of turbidity in a trunk main network, CCWI 2017 - 15th International Conference on Computing and Control for the Water Industry.
  • Younis MC, Keedwell E, Savic D, Raine A. (2017) Evaluating classification algorithms for improved wastewater system calibration, CCWI 2017 - 15th International Conference on Computing and Control for the Water Industry.
  • Meyers G, Kapelan Z, Keedwell E. (2017) Data-Driven study of discolouration material mobilisation in trunk mains, Water (Switzerland), volume 9, no. 10, DOI:10.3390/w9100811.
  • Fayeez A, Keedwell E. (2017) H-ACO: A Heterogeneous Ant Colony Optimisation approach with Application to the Travelling Salesman Problem, Evolution Artificielle 2017, Paris, France, 25th - 27th Oct 2017.
  • Yates W, Keedwell E. (2017) Offline Learning for Selection Hyper-heuristics with Elman Networks, Evolution Artificielle 2017, Paris, 25th - 27th Oct 2017.
  • Meyers G, Kapelan Z, Keedwell E. (2017) Short-Term Forecasting of Turbidity in Trunk Main Networks, Water Research.
  • Yates W, Keedwell EC. (2017) Clustering of Hyper-heuristic Selections using the Smith-Waterman Algorithm for Offline Learning, Genetic and Evolutionary Computation Conference (GECCO) 2017, Berlin, 15th - 19th Jul 2017.
  • Kheiri A, Keedwell E. (2017) A Hidden Markov Model Approach to the Problem of Heuristic Selection in Hyper-Heuristics with a Case Study in High School Timetabling Problems, Evol Comput, volume 25, no. 3, pages 473-501, DOI:10.1162/EVCO_a_00186. [PDF]

2016

  • Meyers G, Kapelan Z, Keedwell E, Randall-Smith M. (2016) Short-term forecasting of turbidity in a UK water distribution system, 12TH INTERNATIONAL CONFERENCE ON HYDROINFORMATICS (HIC 2016) - SMART WATER FOR THE FUTURE, volume 154, pages 1140-1147, DOI:10.1016/j.proeng.2016.07.534. [PDF]
  • Walker DJ, Keedwell, Savic. (2016) Multi-objective Optimisation of a Water Distribution Network with a Sequence-based Selection Hyper-heuristic, Computing and Control for the Water Industry 2016, Amsterdam, 7th - 9th Nov 2017.
  • Bailey J, Harris E, Keedwell E, Djordjevic S, Kapelan Z. (2016) Developing Decision Tree Models to Create a Predictive Blockage Likelihood Model for Real-World Wastewater Networks, Hydroinformatics Conference, Incheon, South Korea, 21st - 26th Aug 2016, Procedia Engineering, volume 154, pages 1209-1216, DOI:10.1016/j.proeng.2016.07.433.
  • Bailey J, Harris E, Keedwell E, Djordjevic S, Kapelan Z. (2016) The Use of Telemetry Data for the Identification of Issues at Combined Sewer Overflows, Hydroinformatics Conference, Incheon, South Korea, 21st - 26th Aug 2016, Procedia Engineering, volume 154, pages 1201-1208, DOI:10.1016/j.proeng.2016.07.524.
  • Walker DJ, Keedwell EK. (2016) Towards Many-objective Optimisation with Hyper-heuristics: Identifying Good Heuristics with Indicators, PPSN 2016 14th International Conference on Parallel Problem Solving from Nature, Edinburgh Scotland, 17th - 21st Sep 2016, Lecture Notes in Computer Science.
  • Walker DJ, Keedwell EK. (2016) Multi-objective Optimisation with a Sequence-based Selection Hyper-heuristic, The Genetic and Evolutionary Computation Conference (GECCO 2016), Denver, Colorado, Usa, 20th - 24th Jul 2016.
  • Guidolin M, Chen AS, Ghimire B, Keedwell EC, Djordjevic S, Savic DA. (2016) A weighted cellular automata 2D inundation model for rapid flood analysis, Environmental Modelling and Software, volume 84, pages 378-394, DOI:10.1016/j.envsoft.2016.07.008. [PDF]

2015

2014

  • Keedwell EC. (2014) Website dedicated to the work undertaken in this grant.
  • Keedwell EC. (2014) Website dedicated to the work on this grant.
  • Mwaura J, Keedwell E. (2014) On using gene expression programming to evolve multiple output robot controllers, IEEE SSCI 2014 - 2014 IEEE Symposium Series on Computational Intelligence - IEEE ICES: 2014 IEEE International Conference on Evolvable Systems, Proceedings, pages 173-180, DOI:10.1109/ICES.2014.7008737.
  • Narayanan A, Keedwell E. (2014) An evolutionary computational approach to phase and synchronization in biological circuits, 2014 10th International Conference on Natural Computation, ICNC 2014, pages 419-424, DOI:10.1109/ICNC.2014.6975872.
  • Sapin E, Keedwell EC, Frayling T. (2014) Ant Colony Optimisation of Decision Trees for the Detection of Gene-Gene Interactions, IEEE International Conference on Bioinformatics and Biomedicine, Belfast, 2nd - 5th Nov 2014.
  • Keedwell EC. (2014) An Analysis of the Area Under the ROC Curve and its Use as a Metric for Comparing Clinical Scorecards, IEEE International Conference on Bioinformatics and Biomedicine, Belfast, 2nd - 5th Nov 2014.
  • Gibson M, Keedwell EC, Savic D. (2014) An investigation of the efficient implementation of Cellular Automata on multi-core CPU and GPU hardware, Journal of Parallel and Distributed Computing, DOI:10.1016/j.jpdc.2014.10.011.
  • Marchi A, Salomons E, Ostfeld A, Kapelan Z, Simpson A, Zecchin A, Maier H, Wu Z, Elsayed S, Song Y. (2014) The Battle of the Water Networks II (BWN-II), Journal of Water Resources Planning and Management, volume 140, no. 7, DOI:10.1061/(ASCE)WR.1943-5452.0000378.
  • Walker D, Keedwell EC, Savic D, Kellagher R. (2014) An Artificial Neural Network-based Rainfall Runoff Model for Improved Drainage Network Modelling, 11th International Conference on Hydroinformatics, New York, 17th - 21st Aug 2014.
  • Gibson M, Keedwell EC, Savic D. (2014) Genetic Programming for Cellular Automata Urban Inundation Modelling, 11th International Conference on Hydroinformatics, New York, 17th - 21st Aug 2014.
  • Johns M, Keedwell EC, Savic D. (2014) Interactive 3D Visualisation of Optimisation for Water Distribution Systems, 11th International Conference on Hydroinformatics, New York, 17th - 21st Aug 2014.
  • Johns M, Keedwell EC, Savic D. (2014) Multi-objective Pipe Smoothing Genetic Algorithm for Water Distribution Network Design, 11th International Conference on Hydroinformatics, New York, 17th - 21st Aug 2014.
  • McClymont K, Keedwell EC, Savic D, Randall-Smith M. (2014) Automated construction of evolutionary algorithm operators for the bi-objective water distribution network design problem using a genetic programming based hyper-heuristic approach, JOURNAL OF HYDROINFORMATICS, volume 16, no. 2, pages 302-318, DOI:10.2166/hydro.2013.226. [PDF]
  • Townsend J, Keedwell E, Galton A. (2014) Artificial Development of Biologically Plausible Neural-Symbolic Networks, Cognitive Computation, volume 6, no. 1, pages 18-34, DOI:10.1007/s12559-013-9217-0.

2013

  • Hart E, Sim K, McClymont K, Keedwell E. (2013) Session details: Workshop on problem understanding and real-world optimisation, Proceedings of the 15th annual conference companion on Genetic and evolutionary computation, DOI:10.1145/3253586.
  • Duncan AP, Tyrrell D, Smart N, Keedwell EC, Djordjevic S, Savic DA. (2013) Comparison of machine learning classifier models for bathing water quality exceedances in UK, IAHR35, Chengdu, China, 8th - 13th Sep 2013, Proceedings of 2013 IAHR Congress, volume 35.
  • Mwaura J, Keedwell EC, Engelbrecht A. (2013) Evolved Linker Gene Expression Programming: A New Technique for Symbolic Regression, 1st BRICS Countries Congress (BRICS-CCI) and 11th Brazilian Congress (CBIC) on Computational Intelligence, Recife, Brazil, 8th - 11th Sep 2013. [PDF]
  • Townsend J, Keedwell EC, Galton A. (2013) Evolution of Connections in SHRUTI Networks, 9th International Workshop on Neural-Symbolic Learning and Reasoning (NeSy13).
  • Kucukkoc I, Zhang DZ, Keedwell EC. (2013) Balancing parallel two-sided assembly lines with ant colony optimisation algorithm, 2nd Symposium on Nature-Inspired Computing and Applications, NICA 2013 - AISB Convention 2013, pages 21-28.
  • Duncan AP, Chen AS, Keedwell EC, Djordjević S, Savić DA. (2013) RAPIDS: Early Warning System for Urban Flooding and Water Quality Hazards, Machine Learning in Water Systems symposium, AISB 2013.
  • Sapin E, Keedwell E, Frayling T. (2013) Ant Colony Optimisation for Exploring Logical Gene-Gene Associations in Genome Wide Association Studies, PROCEEDINGS IWBBIO 2013: INTERNATIONAL WORK-CONFERENCE ON BIOINFORMATICS AND BIOMEDICAL ENGINEERING, pages 449-+. [PDF]
  • Butt E, Foster JAH, Keedwell E, Bell JEA, Titball RW, Bhangu A, Michell SL, Sheridan R. (2013) Derivation and validation of a simple, accurate and robust prediction rule for risk of mortality in patients with Clostridium difficile infection, BMC Infectious Diseases, volume 13, no. 1, DOI:10.1186/1471-2334-13-316.
  • Savić DA, Bicik J, Morley MS, Duncan A, Kapelan Z, Djordjević S, Keedwell EC. (2013) Intelligent urban water infrastructure management, Journal of the Indian Institute of Science, volume 93, no. 2, pages 319-336.
  • Keedwell E, Narayanan A. (2013) Gene expression rule discovery and multi-objective ROC analysis using a neural-genetic hybrid, Int J Data Min Bioinform, volume 7, no. 4, pages 376-396, DOI:10.1504/ijdmb.2013.054225. [PDF]
  • Townsend J, Keedwell EC, Galton A. (2013) Artificial Development of Connections in SHRUTI Networks Using a Multi-Objective Genetic Algorithm, Genetic and Evolutionary Computation Conference, Amsterdam, 5th - 10th Jul 2013.
  • Gibson M, Keedwell EC, Savic D. (2013) Understanding the efficient parallelisation of Cellular Automata on CPU and GPGPU hardware, Genetic and Evolutionary Computation Conference, Amsterdam, 5th - 10th Jul 2013.
  • Johns M, Keedwell EC, Savic D. (2013) Pipe Smoothing Genetic Algorithm for Least Cost Water Distribution Network Design, Genetic and Evolutionary Computation Conference, Amsterdam, 5th - 10th Jul 2013.
  • Sapin E, Keedwell EC, Frayling T. (2013) Subset-Based Ant Colony Optimisation for the Discovery of Gene-Gene Interactions in Genome Wide Association Studies, Genetic and Evolutionary Computation Conference, Amsterdam, 5th - 10th Jul 2013.
  • Ghimire B, Chen AS, Guidolin M, Keedwell EC, Djordjević S, Savić DA. (2013) Formulation of fast 2D urban pluvial flood model using cellular automata approach, Journal of Hydroinformatics, volume 15, no. 3, pages 676-686, DOI:10.2166/hydro.2012.245. [PDF]

2012

  • McClymont K, Keedwell E. (2012) Session details: Understanding problems (GECCO-UP), Proceedings of the 14th annual conference companion on Genetic and evolutionary computation, DOI:10.1145/3245069.
  • Townsend J, Galton A, Keedwell E. (2012) A scalable genome representation for neural-symbolic networks, AISB/IACAP World Congress 2012: 1st Symposium on Nature Inspired Computation and Applications, NICA 2012, Part of Alan Turing Year 2012.
  • Wang Q, Liu H, McClymont K, Johns M, Keedwell E. (2012) A hybrid of multi-phase optimisation and iterated manual intervention for BWN-II, 14th Water Distribution Systems Analysis Conference 2012, WDSA 2012, volume 1, pages 126-132.
  • Keedwell E, Morley M, Croft D. (2012) Continuous Trait-Based Particle Swarm Optimisation (CTB-PSO), Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), volume 7461 LNCS, pages 342-343, DOI:10.1007/978-3-642-32650-9_37.
  • Foster JAH, Butt JEC, Bell J, Goff A, Morgan C, Hancock J, Carmichael C, Keedwell EC, Michell SLI, Sheridan RP. (2012) IMPROVING CLINICAL MANAGEMENT IN CLOSTRIDIUM DIFFICILE: FAECAL CALPROTECTIN DOES NOT PREDICT SEVERITY, RECURRENCE OR MORTALITY, AGE AND AGEING, volume 41, pages 72-72. [PDF]
  • Ghimire B, Chen AS, Guidolin M, Keedwell EC, Djordjević S, Savić DA. (2012) A new two-dimensional cellular automata approach for fast urban flood inundation modelling.
  • Duncan AP, Chen AS, Keedwell EC, Djordjević S, Savić DA. (2012) Urban flood prediction in real-time from weather radar and rainfall data using artificial neural networks, Weather Radar and Hydrology: IAHS Red Book Proceedings.
  • McClymont K, Keedwell EC, Savic D. (2012) AUTOMATED CONSTRUCTION OF HEURISTICS FOR THE WATER DISTRIBUTION NETWORK DESIGN PROBLEM, 10th International Conference on Hydroinformatics, Hamburg, Germany, 14th - 18th Jul 2012.
  • Johns M, Keedwell EC, Savic D. (2012) GRAMMAR CONSTRAINED GENETIC ALGORITHM FOR LEAST-COST WATER DISTRIBUTION NETWORK DESIGN, 10th International Conference on Hydroinformatics, Hamburg, Germany, 14th - 18th Jul 2012.
  • Ghimire B, Chen A, Guidolin M, Keedwell EC, Djordjevic S, Savic D. (2012) A NEW TWO-DIMENSIONAL CELLULAR AUTOMATA APPROACH FOR FAST URBAN FLOOD INUNDATION MODELLING, 10th International Conference on Hydroinformatics, Hamburg, Germany, 14th - 18th Jul 2012.
  • Guidolin M, Duncan A, Ghimire B, Gibson M, Keedwell EC, Djordjevic S, Savic D. (2012) CADDIES: A NEW FRAMEWORK FOR RAPID DEVELOPMENT OF PARALLEL CELLULAR AUTOMATA ALGORITHMS FOR FLOOD SIMULATION, 10th International Conference on Hydroinformatics, Hamburg, Germany, 14th - 18th Jul 2012.

2011

  • McClymont K, Keedwell E, Savić D, Randall-Smith M. (2011) A hyper-heuristic approach to water distribution network design, Urban Water Management: Challenges and Oppurtunities - 11th International Conference on Computing and Control for the Water Industry, CCWI 2011, volume 3.
  • Christmas JT, Keedwell EK, Frayling TM, Perry JR. (2011) Ant colony optimisation to identify genetic variant association with type 2 diabetes, Information Sciences, volume 181, no. 9, pages 1609-1622, DOI:10.1016/j.ins.2010.12.005.
  • McClymont K, Walker D, Keedwell E, Everson R, Fieldsend JE, Savic D, Randall-Smith M. (2011) Novel Methods for Ranking District Metered Areas for Water Distribution Network Maintenance Scheduling, CCWI, Exeter, 5th - 7th Sep 2011. [PDF]
  • Mwaura J, Keedwell E. (2011) Evolving modularity in robot behaviour using gene expression programming, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), volume 6856 LNAI, pages 392-393, DOI:10.1007/978-3-642-23232-9_43.
  • Duncan AP, Chen AS, Keedwell EC, Djordjević S, Savić DA. (2011) Urban flood prediction in real-time from weather radar and rainfall data using artificial neural networks, Weather Radar and Hydrology (WRaH) Symposium 2011.
  • Schellart A, Ochoa S, Simões N, Wang L, Rico-Ramirez M, Liguori S, Duncan A, Chen AS, Keedwell E, Djordjević S. (2011) Urban pluvial flood modelling with real time rainfall information - UK case studies, 12nd International Conference on Urban Drainage.
  • Guidolin M, Duncan A, Keedwell EC, Chen AS, Djordjević S, Savić DA. (2011) Design of a graphical framework for simple prototyping of pluvial flooding cellular automata algorithms, Computing and Control for the Water Industry 2011, Exeter, Uk, 5th - 7th Sep 2011, Urban Water Management - Challenges and Opportunities, volume 1, pages 205-210.
  • Galton A, Keedwell EC, Barclay M. (2011) Reference Object Selection Intelligence Test, AISB Convention, York, Uk, 4th - 7th Apr 2011.
  • McClymont K, Keedwell EC. (2011) Benchmark Multi-objective Optimisation Test Problems with Mixed Encodings, IEEE Congress on Evolutionary Computation, New Orleans, 5th - 8th Jun 2011, 5 Jun 2011.
  • Christmas J, Keedwell E, Frayling T, Perry J. (2011) Ant Colony Optimisation to identify Genetic Disease Association for Type 2 Diabetes, Information Sciences, volume 181, no. 9, pages 1609-1622, DOI:10.1016/j.ins.2010.12.005.

2010

  • Keedwell E. (2010) Towards a Staged Developmental Intelligence Test for Machines, AISB 2010, Leicester, Uk, 30th Mar - 4th Apr 2010.

2009

  • Rogers C, Randall-Smith M, Keedwell E, Diduch R. (2009) Application of Optimization Technology to Water Distribution System Master Planning, World Environmental and Water Resources Congress, Kansas City, Missouri, 17th - 21st May 2009.
  • Keedwell E, Narayanan A. (2009) Gene Expression Classification using Multi-Objective Ensembles, AISB 2009, Edinburgh, Uk, 6th - 9th Apr 2009.
  • Keedwell E, Mwaura J. (2009) Adaptive Gene Expression Programming Using a Simple Feedback Heuristic, AISB 2009, Edinburgh, Uk.

2008

  • Baker L, Keedwell E, Randall-Smith M. (2008) Ant colony optimisation for large-scale water distribution network optimisation, AISB 2008 Convention: Communication, Interaction and Social Intelligence - Proceedings of the AISB 2008 Symposium on Swarm Intelligence Algorithms and Applications, pages 44-50.
  • Gou YE, Walters GA, Khu ST, Keedwell EC. (2008) Efficient multiobjective storm sewer design using cellular automata and genetic algorithm hybrid, Journal of Water Resources Planning and Management, volume 134, no. 6, pages 511-515, DOI:10.1061/(ASCE)0733-9496(2008)134:6(511).

2007

  • Guo Y, Walters GA, Khu ST, Keedwell EC. (2007) A novel cellular automata based approach to storm sewer design, Engineering Optimization, volume 39, no. 3, DOI:10.1080/03052150601128261.
  • Rogers C, Randall-Smith M, Keedwell E, Diduch R. (2007) Application of optimization technology to the development of a water distribution system master plan for the City of Ottawa, American Water Works Association - AWWA Annual Conference and Exposition, ACE 2007.
  • Guo Y, Walters GA, Khu S-T, Keedwell E. (2007) Hybridizing Cellular Automata principles and NSGAII for multi-objective design of urban water networks.
  • Keedwell E, Narayanan A. (2007) Gene finding and rule discovery with a multi-objective neural-genetic hybrid, Proceedings of GECCO 2007: Genetic and Evolutionary Computation Conference, DOI:10.1145/1276958.1277048.

2006

  • Guo Y, Walters GA, Khu S-T, Keedwell E. (2006) Optimal Design of Sewer Networks using hybrid cellular automata and genetic algorithm, Proc. IWA World Water Congress 2006, Beijing, China.
  • Keedwell E, Khu S-T. (2006) A novel evolutionary metaheuristic for the multi-objective optimisation of real-world water distribution networks, Engineering Optimisation.
  • Keedwell E, Khu ST. (2006) A novel cellular automata approach to optimal water distribution network design, Journal of Computing in Civil Engineering, volume 20, no. 1, pages 49-57, DOI:10.1061/(ASCE)0887-3801(2006)20:1(49).
  • Guo Y, Walters GA, Khu S-T, Keedwell E. (2006) A robust hybrid approach for efficient optimal design of sewer systems, 7th International Conference on Hydroinformatics, Nice, France.
  • Randall-Smith M, Rogers C, Keedwell E, Diduch R, Kapelan Z. (2006) Optimized Design of the City of Ottawa Water Network: A Genetic Algorithm Case Study, 8th Annual Water Distribution System Analysis Symposium, Cincinnati, Ohio, Usa.

2005

  • Guo Y, Walters G, Khu ST, Keedwell E. (2005) Transition rule configuration for cellular automata based optimal sewerage design, Proceedings of the 8th International Conference on Computing and Control for the Water Industry, CCWI 2005: Water Management for the 21st Century, volume 2.
  • Keedwell EC, Narayanan A. (2005) Intelligent Bioinformatics: The Application of Artificial Intelligence Techniques to Bioinformatics Problems, Wiley & Sons.
  • Cheung A, Narayanan A, Gamalielsson J, Keedwell EC, Vercellone C. (2005) Artificial Neural Networks for Reducing the Dimensionality of Gene Expression Data, Bioinformatics Using Computational Intelligence Paradigms.
  • Keedwell E, Narayanan A. (2005) Discovering gene networks with a neural-genetic hybrid, IEEE/ACM Trans Comput Biol Bioinform, volume 2, no. 3, pages 231-242, DOI:10.1109/TCBB.2005.40. [PDF]
  • Jonkergouw P, Keedwell E, Khu S-T. (2005) Modelling chlorine decay in water networks using genetic programming, International conference on natural and adaptive computing algorithms ICANNGA 2005.
  • Juliusdottir T, Keedwell E, Corne D, Narayanan A. (2005) Two-phase EA/k-NN for feature selection and classification in cancer microarray datasets, Proceedings of the 2005 IEEE Symposium on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB '05, volume 2005, DOI:10.1109/cibcb.2005.1594891.
  • Khu ST, Keedwell E. (2005) Introducing choices (flexibility) in upgrading of water distribution network: The New York City Tunnnel Network example, Engineering Optimization, volume 37, no. 3, pages 291-305.
  • Khu S-T, Han D, Keedwell E, Pollard O. (2005) An improved operational rainfall-runoff model using a coupled deterministic-based model and data-driven technique, Seoul.
  • Keedwell E, Khu ST. (2005) A Hybrid Genetic Algorithm for the Design of Water Distribution Networks, Engineering Applications of Artificial Intelligence, volume 18, no. 4, pages 461-472.
  • Krishna A, Narayanan A, Keedwell EC. (2005) Dissecting the Biological Motherboard,Systems Biology and Beyond, "International conference on natural and adaptive computing algorithms,Springer Wein NewYork", pages 325-328.
  • Krishna A, Narayanan A, Keedwell EC. (2005) Reverse engineering gene networks with artificial neural networks, Adaptive and Natural Computing Algorithms, pages 325-328, DOI:10.1007/3-211-27389-1_78. [PDF]
  • Krishna A, Narayanan A, Keedwell EG. (2005) Neural networks and temporal gene expression data, Lecture Notes in Computer Science, volume 3449, pages 64-73, DOI:10.1007/978-3-540-32003-6_7.

2004

  • Narayanan A, Evangelia, Keedwell E. (2004) Analyzing gene expression data for childhood medulloblastoma survival with artificial neural networks, Proceedings of the 2004 IEEE Symposium on Computational Intelligence in Bioinformatics and Computational Biology, CIBCB'04, pages 9-16.
  • Khu ST, Keedwell EC, Pollard O. (2004) An evolutionary-based real-time updating technique for an operational rainfall-runoff forecasting model, International Environmental Modelling and Software Society, IEMSS2004, Osnabruck Germany. [PDF]
  • Keedwell E, Khu ST. (2004) Hybrid genetic algorithms for multi-objective optimisation of water distribution networks, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), volume 3103, pages 1042-1053, DOI:10.1007/978-3-540-24855-2_115.
  • Khu ST, Keedwell E. (2004) Using multi-objective genetic algorithm to achieve design flexibility for water distribution systems, Singapore.
  • Narayanan A, Keedwell EC, Gamalielsson J, Tatineni S. (2004) Single-layer artificial neural networks for gene expression analysis, Neurocomputing, volume 61, no. 1-4, pages 217-240, DOI:10.1016/j.neucom.2003.10.017.

2003

  • Bragalli C, Savic DA, Keedwell E. (2003) Water quality in multi-objective genetic algorithms for Integrated Design Support of water distribution networks, "'Proc. of the IWA-IAHR joint conference: Pumps, Electromechanical Devices and Systems PEDS 2003, Netherlands: Balkema'", pages 27-35.
  • Keedwell E, Narayanan A. (2003) Genetic algorithms for gene expression analysis, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), volume 2611, pages 76-86, DOI:10.1007/3-540-36605-9_8.
  • Keedwell E, Khu S-T. (2003) More choices in water system design through hybrid optimisation, Proc. Int. Conf. on Computing and Control for the Water Industry, pages 257-264.

2002

  • Keedwell E, Narayanan A, Savic DA. (2002) Modelling Gene Regulatory Data Using Artificial Neural Networks, "'Proceedings of the International Joint Conference on Neural Networks (IJCNN), The 2002 IEEE World Congress on Computational Intelligence, Honolulu, Hawaii 12-17 May, Proceedings published on CD by the IEEE'", pages 183-188.
  • Narayanan A, Keedwell EC, Olsson B. (2002) Artificial intelligence techniques for bioinformatics, Appl Bioinformatics, volume 1, no. 4, pages 191-222. [PDF]
  • Keedwell E, Narayanan A, Savic DA. (2002) Modelling Gene Regulatory Data Using Artificial Neural Networks, "'Proceedings of the International Joint Conference on Neural Networks (IJCNN), The 2002 IEEE World Congress on Computational Intelligence, Honolulu, Hawaii 12-17 May, Proceedings published on CD by the IEEE'", pages 183-188.

2000

  • Keedwell E, Bessler F, Narayanan A, Savic DA. (2000) From data mining to rule refining A new tool for post data mining rule optimisation, ICTAI, pages 82-85, DOI:10.1109/TAI.2000.889849. [PDF]
  • Keedwell EC, Narayanan A. (2000) Creating rules from trained networks using genetic algorithms, International Journal of Computers Systems and Signals, volume 1.
  • Keedwell E, Narayanan A, Savic D. (2000) Evolving rules from neural networks trained on continuous data, PROCEEDINGS OF THE 2000 CONGRESS ON EVOLUTIONARY COMPUTATION, VOLS 1 AND 2, pages 639-645. [PDF]

1999

  • Keedwell E, Narayanan A, Savic D. (1999) Using genetic algorithms to extract rules from trained neural networks, GECCO-99: PROCEEDINGS OF THE GENETIC AND EVOLUTIONARY COMPUTATION CONFERENCE, pages 793-793. [PDF]

Back to top


Further information

Teaching

Prof. Keedwell currently teaches/supervises projects in:

  • ECMM423 Evolutionary Computation and Optimisation
  • ECM3001 Commercial and Industrial Experience
  • ECM3401 Individual Literature Review and Project
  • ECMM4xx MSc Projects

He has previously taught all or parts of:

  • ECM3412/ECMM409 Nature Inspired Computation
  • ECM3408 Enterprise Computing
  • ECM2417 Frontiers of Computing
  • ECMM214 Hydroinformatics Tools

Older taught modules (no longer running) include:

  • COM3521 Artificial Intelligence and Software Engineering
  • BIOM505  Machine learning techniques and information systems
  • COM2408 Symbolic Artificial Intelligence
  • ECMM418 ITMB Case Studies
  • ECMM412 Machine Learning and Optimisation

Professional Activities

Prof. Keedwell has acted as reviewer/member of the programme committee for a number of conference series and journals, including the following:

Conferences

  • Genetic and Evolutionary Computing Conference (GECCO) 2005-2007 & 2014-
  • Computational Intelligence in Bioinformatics and Computational Biology (CIBCB) 2005-Present
  • Congress on Evolutionary Computing (CEC) 2006-Present
  • International Symposium on Bioinformatics Research and Applications (ISBRA) 2007, 2008
  • International Conference on Evolutionary Computing Theory and Applications 2007-Present

Journals

  • Editor: Cogent Engineering
  • Guest Editor: Journal of Hydroinformatics "Making Water Smart"
  • Reviewer for a range of journals including:
  • Artificial Intelligence
  • Biosystems
  • IEEE/ACM Transactions on Computational Biology and Bioinformatics,
  • IEEE Transactions on Systems, Man and Cybernetics
  • IEEE Transactions on Evolutionary Computation
  • IEEE Transactions on Nanobioscience
  • Information Sciences
  • Natural Computing
  • ASCE Journal of Computing in Civil Engineering,
  • ASCE Journal of Water Resources Planning and Management
  • Environmental Modelling and Software
  • Journal of Biomedical Informatics
  • Knowledge and Information Systems: An International Journal
  • BMC Systems Biology
  • Urban Water

Research

Publications

Prof. Keedwell has published one book 'Intelligent Bioinformatics' and over 120 papers in journals and conferences in the fields of computer science, bioinformatics and hydroinformatics.  Details of these are available on the publications page.  Further details are also available on Google Scholar and Researchgate

Funded Projects

Prof. Keedwell has successfully applied for projects worth over £2m in total.  This funding has predominantly been provided by the EPSRC, with projects also funded by Innovate UK (was the Technology Strategy Board) and by industry. 

Previous/Current Project Websites

Ant Colony Optimisation for Genome-Wide Association Studies (ACOGWAS)

This EPSRC funded project is investigating the use of a modern nature-inspired optimisation technique known as ant colony optimisation to discover associations between groups of small changes in DNA and a variety of diseases including type I and type II diabetes, inflammatory bowel disease and rheumatoid arthritis.  The developed techniques are designed to search the billions of combinations of small changes in DNA over populations of thousands of people to determine those that might lead to a greater susceptability to disease.
In many cases, it is not possible to search every combination, even with very powerful computers and so this project is using one of a family of stochastic optimisation algorithms to perform this task.  You can read more about the biology and computer science of this project, as well as read about the progress we are making in the sections below.

Ant Colony Optimisation (ACO)

The first ACO algorithm, proposed by Marco Dorigo in 1996 [1] was Ant System and since then a number of modifications have been proposed to improve performance of the algorithm.  ACO is inspired by the way that populations of ants in nature can find a short path between the nest and a food source by sensing the pheromone trails of previous ants. The computational algorithm requires ants to traverse a topology which is either a direct representation of the problem (e.g. in the travelling salesman problem) or a landscape of variable choices known as a construction graph. We have adapted the algorithm to work with the specific challenges of Genome-Wide Association Studies.

Subset-Based ACO

Standard ACO works by laying pheromone on the links between variables choices, which works well for combinatorial problems.  However, for subset based problems (i.e. selecting a small number of N from a larger set), this method does not work well and incurs high computational complexity.  We have therefore made use of a method first proposed by Leguizamon and Michaelawicz (1999) [2] which lays pheromone on the variables (SNPs) themselves rather than the links between them, reducing the search space considerably and speeding execution.

Tournament Selection ACO (T-ACO)

Standard ACO path selection works by constructing a roulette wheel of variable choices for the ants to consider, weighted by the pheromone (and possibly a local heuristic) on each choice. This works well for a wide range of problems, but the specifics of GWAS mean that such a roulette wheel would require ~400,000 such segments, leading to poor path selection pressure and large computational loads. To counter this, we developed a tournament-based ACO algorithm [3] that chooses paths by randomly selecting N SNPs and choosing the SNP with the best pheromone value for the ant to explore. Experimentation has shown the effectiveness of this method on high-dimensional problems and improved execution times for reasonable N.

P-Values of randomly generated gene-gene interactions in Type-II diabetes data.

GWAS Using ACO

We have applied the above methods to data from the Wellcome Trust Case Control Consortium involving a number of different diseases. The datasets consist of around 400,000 SNPs and over 5000 individuals (~3000 controls, ~2000 cases) that must be explored by the ACO approaches. The size of this dataset has presented some challenges from a computational perspective and the flexibility of the ACO approach is such that we have been able to explore interesting relationships between SNPs.  These are presented in more detail below:

Byte-Level Data Compression

With between 400,000 and 500,000 SNPs and 5000 individuals, storing the datasets in working memory became a problem. Using a single byte for each SNP would require between 2 and 2.5 GB of free space in RAM which can result in the OS paging to disk even with 4 GB or more of memory. We have therefore developed a compression technique that allows 4 SNPs to be stored in a single byte that reduces runtimes and allows for multiple runs to be made in parallel on the same machine if required [4].

Logical Operations

The traditional view of gene-gene interactions is that they operate with an AND operation, for instance - SNP1=GENOTYPE X AND  SNP2=GENOTYPE Y is commonly used. We have explored some of the relationships that might exist between SNPs that are not represented by this AND operation, and have included OR, XOR and NOT to provide the algorithm more flexibility in the way in which it allows SNPs to interact [5].  Results have shown that these additional operations provide additional power to the algorithm to find statistically important associations in the data.

Assessing Statistical Significance

A key issue with developing these methods is the correction for multiple testing. With single associations, it is relatively simple to correct p-values using Bonferroni correction, but a stochastic algorithm discovering gene-gene interactions is more difficult. We have used monte-carlo methods to select 1,000,000 random combinations from the data and have also run the algorithm thousands of time on shuffled data to determine the statistical significance of results.

Summary

This ongoing EPSRC project has yielded interesting developments in ACO, the representation of data and significance testing and has shown promise in discovering gene-gene interactions in large-scale GWAS data.

References

  1. M. Dorigo, V. Maniezzo, et A. Colorni (1996) Ant system: optimization by a colony of cooperating agents, IEEE Transactions on Systems, Man, and Cybernetics--Part B , volume 26, numéro 1, pages 29-41
  2. G. Leguizamon and Z. Michalewicz, (1999) "A new version of ant system for subset problems," Proceedings of the 1999  Congress on Evolutionary Computation(CEC 99), vol.2, pp.1458-1464, 1999.
  3. Sapin E, Keedwell EC. (2012) "T-ACO - Tournament Ant Colony Optimisation for High Dimensional Problems", ECTA 2012 - 4th International Conference on Evolutionary Computation Theory and Applications, Barcelona, Spain, 5th - 7th Oct 2012.
  4. Sapin E, Keedwell EC., Frayling T (2013) "Subset-Based ACO for Genome Wide Association Study:  Discovery of Promising Combinations" to appear at Genetic and Evolutionary Computing Conference GECCO 2013
  5. Sapin E, Keedwell EC., Frayling T (2013) "Ant Colony Optimisation for Exploring Logical Gene-Gene Associations in Genome Wide Association Studies", International Work Conference on Bioinformatics (IWBBIO), Granada, Spain.

Pheromone trail for initial ACO algorithm over time note the inflection point B2

Sequence Analysis Based Hyperheuristics for Real World Problems (SEQAH)

Hyperheuristics are a set of techniques that optimise at the level above metaheuristics (such as evolutionary algorithms and particle swarm optimisation).  They can be used to discover near-optimal combinations of low level heuristics (e.g. mutation and crossover operations) and are known as selection hyperheuristics or can generate new heuristics to aid in solving difficult problems and are known as generation hyperheuristics. We have developed new methods for both types of hyperheuristics and have applied the resulting algorithms to a wide variety of problems, including those in the water industry.

Sequence Analysis Based Hyperheuristics for Real-World Problems

This EPSRC-funded project is investigating the use of sequence analysis techniques taken from bioinformatics in the development of hyperheuristics.  Hyper-heuristics are optimisation approaches that operate at the level above traditional metaheuristics (e.g. evolutionary algorithms) and must either select from a pre-defined set of low level heuristics (e.g. mutation, crossover and hillclimbing operators) or generate them.  The focus in the SEQAH project is on developing selection hyperheuristics by analysing and creating sequences of heuristic usage during an optimisation.  The first method developed in this project is known as SSHH and uses a hidden markov model to analyse and produce sequences of heuristics.

Sequence Based Selection Hyperheuristic (SSHH)

SSHH uses a hidden markov model where hidden states are replaced with low level heuristics (LLH) and a matrix of transition probabilities determine the movement between these states.  A further set of emission probabilities determine whether, given a particular LLH selection, the heuristic will be applied or will be coupled with another LLH to form a sequence.  The method can be seen in the diagram below and more detail can be found in this paper (best paper nominee, GECCO 2015).

Promising early results obtained with minimal tuning or problem specific expertise:

  1. SSHH obtains better performance than the winner of the Cross Domain Heuristic Challenge (CHESC) 2011
  2. Ranked 3rd in the Windfarm Optimisation Challenge at GECCO 2015
  3. Outperformed genetic algorithm formulations on the water distribution network optimisation problem

SSHH Operation.  AS= Acceptance Strategy (1=evaluate new solution, 2=move to next LLH), LLH = Low Level Heuristic

Multi-objective hyperheuristics and discolouration propensity modelling

In this project, we are investigating new hyperheuristics for continuous & mixed encoding multi-objective problems and are applying them to problems in the water industry, including the mitigation of discolouration risk.  The project is supported by EPSRC and consultants Mouchel Ltd.

Heuristics evolved for solving multi-objective water distribution network problems with discolouration risk taken from the pareto-front of operations.  The probability distributions show the probability of mutation around the current point and the point clouds demonstrate the resulting behaviours in optimisation space.

  • Kheiri A., Keedwell E. (2015) Markov Chain Selection Hyper-heuristic for the Optimisation of Constrained Magic Squares accepted to UKCI 2015, Exeter, UK
  • Kheiri A., Keedwell E., Gibson, M., Savic, D., (2015) Sequence analysis-based hyper-heuristics for water distribution network optimisation accepted to Computing and Control in the Water Industry (CCWI) 2015, Leicester, UK
  • Kheiri, A., Keedwell E., (2015) A sequence-based selection hyper-heuristic utilising a hidden Markov model in the proceedings of Genetic and Evolutionary Computation Conference (GECCO) 2015, Madrid, Spain, p417-424 (best paper nominee)
  • McClymont K, Keedwell E, Savic D. (2015) An analysis of the interface between evolutionary algorithm operators and problem features for water resources problems. A case study in water distribution network design, Environmental Modelling and Software, DOI:10.1016/j.envsoft.2014.12.023
  • McClymont K., Keedwell E., (2011) “Deductive Sort and Climbing Sort: New Methods for Non-Dominated Sorting” Evolutionary Computation 2011, MIT Press
  • McClymont K., Keedwell E., (2011) “Benchmark Multi-objective Optimisation Test Problems with Mixed Encodings” IEEE Congress on Evolutionary Computation 2011, New Orleans
  • McClymont K., Keedwell E., (2011) “Markov chain Hyper-heuristic (MCHH): an Online Selective Hyper-heuristic for Multi-objective Continuous Problems” Genetic and Evolutionary Computing Conference 2011, Dublin, Ireland
  • McClymont K., Keedwell E., (2010) “Optimising Multi-Modal Polynomial Mutation Operators for Multi-Objective Problem Classes” IEEE Congress on Evolutionary Computation 2010, Barcelona

Back to top