Research
Peer-reviewed work spanning evolutionary algorithms, optimisation, machine learning, bioinformatics and artificial life.
A modern historical resource exploring the people, research and ideas of the 2004 Congress on Evolutionary Computation — and why they still matter to artificial intelligence today.
CEC2004 brought researchers, practitioners and students from around the world to Portland to discuss advances in evolutionary computation: systems that search, adapt and improve through mechanisms inspired by evolution and collective behaviour.
Peer-reviewed work spanning evolutionary algorithms, optimisation, machine learning, bioinformatics and artificial life.
Plenary lectures, tutorials, special sessions, competitions and conversations across a global technical community.
Ideas that continue to influence optimisation, adaptive systems, robotics and modern approaches to AI.
CEC2004 treated evolutionary computation broadly. Its themes now read like a map of several enduring areas in computational intelligence.
Explore the conceptsPopulations of candidate solutions improve through variation, evaluation and selection.
Simple agents interact locally, producing useful global behaviour without central control.
Methods explore not one perfect answer, but families of solutions balancing competing goals.
Computational models investigate emergence, adaptation and life-like behaviour in complex systems.
CEC2004 marked the twentieth anniversary of the iterated Prisoner’s Dilemma work with a special plenary lecture by political scientist Robert Axelrod. Documented in the original programme and proceedings
The lecture, recorded as The Future of Cooperation, connected game theory, adaptation and social behaviour — a reminder that evolutionary computation has always reached beyond optimisation alone.
Meet the speakersThe congress gathers a broad international community around evolutionary algorithms, swarm methods, games, bioinformatics and adaptive systems.
Improved computing and parallel hardware expand experiments in robotics, automated design, optimisation and neuroevolution.
Evolutionary search complements gradient-based learning in architecture search, agent behaviour, quality-diversity and difficult black-box problems.
This restoration is grounded in the original conference pages, IEEE material, bibliographic records and surviving university pages. Where evidence is incomplete, the archive says so.