Tomlinson, A. and Bull, L.
Symbiogenesis in learning classifier systems.
Artificial Life, 7 (1).
- Published Version
Publisher's URL: http://dx.doi.org/10.1162/106454601300328016
|Additional Information:||The class of problems which require the memory of previous inputs or actions - non-Markov problems - remain amongst the most difficiult for machine learning techniques. This paper describes how an evolutionary computing mechanism based on the hypothesised process leading to the evolution of eukaryotic cells from simpler organisms enables the solution of such problems for reinforcement learning. The research was undertaken as part of an EPSRC project (GR/R067848/01) exploring the use of Learning Classifier Systems for distributed road traffic junction control - a task in which memory is potentially useful.|
|Uncontrolled Keywords:||animat, cooperation, evolution, genetic linkage, symbiosis|
|Faculty/Department:||~Pre-2010 Faculty Structure > Environment and Technology > Bristol Institute of Technology|
~Pre-2012 Faculty Structure > Faculty of Environment and Technology > Artificial Intelligence Group
R. Upload account
|Deposited On:||22 Jan 2010 15:10|
|Last Modified:||24 Nov 2012 23:46|
Repository Staff Only: item control page
Total Document Downloads in Past 12 Months