%0 Conference Paper %1 veenhuis:2005:CEC %A Veenhuis, Christian %A Köppen, Mario %A Krüger, Jörg %A Nickolay, Bertram %B Proceedings of the 2005 IEEE Congress on Evolutionary Computation %C Edinburgh, UK %D 2005 %E Corne, David %E Michalewicz, Zbigniew %E Dorigo, Marco %E Eiben, Gusz %E Fogel, David %E Fonseca, Carlos %E Greenwood, Garrison %E Chen, Tan Kay %E Raidl, Guenther %E Zalzala, Ali %E Lucas, Simon %E Paechter, Ben %E Willies, Jennifier %E Guervos, Juan J. Merelo %E Eberbach, Eugene %E McKay, Bob %E Channon, Alastair %E Tiwari, Ashutosh %E Volkert, L. Gwenn %E Ashlock, Dan %E Schoenauer, Marc %I IEEE Press %K PSO, TSO algorithms, genetic programming, %P 1238--1245 %T Tree Swarm Optimization: An Approach to PSO-based Tree Discovery %V 2 %X In recent years a swarm-based optimisation methodology called Particle Swarm Optimisation (PSO) has developed. PSO is highly explorative and primarily used in function optimisation. proposes a swarm-based learning algorithm based on PSO which is able to discover trees in tree spaces. Particles are flying through a tree space forming flocks around peaks of a fitness function. Because it inherits the explorative property of PSO, it needs only few evaluations to find suitable trees. %@ 0-7803-9363-5

@inproceedings{veenhuis:2005:CEC, abstract = {In recent years a swarm-based optimisation methodology called Particle Swarm Optimisation (PSO) has developed. PSO is highly explorative and primarily used in function optimisation. proposes a swarm-based learning algorithm based on PSO which is able to discover trees in tree spaces. Particles are flying through a tree space forming flocks around peaks of a fitness function. Because it inherits the explorative property of PSO, it needs only few evaluations to find suitable trees.}, added-at = {2008-06-19T17:46:40.000+0200}, address = {Edinburgh, UK}, author = {Veenhuis, Christian and K{\"o}ppen, Mario and Kr{\"u}ger, J{\"o}rg and Nickolay, Bertram}, biburl = {https://www.bibsonomy.org/bibtex/24cd00ffec922b9b2e9450a1da030bea5/brazovayeye}, booktitle = {Proceedings of the 2005 IEEE Congress on Evolutionary Computation}, editor = {Corne, David and Michalewicz, Zbigniew and Dorigo, Marco and Eiben, Gusz and Fogel, David and Fonseca, Carlos and Greenwood, Garrison and Chen, Tan Kay and Raidl, Guenther and Zalzala, Ali and Lucas, Simon and Paechter, Ben and Willies, Jennifier and Guervos, Juan J. Merelo and Eberbach, Eugene and McKay, Bob and Channon, Alastair and Tiwari, Ashutosh and Volkert, L. Gwenn and Ashlock, Dan and Schoenauer, Marc}, interhash = {ec8c9c738c7b09f66705e7cef2aa21f2}, intrahash = {4cd00ffec922b9b2e9450a1da030bea5}, isbn = {0-7803-9363-5}, keywords = {PSO, TSO algorithms, genetic programming,}, month = {2-5 September}, notes = {CEC2005 - A joint meeting of the IEEE, the IEE, and the EPS. symbolic regression, artificial ant, TSO. 'Tree swarm optimisation uses full trees for the particles.' All members of the swarm have the same structure and length. Number of dimensions in continious PSO space = size of terminal set + size of function set. No attempt at placing similar symbols near each other. 'position trees' and 'velocity trees'. Section 3.5 'Distances between full symbol trees' Four symbolic regression. 'But as in PSO the exact global optimum cannot be found very often for difficult problems.' Santa Fe artificial ant. 'In 64 percent of the runs, solutions collecting all food items are found.' UCI iris}, organisation = {IEEE Computational Intelligence Society, Institution of Electrical Engineers (IEE), Evolutionary Programming Society (EPS)}, pages = {1238--1245}, publisher = {IEEE Press}, publisher_address = {445 Hoes Lane, P.O. Box 1331, Piscataway, NJ 08855-1331, USA}, timestamp = {2008-06-19T17:53:38.000+0200}, title = {Tree Swarm Optimization: An Approach to {PSO}-based Tree Discovery}, volume = 2, year = 2005 }