FDA - a scalable evolutionary algorithm for the optimization of additively decomposed functions. Export

@article{citeulike:348976, abstract = {The Factorized Distribution Algorithm (FDA) is an evolutionary algorithm which combines mutation and recombination by using a distribution. The distribution is estimated from a set of selected points. In general, a discrete distribution defined for n binary variables has 2(n) parameters. Therefore it is too expensive to compute. For additively decomposed discrete functions (ADFs) there exist algorithms which factor the distribution into conditional and marginal distributions. This factorization is used by FDA. The scaling of FDA is investigated theoretically and numerically. The scaling depends on the ADF structure and the specific assignment of function values. Difficult functions on a chain or a tree structure are solved in about O(n radical n) operations. More standard genetic algorithms are not able to optimize these functions. FDA is not restricted to exact factorizations. It also works for approximate factorizations as is shown for a circle and a grid structure. By using results from Bayes networks, FDA is extended to LFDA. LFDA computes an approximate factorization using only the data, not the ADF structure. The scaling of LFDA is compared to the scaling of FDA.}, address = {Theoretical Foundation GMD Lab., Real World Computing Partnership, GMD FZ Informationstechnik, 53754 St. Augustin, Germany. muehlenbein@gmd.de}, author = {Muehlenbein, H. and Mahnig, T.}, citeulike-article-id = {348976}, citeulike-linkout-0 = {http://view.ncbi.nlm.nih.gov/pubmed/10578027}, citeulike-linkout-1 = {http://www.hubmed.org/display.cgi?uids=10578027}, issn = {1063-6560}, journal = {Evol Comput}, keywords = {eda-theory, edas, fda}, number = {4}, pages = {353--376}, posted-at = {2007-02-08 16:57:23}, priority = {2}, title = {FDA - a scalable evolutionary algorithm for the optimization of additively decomposed functions.}, url = {http://view.ncbi.nlm.nih.gov/pubmed/10578027}, volume = {7}, year = {1999} }

TY - JOUR ID - citeulike:348976 L3 - citeulike-article-id:348976 N2 - The Factorized Distribution Algorithm (FDA) is an evolutionary algorithm which combines mutation and recombination by using a distribution. The distribution is estimated from a set of selected points. In general, a discrete distribution defined for n binary variables has 2(n) parameters. Therefore it is too expensive to compute. For additively decomposed discrete functions (ADFs) there exist algorithms which factor the distribution into conditional and marginal distributions. This factorization is used by FDA. The scaling of FDA is investigated theoretically and numerically. The scaling depends on the ADF structure and the specific assignment of function values. Difficult functions on a chain or a tree structure are solved in about O(n radical n) operations. More standard genetic algorithms are not able to optimize these functions. FDA is not restricted to exact factorizations. It also works for approximate factorizations as is shown for a circle and a grid structure. By using results from Bayes networks, FDA is extended to LFDA. LFDA computes an approximate factorization using only the data, not the ADF structure. The scaling of LFDA is compared to the scaling of FDA. IS - 4 JF - Evol Comput SN - 1063-6560 AD - Theoretical Foundation GMD Lab., Real World Computing Partnership, GMD FZ Informationstechnik, 53754 St. Augustin, Germany. muehlenbein@gmd.de EP - 376 TI - FDA - a scalable evolutionary algorithm for the optimization of additively decomposed functions. VL - 7 SP - 353 KW - eda-theory KW - edas KW - fda AU - Muehlenbein, H AU - Mahnig, T PY - 1999/// UR - http://view.ncbi.nlm.nih.gov/pubmed/10578027 ER -