Enhancing the Efficiency of the ECGA

Duque, Thyago S.P.C.; Goldberg, David E.; Sastry, Kumara

doi:10.1007/978-3-540-87700-4_17

Cited by 17 publications

(13 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Another study had shown that a simple bit-flipping hillclimber applied to all individuals can reduce the population requirements by an order of magnitude even on deceptive problems [3]. This kind of hybridization is easily parallelizable as independently improving each individual through local-search is an embarrassingly parallel task.…”

Section: Hybridization Of Ecgamentioning

confidence: 98%

“…The MPM concept is illustrated in Table 3 over a 3 bit problem with [1,3], [2] as partitions. The first and third bit are jointly distributed while variable [2] is independent.…”

Section: Case Study On Ecgamentioning

confidence: 99%

“…The first and third bit are jointly distributed while variable [2] is independent. The compound partition [1,3] can have four settings: {00, 01, 10, 11}. The probability distribution for the partitions is given Starting from a random population, the eCGA applies the process of evaluation, selection, MPMs based model-building and sampling until a halting criterion is met.…”

Section: Case Study On Ecgamentioning

confidence: 99%

“…Given the implied population sizes as the dimension of the problems increases, the computational cost of model building may quickly exceed economical practicality. Recent benchmarking and profiling results showed that easily more than 90% of EDAs running time may be spent in the model building phase [3].…”

Section: Introductionmentioning

confidence: 99%

“…More direct approaches aim to reduce the complexity of model building by restricting the search over a reduced set of variables in each epoch [3]. Heuristics of this type work well for problems where all or most of the variables are engaged in dependencies as randomly choosing and analyzing a partition will discover and exploit dependencies with a high probability.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Correlation guided model building

Icǎlnzan

Dumitrescu

Hirsbrunner

2009

Proceedings of the 11th Annual Conference on Genetic and Evolutionary Computation

View full text Add to dashboard Cite

The intrinsic feature of Estimation of Distribution Algorithms lies in their ability to learn and employ probabilistic models over the input spaces. Discovery of the appropriate model usually implies a computationally expensive comprehensive search, where many models are proposed and evaluated in order to find the best value of some model discriminative scoring metric.This paper presents basic results demonstrating how simple variable correlation data can be extended and used to efficiently guide the model search, decreasing the number of model evaluations by several orders of magnitude and without significantly affecting model quality.As a case study, the O(n 3 ) model building of the Extended Compact Genetic Algorithm is successfully replaced by a correlation guided search of linear complexity which infers the perfect problem structures on the test suites.

show abstract

Section: Hybridization Of Ecgamentioning

confidence: 98%

“…The MPM concept is illustrated in Table 3 over a 3 bit problem with [1,3], [2] as partitions. The first and third bit are jointly distributed while variable [2] is independent.…”

Section: Case Study On Ecgamentioning

confidence: 99%