Genetic Algorithms, Tournament Selection, and the Effects of Noise

By Brad L. Miller, David E. Goldberg. 1995

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Optimization & Training

Citations in the corpus

A Multimodal Framework for the Detection of Hateful Memes (2020, methodology): “We experimented with different number of cross-validation folds, but our final model is trained under a 15-fold cross-validation setting and we optimize the ensemble weights using an evolutionary algorithm with tournament selection, Gaussian noise mutation and uniform crossover [8, 24] strategies.”
Weight Agnostic Neural Networks (2019, background): “…multiple rollouts, with a different shared weight value assigned at each rollout, (3) networks are ranked according to their performance and complexity, and (4) a new population is created by varying the highest ranked network topologies, chosen probabilistically through tournament selection [74].”
An Adaptive Random Path Selection Approach for Incremental Learning. (2019, methodology): “Comparison with a Genetic Algorithm: We compare our random selection approach with a genetic algorithm i.e., Binary Tournament Selection (BTS) [36] for 25 maximum generations, on MNIST with 5 tasks (each of 2 classes), using a simple 2 layer (100 neurons) MLP with M = 8 , J = 1 .”
Global Optimization Algorithms -- Theory and Application (2009, background): “Its features are well-known and have been analyzed by a variety of researchers such as Blickle and Thiele [485; 478], Miller and Goldberg [487], Lee et al.”
A Genetic Algorithm to Optimize Lazy Learning Parameters for the Prediction of Customer Demands (2013, methodology): “In the following, we describe the results of the parameter optimization by the GA. Subsequently, we show the prediction results where we compare the nonlinear dynamics prediction methods to linear standard methods regarding forecasting accuracy.”
Development of a fast fluid dynamics-based adjoint method for the inverse design of indoor environments (2017, methodology): “According to Liu et al. (2015b), the computational fluid dynamics (CFD)-based genetic algorithm (Miller and Goldberg 1995; Xue, Zhai, and Chen 2013), CFD-based adjoint method (Liu and Chen 2015), and proper orthogonal decomposition (POD) analysis (Li et al. 2013) can be used for such a purpose.”
Evolving Agents for the Hanabi 2018 CIG Competition (2018, methodology): “Individuals are chosen for crossover using tournament selection [17] with tournament size t.”
Automatically Designing CNN Architectures Using the Genetic Algorithm for Image Classification (2018, background): “To this end, the binary tournament selection is typically used for such a purpose [51], [56].”
MOEA/D with Adaptive Weight Adjustment (2014, background): “For each i = 1 , · · · , N , set the neighborhood list of the i -th subproblem as (Miller and Goldberg, 1995) according to π i , and add them to I .”
IGD Indicator-Based Evolutionary Algorithm for Many-Objective Optimization Problems (2018, methodology): “The binary tournament selection [57] approach is employed in Algorithm 5 to select individuals from the current population to ﬁll up the gene pool.”
Automatically Designing CNN Architectures Using Genetic Algorithm for Image Classification (2018, background): “To this end, the binary tournament selection is typically used for such a purpose [50], [55].”
Completely Automated CNN Architecture Design Based on Blocks (2020, methodology): “In the proposed AE-CNN algorithm, the binary tournament selection [50] is used for this purpose [50], [51] based on the conventions of the GA community.”