China 
Africa 
Explore chapters and articles related to this topic
Performance Assessment and Metric Indices
Published in Yi Chen, Yun Li, Computational Intelligence Assisted Design, 2018
Moreover, the fitness function should not only correlate closely to the designer's goal but also be computationally very efficient for the fitness approximation, especially in those cases that require large amounts of CPU time, including where (1) the computation time of a single solution is extremely high, (2) the precise model for the fitness computation is missing or is too complex to define, or (3) the fitness function is uncertain or noisy.
Using a Random Forest Model to Choose Optimized Group Structures
Published in Nuclear Science and Engineering, 2023
Thomas G. Saller, Vishnu Nair, Andrew Till, Nathan Gibson
In Ref. 10, Yi et al. switch from using a simulation-driven fitness function (i.e., performing a full transport calculation to get k-eff) to a physics-based fitness approximation, which was derived from contribution transport theory. The advantage of the physics-based fitness approximation is that it does not require any coarse-group transport calculations, significantly decreasing the computational burden. All it requires is the forward and adjoint fine-group fluxes. Yi et al. found similar results between the previous simulation-driven fitness function and the new physics-based fitness approximation. This outcome indicates that in the future we should investigate ways to use the fine-group fluxes to estimate coarse-group structure effectiveness.
BCDU-Net and chronological-AVO based ensemble learning for lung nodule segmentation and classification
Published in Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2023
The most widely utilised neural network (NN) is the feed-forward back-propagation MLP for pattern classification problems. MLP generally utilises a standard back-propagation algorithm for all supervised pattern recognition steps. MLP has good solving capability even for complex optimisation problems. Moreover, it is highly useful in solving constraints like fitness approximation. Here, MLP is employed with three layers, such as input, hidden and output layer in classifying lung nodules. Figure 3 implies the framework of MLP classifier.