A Fuzzy Multi-Objective Evolutionary Framework with Bargaining-Based Selection for Interpretable Body Fat Prediction

Transfer learning-assisted multi-objective evolutionary clustering framework with decomposition for high-dimensional data

Design of multi-component structures can be a challenging task. While having multiple components in a complex structure is often necessary in order to reduce the manufacturing cost, multiple components need joining operations. Optimal design of joints is not a decoupled problem from designing the base structure, and often comes at balancing trade-offs in assembly cost, weight and structural performance. Thus, the problem is posed in a multi-objective framework. Since some of the objectives are inherently discrete, non-gradient optimization methods are needed. Previous work has adopted a Kriging-interpolated levelset (KLS) formulation for implicit definition of the base topology as well as its decomposition into multiple components. While the number of design variables in KLS formulation is significantly smaller than explicit formulations, it can still be a challenge for general-purpose non-gradient multi-objective algorithms. This paper proposes a systematic approach for the problem in order to efficiently generate a well-seeded initial population to be used in multi-objective evolutionary algorithms. A multi-component cantilever is used as a basis for comparison between a basic NSGA-II algorithm, versus the proposed optimization framework. The results demonstrate its superiority and capability in obtaining multi-component complex topologies with desirable quality, which are not achieved by the basic algorithm.

With the advent of microarray technology it has been possible to measure thousands of expression values of genes in a single experiment. Biclustering or simultaneous clustering of both genes and conditions is challenging particularly for the analysis of high-dimensional gene expression data in information retrieval, knowledge discovery, and data mining. The objective here is to find sub-matrices, i.e., maximal subgroups of genes and subgroups of conditions where the genes exhibit highly correlated activities over a range of conditions while maximizing the volume simultaneously. Since these two objectives are mutually conflicting, they become suitable candidates for multi-objective modeling. In this study, we will describe some recent literature on biclustering as well as a multi-objective evolutionary biclustering framework for gene expression data along with the experimental results.

Multi-objective evolutionary co-learning framework for energy-efficient hybrid flow-shop scheduling problem with human-machine collaboration

The asset-based weapon target assignment (ABWTA) problem is one of the important branches of the weapon target assignment (WTA) problem. Due to the current large-scale battlefield environment, the ABWTA problem is a multi-objective optimization problem (MOP) with strong constraints, large-scale and sparse properties. The novel model of the ABWTA problem with the operation error parameter is established. An evolutionary algorithm for large-scale sparse problems (SparseEA) is introduced as the main framework for solving large-scale sparse ABWTA problem. The proposed framework (SparseEA-ABWTA) mainly addresses the issue that problem-specific initialization method and genetic operators with a reward strategy can generate solutions efficiently considering the sparsity of variables and an improved non-dominated solution selection method is presented to handle the constraints. Under the premise of constructing large-scale cases by the specific case generator, two numerical experiments on four outstanding multi-objective evolutionary algorithms (MOEAs) show Runtime of SparseEA-ABWTA is faster nearly 50% than others under the same convergence and the gap between MOEAs improved by the mechanism of SparseEA-ABWTA and SparseEA-ABWTA is reduced to nearly 20% in the convergence and distribution.

Broad learning approach to Surrogate-Assisted Multi-Objective evolutionary fuzzy clustering algorithm based on reference points for color image segmentation

1. Body-weight, height, skinfold thickness and body density measurements were made on 245 men and 324 women aged between 17 and 72 years. The body fat content of each individual was calculated from his density measurement using equations similar to that of Siri (1956) but appropriate to age, muscular development, and extent of obesity. Regression equations were then derived for separate age-groups for the prediction of body fat from (1) weight–height (W–H) relationships (W:H, W:H2 W:H3, W0.33:H, H:W0.33), (2) ‘percentage overweight’ (body-weight:mean body-weight; W:W;) and ‘percentage desirable weight’ body-weight: ‘desirable weight’ (mean weights for men and women aged 20-24 years); (W:DES), (3) the independent variables weight and height incorporated in a regression equation, (4) skinfold measurements.2. The correlations between height and indices, W:H, W:H3, W0.33:H and H:W0.33: were substantially different from those between height and body fat estimated by densitometry.3. The method having the highest correlation with body fat estimated by densitometry was the skinfold method, although in the older groups of women other methods sometimes gave equally good correlations. The index H:W0.33 had a negative correlation with body fat, and the indices W:H, W0.33:H and W:H3 had a lower correlation with body fat estimated by densitometry than did the other indices. These indices are therefore unsuitable measures of obesity, and this is in agreement with the findings of other workers.4. The correlations betweeen body fat estimated by densitometry and the indices W:H2, W:W, W:DES and the equations incorporating the independent variable weight and height are all very similar, although it is not appropriate to use the index W:W; as a measure of obesity in groups of people of widely different ages.5. The standard error of prediction of body fat from skinfold measurement may be of the same order of magnitude as the standard error of prediction of body fat by densitometry. It is therefore probably inappropriate to assess the accuracy of the skinfold method by comparison with the density method alone.6. From the distribution patterns obtained, it was evident that compared with the density method, all methods tended to over-estimate body fat in very lean individuals. The skinfold method however showed this tendency to only a relatively small extent.

Background: Although the body mass index (BMI, kg m−2) is widely used as a measure of adiposity, it is a measure of excess weight, rather than excess body fat. It has been suggested that skinfold thicknesses be measured among overweight children to confirm the presence of excess adiposity.Objective: The present study examined the additional information provided by skinfold thicknesses on body fatness, beyond that conveyed by BMI-for-age, among healthy 5- to 18-years old (n = 1196).Methods and procedures: Total body dual-energy X-ray absorptiometry (DXA) provided estimates of % body fat, and the sum of two skinfolds (triceps and subscapular) was used as an indicator of the overall skinfold thickness.Results: As assessed by the multiple R2s and the residuals of various regression models, information on the skinfold sum significantly ( p < 0.001) improved the prediction of body fatness beyond that obtained with BMI-for-age. For example, the use of the skinfold sum, in addition to BMI-for-age, increased the multiple R2s for predicting % body fat from 0.81 to 0.90 (boys), and from 0.82 to 0.89 (girls). The use of the skinfold sum also reduced the overall prediction errors (absolute value of the residuals) for % body fat by 20–30%, but these reductions varied substantially by BMI-for-age. Among overweight children, defined by a BMI-for-age ≥95th percentile, the skinfold sum reduced the predication errors for % body fat by only 7–9%.Conclusions: Although skinfold thicknesses, when used in addition to BMI-for-age, can substantially improve the estimation of body fatness, the improvement among overweight children is small.

Biological networks such as protein-protein interaction networks (PPINs) play a crucial role in cell biology of an organism and are widely studied in bioinformatics and biostatistics. Scores of high throughput techniques are used to identify PPIs which are then used to compile PPINs of an organism. Detection of protein complexes that are associated with diseases is another ongoing challenge. Consequently, methods have been developed that use PPINs in order to predict or identify disease associated protein complexes. Here we have modeled this task as a multi-objective optimization problem and developed a multi-objective evolutionary framework called MODAPROC to detect protein complexes associated with human diseases. MODAPROC considered three objective functions viz., density, number of outward interactions, and disease associations of a protein complex. Density and disease associations are to be maximized while outward interactions are minimized here. We have analyzed the resulting protein complexes and their association with different disorders. Moreover, we have detected some key properties of the disease associated protein complexes as well as classified them according to different disorders. MODAPROC is fast, robust, and provides a good analysis tool for finding and discovering obscure disease specific gene clusters and can be extended using other domain specific objective functions.

In the electric vehicle (EV) application, increasing power demand requires a reliable power supply to support continuous charging operation. In the previous optimisation framework, grid input source power quality, total harmonic distortions, the need to control energy management, and high EV load demand were limitation in existing work. To solve the limitation, a novel Multi Objective Evolutionary Algorithm for prediction based proper grid energy management with two modes of operation is proposed. An input source is a Grid and an alternative renewable PV source combined for an EV charging station, providing power to the EV and ensuring proper flow in both directions. Grid-to-Vehicle (G2V) charging provides power to charge an electric vehicle’s batteries, and Vehicle-to-Grid (V2G) charging returns power to the grid for discharge. The battery power is periodically evaluated at a PWM duty cycle over a specified time period (in seconds) during steady-state operation in two different modes. The Multi-Objective Evolutionary Algorithm analyses the input Grid and alternative solar renewable sources with proper charge and discharge functions. The EV load output, G2V and V2G output, SOC, and Battery energy management are calculated with an appropriate description of the dataset. The result improved energy management, effectively manages load demand, and improves battery backup for solar renewable sources and reliable EV output.

Defect recognition is an essential aspect of intelligent manufacturing, but it is a challenging task with noise and unpredictable uncertainties, where convolutional neural networks (CNNs) struggle to achieve good performance. The fuzzy neural network (FNN) emerges as a promising approach to handle uncertainties. However, conventional methods for designing FNNs are tedious and error-prone. A solution is to automatically search for efficient FNN, which can be achieved by neural architecture search (NAS). To achieve NAS for FNN, we propose an efficient classifier-assisted evolutionary multiobjective FNN framework for defect recognition. Considering the characteristics of FNN (e.g., difficult to train and prone to overfitting), we first construct the architecture search as a constrained multiobjective optimization problem, the network accuracy and the architecture size are two conflicting objectives, and the constraint is used to filter out low-quality architectures. Then, we design the search space to incorporate the fuzzy module and develop the corresponding architectural representation and evolutionary operators. Furthermore, the complex regression task of performance evaluation is transformed into a classification task, and a classifier is designed to simplify the performance evaluation process. Massive experiments on four defect recognition datasets (i.e., ELPV, CODEBRIM, MIXEDWM38, and WM-811K) show that the architectures can effectively handle inherent uncertainties from datasets. Our method achieves 94.77% accuracy on ELPV, 81.82% accuracy on CODEBRIM, 98.99% accuracy on MIXEDWM38, and 98.22% accuracy on WM-811K, respectively.

A conjunctive management framework for the optimal design of pumping and injection strategies to mitigate seawater intrusion

One of the critical public safety roles for water distribution systems (WDS) is suppression of urban fire events. Previous studies have investigated WDS rehabilitation for mitigation of potential fire events with a major focus on improving fire flows by pipe enlargement. However, pipe enlargement can cause water quality problems and place public health at risk during normal operational periods. Thus a novel approach is required to effectively address the conflicting goals of the WDS: reliable delivery of water during normal as well as emergency conditions, meeting water quality standards, and finding cost-effective design and rehabilitation options. In this study an evolutionary computation-based multi-objective optimization-simulation framework is developed to design effective mitigation strategies for urban fire events for water distribution systems with three objectives: (1) minimizing fire damages, (2) minimizing water quality deficiencies, and (3) minimizing the cost of mitigation. An elitist non-dominated sorting genetic algorithm (NSGA-II) is modified by incorporating an evolution strategy (ES) to address difficulties for heuristic algorithms posed by WDS problems. Implementation of this methodology generates Pareto-optimal solution surfaces that express the trade-off relationship between fire damage, water quality, and least cost objectives. Thus, the method provides decision makers with the flexibility to choose a mitigation plan for urban fire events best suited for their circumstances. Each Paretooptimal solution comprises a set of pipes to be enlarged to achieve increased fire flow and the corresponding diameters of these pipes. The algorithm is illustrated with several test functions. The Micropolis virtual city is then used to demonstrate the application of the proposed methodology to a complex WDS.

In this paper, a new real-world application problem, i.e., the vehicle fleet maintenance scheduling optimization problem, is defined and a specialized multi-objective evolutionary algorithm framework (grouping strategy, three vector chromosome and corresponding genetic operators) is proposed to solve the problem. State-of-the-art multi-objective evolutionary algorithms such as NSGA-III, SMS-EMOA, DI-MOEA are employed in the proposed algorithm framework to solve the problem, and their behavior is investigated. Although DI-MOEA is used the first time for a real-world application problem, its performance is better than other algorithms for some instances.

In the field of preference-based evolutionary multiobjective optimization, optimization algorithms are required to search for the Pareto optimal solutions preferred by the decision maker (DM). The reference point is a type of techniques that effectively describe the preferences of DM. So far, the reference point is either static or interactive with the evolutionary process. However, the existing reference point techniques do not cover all application scenarios. A novel case, i.e., the reference point changes over time due to the environment change, has not been considered. This paper focuses on the multiobjective optimization problems with dynamic preferences of the DM. First, we propose a change model of the reference point to simulate the change of the preference by the DM over time. Then, a dynamic preference-based multiobjective evolutionary algorithm framework with a clonal selection algorithm (ĝa-NSCSA) and a genetic algorithm (ĝa-NSGA-II) is designed to solve such kind of optimization problems. In addition, in terms of practical applications, the experiments on the portfolio optimization problems with the dynamic reference point model are tested. Experimental results on the benchmark problems and the practical applications show that ĝa-NSCSA exhibits better performance among the compared optimization algorithms.