Evolutionary algorithms and their applications to engineering problems

Jellyfish Search Optimization Algorithm for MPP Tracking of PV System

Because of the rapid increase in the depletion rate of conventional energy sources, the energy crisis has become a central problem in the contemporary world. This issue opens the gateway for exploring and developing renewable energy sources to fulfill the exigent energy demand. Solar energy is an abundant source of sustainable energy and hence, nowadays, solar photovoltaic (PV) systems are employed to extract energy from solar irradiation. However, the PV systems need to work at the maximum power point (MPP) to exploit the highest accessible power during varying operating conditions. For this reason, maximum power point tracking (MPPT) algorithms are used to track the optimum power point. Furthermore, the efficient utilization of PV systems is hindered by renowned partial shading conditions (PSC), which generate multiple peaks in the power-voltage characteristic of the PV array. Thus, this article addresses the performance of the newly developed jellyfish search optimization (JSO) strategy in the PV frameworks to follow the global maximum power point (GMPP) under PSC.

Time-Lag Selection for Time-Series Forecasting Using Neural Network and Heuristic Algorithm

The time-series forecasting is a vital area that motivates continuous investigate areas of intrigued for different applications. A critical step for the time-series forecasting is the right determination of the number of past observations (lags). This paper investigates the forecasting accuracy based on the selection of an appropriate time-lag value by applying a comparative study between three methods. These methods include a statistical approach using auto correlation function, a well-known machine learning technique namely Long Short-Term Memory (LSTM) along with a heuristic algorithm to optimize the choosing of time-lag value, and a parallel implementation of LSTM that dynamically choose the best prediction based on the optimal time-lag value. The methods were applied to an experimental data set, which consists of five meteorological parameters and aerosol particle number concentration. The performance metrics were: Mean Absolute Error (MAE), Root Mean Square Error (RMSE), Mean Absolute Percentage Error (MAPE) and R-squared. The investigation demonstrated that the proposed LSTM model with heuristic algorithm is the superior method in identifying the best time-lag value.

Application of Machine Learning to Solid Particle Erosion of APS-TBC and EB-PVD TBC at Elevated Temperatures

Machine learning (ML) and deep learning (DL) for big data (BD) management are currently viable approaches that can significantly help in high-temperature materials design and development. ML-DL can accumulate knowledge by learning from existing data generated through multi-physics modelling (MPM) and experimental tests (ETs). DL mainly involves analyzing nonlinear correlations and high-dimensional datasets implemented through specifically designed numerical algorithms. DL also makes it possible to learn from new data and modify predictive models over time, identifying anomalies, signatures, and trends in machine performance, develop an understanding of patterns of behaviour, and estimate efficiencies in a machine. Machine learning was implemented to investigate the solid particle erosion of both APS (air plasma spray) and EB-PVD (electron beam physical vapour deposition) TBCs of hot section components. Several ML models and algorithms were used such as neural networks (NNs), gradient boosting regression (GBR), decision tree regression (DTR), and random forest regression (RFR). It was found that the test data are strongly associated with five key factors as identifiers. Following test data collection, the dataset is subjected to sorting, filtering, extracting, and exploratory analysis. The training and testing, and prediction results are analysed. The results suggest that neural networks using the BR model and GBR have better prediction capability.

Parametric Design and Comfort Optimization of Dynamic Shading Structures

This work implements parametric tools to optimize the environmental design of urban adaptive shadings through multiobjective evolutionary algorithms that look for solutions of dynamic (time-changing) structures used in open public spaces. The proposal is located in Malecon Cancun Tajamar in the southeast part of Mexico, and the main objective is to enhance the thermal comfort of users as well as to become part of the social dynamics of the place reinforcing identity through appropriation. The proposed workflow includes four steps: (1) geometric modelling by parametric modelling tools; (2) simulation of environmental parameters by using BPS tools; (3) shape optimization by using an evolutionary algorithm; and (4) environmental verification of the results. The Universal Thermal Climate Index (UTCI) was used to assess the outdoor thermal comfort derived from the dynamic shadings. The results showed a significant improvement in the thermal comfort with absolute UTCI differences of 3.9, 7.4, and 3.1 °C at 8, 12, and 16 h, respectively, during the summer; and absolute differences of 1.4, 3.5, and 2 °C at 8, 12, and 16 h, respectively, during the winter. The proposed workflow can help to guide the early design process of dynamic shadings by finding optimal solutions that enhance outdoor thermal comfort.

A Computational Approach to Solve a System of Transcendental Equations with Multi-Functions and Multi-Variables

A system of transcendental equations (SoTE) is a set of simultaneous equations containing at least a transcendental function. Solutions involving transcendental equations are often problematic, particularly in the form of a system of equations. This challenge has limited the number of equations, with inter-related multi-functions and multi-variables, often included in the mathematical modelling of physical systems during problem formulation. Here, we presented detailed steps for using a code-based modelling approach for solving SoTEs that may be encountered in science and engineering problems. A SoTE comprising six functions, including Sine-Gordon wave functions, was used to illustrate the steps. Parametric studies were performed to visualize how a change in the variables affected the superposition of the waves as the independent variable varies from x1 = 1:0.0005:100 to x1 = 1:5:100. The application of the proposed approach in modelling and simulation of photovoltaic and thermophotovoltaic systems were also highlighted. Overall, solutions to SoTEs present new opportunities for including more functions and variables in numerical models of systems, which will ultimately lead to a more robust representation of physical systems.

Non-Deterministic Methods and Surrogates in the Design of Rockfill Dams

The parameters of the constitutive models used in the design of rockfill dams are associated with a high degree of uncertainty. This occurs because rockfill dams are comprised of numerous zones, each with different soil materials, and it is not feasible to extract materials from such structures to accurately ascertain their behavior or their respective parameters. The general approach involves laboratory tests using small material samples or empirical data from the literature. However, such measures lack an accurate representation of the actual scenario, resulting in uncertainties. This limits the suitability of the model in the design process. Inverse analysis provides an option to better understand dam behavior. This procedure involves the use of real monitored data, such as deformations and stresses, from the dam structure via installed instruments. Fundamentally, it is a non-destructive approach that considers optimization methods and actual performance data to determine the values of the parameters by minimizing the differences between simulated and observed results. This paper considers data from an actual rockfill dam and proposes a surrogate assisted non-deterministic framework for its inverse analysis. A suitable error/objective function that measures the differences between the actual and simulated displacement values is defined first. Non-deterministic algorithms are used as the optimization technique, as they can avoid local optima and are more robust when compared to the conventional deterministic methods. Three such approaches, the genetic algorithm, differential evolution, and particle swarm optimization are evaluated to identify the best strategy in solving problems of this nature. A surrogate model in the form of a polynomial regression is studied and recommended in place of the actual numerical model of the dam to reduce computation cost. Finally, this paper presents the relevant dam parameters estimated by the analysis and provides insights into the performance of the three procedures to solve the inverse problem.

Inference of dynamic spatial GRN models with multi-GPU evolutionary computation

Reverse engineering mechanistic gene regulatory network (GRN) models with a specific dynamic spatial behavior is an inverse problem without analytical solutions in general. Instead, heuristic machine learning algorithms have been proposed to infer the structure and parameters of a system of equations able to recapitulate a given gene expression pattern. However, these algorithms are computationally intensive as they need to simulate millions of candidate models, which limits their applicability and requires high computational resources. Graphics processing unit (GPU) computing is an affordable alternative for accelerating large-scale scientific computation, yet no method is currently available to exploit GPU technology for the reverse engineering of mechanistic GRNs from spatial phenotypes. Here we present an efficient methodology to parallelize evolutionary algorithms using GPU computing for the inference of mechanistic GRNs that can develop a given gene expression pattern in a multicellular tissue area or cell culture. The proposed approach is based on multi-CPU threads running the lightweight crossover, mutation and selection operators and launching GPU kernels asynchronously. Kernels can run in parallel in a single or multiple GPUs and each kernel simulates and scores the error of a model using the thread parallelism of the GPU. We tested this methodology for the inference of spatiotemporal mechanistic gene regulatory networks (GRNs)-including topology and parameters-that can develop a given 2D gene expression pattern. The results show a 700-fold speedup with respect to a single CPU implementation. This approach can streamline the extraction of knowledge from biological and medical datasets and accelerate the automatic design of GRNs for synthetic biology applications.

Optimal Operation of the Voltage-Doubler Boost Converter through an Evolutionary Algorithm

This manuscript presents the numerical optimization (through a mathematical model and an evolutionary algorithm) of the voltage-doubler boost converter, also called the series-capacitor boost converter. The circuit is driven by two transistors, each of them activated according to a switching signal. In the former operation, switching signals have an algebraic dependence from each other. This article proposes a new method to operate the converter. The proposed process reduces the input current ripple without changing any converter model parameter, only the driving signals. In the proposed operation, switching signals of transistors are independent of each other, providing an extra degree of freedom, but on the other hand, this produces an infinite number of possible combinations of duty cycles (the main parameter of switching signals) to achieve the desired voltage gain. In other words, this leads to a problem with infinite possible solutions. The proposed method utilizes an evolutionary algorithm to determine the switching functions and, at the same time, to minimize the input current ripple of the converter. A comparison made between the former and the proposed operation shows that the proposed process achieves a lower input current ripple while achieving the desired voltage gain.

A Genetic Algorithm Procedure for the Automatic Updating of FEM Based on Ambient Vibration Tests

The dynamic identification of the modal parameters of a structure, in order to gain control of its functionality under operating conditions, is currently under discussion from a scientific and technical point of views. The experimental observations obtained through structural health monitoring (SHM) are a useful calibration reference of numerical models (NMs). In this paper, the procedures for the identification of modal parameters in historical bell towers using a stochastic subspace identification (SSI) algorithm are presented. Then, NMs are manually calibrated on the identification's results. Finally, the applicability of a genetic algorithm for the automatic calibration of the elastic parameters is considered with the aim of searching for the properties of the autochthonous material, in order to reduce modelling error following the model assurance criterion (MAC). In this regard, several material values on the same model are examined to see how to approach the evolution and the distribution of these features, comparing the characterization proposed by the genetic algorithm with the results considered by the manual iterative procedure.