Genetic Algorithm Algorithm Research Articles

Mathematical Physics Approaches to Nanotechnology and Material Science

This work examines the use of mathematical physics tools and artificial intelligence in enhancing nanotechnology and materials science studies. Using AI algorithms interacting with the basic physical models the paper investigates how these combined methods may enhance the design and characteristics of nanomaterials. Support Vector Machine (SVM), Artificial Neural Networks (ANN), Genetic Algorithm (GA), and Particle Swarm Optimization (PSO) were employed, whereby predictions of material properties were made, synthesis processes optimized and characterization methods improved. As a result, the authors reported that AI-based models ‘ raised the bars’ in the prediction of material properties with respective accuracies of 92% SVM and 89% ANN. Based on optimization, result showed that the GA algorithm provided 35% improvement in material while the PSO provided increased energy efficiency of 40% on nanomaterial application. In experimental research outcomes, the efficiency of AI technique also proved that AI methodologies were 25% faster process than synthesis process and cost 20% less compared to standard methodologies. These outcomes show that AI and mathematical physics are versatile tools to transform material synthesis and advance the search for environment-friendly, high-performance materials at a much faster pace. This study reveals that other fields such as AI and physics can collaborate to enhance developments in the nanotechnology and materials science fields.

Conceptual Framework for Adaptive Bacterial Memetic Algorithm Parameterization in Storage Location Assignment Problem

Recognized as an NP-hard combinatorial challenge, Storage Location Assignment Problem (SLAP) demands heuristic or algorithmic solutions for effective optimization. This paper specifically examines the enhancement of SLAP through the utilization of evolutionary algorithms, as they are particularly suitable for complex cases. Among others, the genetic algorithm (GA) is typically applied to solve this problem. This paper investigates the Bacterial Memetic Algorithm (BMA) as a possible solution for optimization. Though the comparative analysis of the BMA with the previously well-used GA algorithm under certain test parameters reveals that BMA is suitable for SLA optimization, BMA failed to achieve better results. We attribute the unsatisfactory results to the parameter settings, as illustrated by a few specific examples. However, the complexity of the problem and the parameterization does not allow for continuous manual parameter adjustment, which is why we have identified the need for a concept that automatically and adaptively adjusts the parameter settings based on the statistics and fitness values obtained during the execution. The novelty of this paper is to specify the concept of adaptive BMA parameterization and rules.

Research on time-varying path optimization for multi-vehicle type fresh food logistics distribution considering energy consumption

With the increasing demand for fresh food markets, refrigerated transportation has become an essential component of logistics operations. Currently, fresh food transportation frequently faces issues of high energy consumption and high costs, which are inconsistent with the development needs of the modern logistics industry. This paper addresses the optimization problem of multi-vehicle type fresh food distribution under time-varying conditions. It comprehensively considers the changes in road congestion at different times and the quality degradation characteristics of fresh goods during distribution. The objectives include transportation cost, dual carbon cost, and damage cost, subject to constraints such as delivery time windows and vehicle capacity. A piecewise function is used to depict vehicle speeds, proposing a dynamic urban fresh food logistics vehicle routing optimization method. Given the NP-hard nature of the problem, a hybrid Tabu Search (TS) and Genetic Algorithm (GA) approach is designed to compute an optimal solution. Comparison with TS and GA algorithm results shows that the TS-GA algorithm provides the best optimization efficiency and effectiveness for solving large-scale distribution problems. The results indicate that using the TS-GA algorithm to optimize a distribution network with one distribution center and 30 delivery points resulted in a total cost of CNY 12,934.02 and a convergence time of 16.3 s. For problems involving multiple vehicle types and multiple delivery points, the TS-GA algorithm reduces the overall cost by 2.94–7.68% compared to traditional genetic algorithms, demonstrating superior performance in addressing multi-vehicle, multi-point delivery challenges.

Voice recognition enhancement by genetic algorithm

This research paper investigates the application of Genetic Algorithms (GA) in optimizing Artificial Neural Networks (ANN) for phoneme recognition. The study examines the formalism of GA, their parameters, and operators, and describes the genetic strategy adopted for phoneme recognition using the TIMIT sound database. The paper presents the outcomes of experiments conducted on the phonemes of the test base and the DR1 dialect learning base of the TIMIT base, and compares the recognition rates obtained by learning and testing with those guided by Self-Organizing Map (SOM) experiments. The findings suggest that the GA algorithm provides improved recognition rates and extends the search space to a global optimum, but can sometimes produce unacceptable results due to rapid and premature convergence and the overfitting problem. The study proposes a hybrid model between SOM and GA to take advantage of each of their properties and improve recognition rates. The results show that the proposed approach achieves an average recognition rate of over 95 %.

Effect of DG penetration on hybrid relay coordination using user-defined dual-setting directional overcurrent relays and distance relays

ABSTRACT The efficient coordination of the distance and directional overcurrent relays (DOCRs) is jeopardized by the massive deployment of renewable distributed generation (DG) into electrical power networks. The two key aspects influencing the relay coordination are the bidirectional flow of current and increased short-circuit currents due to DG integration. This study investigates the consequences of variable DG penetration on the distance and dual-setting DOCRs (DS-DOCRs) coordination. As a result, common optimal relay settings have been obtained that can be utilized for variable DG penetration levels in modified IEEE-14 bus test system. For the distance relays (DRs), a fixed zone 1 and variable zone 2 operation have been considered. The proposed combined protection scheme is tested on a modified IEEE-14 bus test system using DS-DOCRs and DRs. The relay coordination problem is formulated as a constrained non-linear optimization problem and solved by genetic algorithm (GA) and gray wolf optimization (GWO). In this paper, DS-DOCRs with user-defined relay characteristics have demonstrated superior performance in line protection compared to conventional DOCRs with normal inverse (NI) characteristics. In addition, a comprehensive comparative analysis of optimization algorithms has been performed between GA and GWO algorithms.

Evolutionary Algorithm for the Earliness - Tardiness Project Scheduling Problem

Completing too early or too late can cause damage to the project. For example, in transporting goods to the seaports, if the shipping arrives too early, it will take time and cost to stay at the dock. In the other case, shipping too late will reduce the quality of the product and break the contract. The goal of this paper is to minimize the tardiness and earliness of those projects. First, some related studies are briefly introduced, thereby showing the role and practical applications of the MS-RCPSP (Multi-Skill Resource Constrained Project Scheduling Problem) problem and the two factors of earliness and delay in scheduling problems. The roles of earliness and tardiness in scheduling problems are also described. At the same time, the new contribution of this research is also highlighted, which is the proposal for the latest problem E-RCPSP (Earliness tardiness MS-RCPSP). This problem is the first problem in the MS-RCPSP family that considers the sum of earliness and tardiness as the objective, which has never been considered in previous studies. Then, the mathematical model of the proposed E-RCPSP problem, which includes the objective function, project components and constraints is introduced. To solve the proposed problem, two approximate algorithms based on the evolutionary algorithms GA (Genetic Algorithm) and DE (Differential Evolution) are introduced. In the last section of the article, the performance of these two algorithms is verified and compared with each other through experiments conducted on the iMOPSE benchmark dataset.}\Keywords{Network Resources Management, Earliness-Tardiness cost, Resource-constrained project scheduling, Evolutionary Algorithms, Differential evolution algorithms.

Comparative study of optimization methods for building energy consumption and daylighting performance

When striving for sustainable development, designers encounter the formidable task of integrating the advantages of daylighting into energy-efficient designs. The process of investigating a multitude of design variables can be arduous and time-intensive. However, computational simulations provide indispensable insights into forecasting performance. A considerable number of these design challenges do not have explicit mathematical solutions; thus, black-box optimization techniques are favored for the purpose of determining the most effective design solutions. Popular meta-heuristic optimization algorithms, including the Genetic Algorithm (GA), Particle Swarm Optimization (PSO), Whale Optimization Algorithm (WOA), Sine Cosine Algorithm (SCA), Grey Wolf Optimizer (GWO), Ant Lion Optimizer (ALO), and Multi-Verse Optimizer (MVO), are described and thoroughly analyzed in this article. In addition, the present study run empirical studies to compare these optimization algorithms by applying them to tackle five distinct daylighting and energy optimization problems associated with enhancing building performance. In the test cases, the present study results demonstrate that the PSO and MVO algorithms exhibit better performance compared to the GA and WOA algorithms. The selection of the most suitable algorithm is contingent upon the distinctive attributes of the problem, and the intricacy of the optimization landscape.

Optimizing column location in flat slab system using Particle Swarm Optimization vs Genetic Algorithms

Abstract One of the key architectural advantages of a concrete flat slab system is its flexibility in design. The absence of beams and columns allows for larger, open spaces with minimal obstructions, giving architects the freedom to create unique and innovative layouts. However, the location of the middle columns (one of the main factors affecting the construction cost) is commonly aligned based on experience which may entail an uneconomic design alternative. This paper aims to propose different machine learning optimization algorithms, namely, particle swarm optimization (PSO) and genetic algorithms (GA) to optimize the location of columns within a flat slab plan considering different serviceability requirements. Both optimization systems showed good performance for allocating the concrete columns, However, the GA algorithm showed better results in terms of maximum deflection with less computational effort.

Machine Learning Approaches for Short-Term Photovoltaic Power Forecasting

A photovoltaic (PV) power forecasting prediction is a crucial stage to utilize the stability, quality, and management of a hybrid power grid due to its dependency on weather conditions. In this paper, a short-term PV forecasting prediction model based on actual operational data collected from the PV experimental prototype installed at the engineering college of Misan University in Iraq is designed using various machine learning techniques. The collected data are initially classified into three diverse groups of atmosphere conditions—sunny, cloudy, and rainy meteorological cases—for various seasons. The data are taken for 3 min intervals to monitor the swift variations in PV power generation caused by atmospheric changes such as cloud movement or sudden changes in sunlight intensity. Then, an artificial neural network (ANN) technique is used based on the gray wolf optimization (GWO) and genetic algorithm (GA) as learning methods to enhance the prediction of PV energy by optimizing the number of hidden layers and neurons of the ANN model. The Python approach is used to design the forecasting prediction models based on four fitness functions: R2, MAE, RMSE, and MSE. The results suggest that the ANN model based on the GA algorithm accommodates the most accurate PV generation pattern in three different climatic condition tests, outperforming the conventional ANN and GWO-ANN forecasting models, as evidenced by the highest Pearson correlation coefficient values of 0.9574, 0.9347, and 0.8965 under sunny, cloudy, and rainy conditions, respectively.

Study on Multi-Objective Optimization of Construction Project Based on Improved Genetic Algorithm and Particle Swarm Optimization

Construction projects require concurrent consideration of the three major objectives of construction period, cost, and quality. To address the multi-objective optimization issues of construction projects, mathematical models of construction period, quality, and cost are established, respectively, and multi-objective optimization models are constructed for different construction objectives. A hybrid optimization method combining an improved genetic algorithm (GA) with a time-varying mutation rate and a particle swarm algorithm (PSO) is proposed to optimize construction projects, which overcomes the shortcomings of the original GA and improves the global optimality and stability of results. Various construction projects were considered, and different construction objectives were analyzed individually. Finally, an uncertainty analysis is developed for the proposed GA-PSO algorithm and compared with GA and PSO. The results indicate that the proposed hybrid approach outperforms the PSO and GA algorithms in providing a better and more stable multi-objective optimized construction solution, with performance improvements of 4.3–8.5% and volatility reductions of 37.5–64.4%. This provides a reference for the optimal design of wind farms, buildings, and other construction projects.

A computationally intelligent framework for traffic engineering and congestion management in software-defined network (SDN)

Software-defined networking (SDN) revolutionizes network administration by centralizing control and decoupling the data plane from the control plane. Despite its advantages, the escalating volume of network traffic induces congestion at nodes, adversely affecting routing quality and overall performance. Addressing congestion has become imperative due to its emergence as a fundamental challenge in network management. Previous strategies often faced drawbacks in handling congestion, with issues arising from the inability to efficiently manage heavy packet surges in specific network regions. In response, this research introduces a novel approach integrating a multiplicative gated recurrent neural network with a congestion-aware hunter prey optimization (HPO) algorithm for effective traffic management in SDN. The framework leverages machine learning and deep learning techniques, acknowledged for their proficiency in processing traffic data. Comparative simulations showcase the congestion-aware HPO algorithm's superiority, achieving a normalized throughput 3.4–7.6% higher than genetic algorithm (GA) and particle swarm optimization (PSO) alternatives. Notably, the proposed framework significantly reduces data transmission delays by 58–65% compared to the GA and PSO algorithms. This research not only contributes a state-of-the-art solution but also addresses drawbacks observed in existing methodologies, thereby advancing the field of traffic engineering and congestion management in SDN. The proposed framework demonstrates notable enhancements in both throughput and latency, providing a more robust foundation for future SDN implementations.

Design of electrostatic lenses through genetic algorithm and particle swarm optimisation methods integrated with differential algebra

Genetic algorithm (GA) and particle swarm optimisation (PSO) techniques have been integrated with the differential algebra (DA) method in charged particle optics to optimise an Einzel lens. The DA method is a robust and efficient tool for the calculation of aberration coefficients of electrostatic lenses, which makes use of nonstandard analysis for ray tracing a particle as it is subjected to the field generated by a lens. In this study, initial populations of lenses with random geometrical configurations are generated. These initial populations are then subjected to GA and PSO algorithms to alter the geometry of each lens for a set number of iterations. The lens performance is evaluated by calculating the spot size using the aberrations coefficients up to third-order generated by the DA method. Moreover, a focusing column comprising two lenses and a Wien filter was optimised using GA method.

Trajectory tracking control based on genetic algorithm and proportional integral derivative controller for two-wheel mobile robot

This paper uses the genetic algorithm (GA) to optimize the proportional integral derivative (PID) controller parameters to present the motion control design for a two-wheeled mobile robot autonomous system. The GA algorithm determines a collision-free travel curve for a robot with a tangential velocity restriction constraint. A trajectory-tracking controller based on the PID control structure is developed to monitor the calculated route curves for the mobile robot. Simulation results show the effectiveness of the GA-PID controller compared to the PID controller. The GA-PID controller demonstrates improved performance in trajectory tracking and collision avoidance, making it suitable for controlling the motion of two-wheeled mobile robots. The GA's optimization process allows for better tuning of the PID controller parameters, resulting in more efficient and accurate robot motion control. The results suggest that the proposed GA-PID controller is a promising approach for enhancing mobile robots' autonomous navigation capabilities.

Enhancing SILCS-MC via GPU Acceleration and Ligand Conformational Optimization with Genetic and Parallel Tempering Algorithms.

In the domain of computer-aided drug design, achieving precise and accurate estimates of ligand-protein binding is paramount in the context of screening extensive drug libraries and performing ligand optimization. A fundamental aspect of the SILCS (site identification by ligand competitive saturation) methodology lies in the generation of comprehensive 3D free-energy functional group affinity maps (FragMaps), encompassing the entirety of the target molecule structure. These FragMaps offer an intricate landscape of functional group affinities across the protein, bilayer, or RNA, acting as the basis for subsequent SILCS-Monte Carlo (MC) simulations wherein ligands are docked to the target molecule. To augment the efficiency and breadth of ligand sampling capabilities, we implemented an improved SILCS-MC methodology. By harnessing the parallel computing capability of GPUs, our approach facilitates concurrent calculations over multiple ligands and binding sites, markedly enhancing the computational efficiency. Moreover, the integration of a genetic algorithm (GA) with MC allows us to employ an evolutionary approach to perform ligand sampling, assuring enhanced convergence characteristics. In addition, the potential utility of parallel tempering (PT) to improve sampling was investigated. Implementation of SILCS-MC on GPU architecture is shown to accelerate the speed of SILCS-MC calculations by over 2-orders of magnitude. Use of GA and PT yield improvements over Markov-chain MC, increasing the precision of the resultant docked orientations and binding free energies, though the extent of improvements is relatively small. Accordingly, significant improvements in speed are obtained through the GPU implementation with minor improvements in the precision of the docking obtained via the tested GA and PT algorithms.

Near-infrared spectroscopy and multivariate analysis as effective, fast, and cost-effective methods to discriminate Candida auris from Candida haemulonii.

Candida auris and Candida haemulonii are two emerging opportunistic pathogens that have caused an increase in clinical cases in the recent years worldwide. The differentiation of some Candida species is highly laborious, difficult, costly, and time-consuming depending on the similarity between the species. Thus, this study aimed to develop a new, faster, and less expensive methodology for differentiating between C. auris and C. haemulonii based on near-infrared (NIR) spectroscopy and multivariate analysis. C. auris CBS10913 and C. haemulonii CH02 were separated in 15 plates per species, and three isolated colonies of each plate were selected for Fourier transform near-infrared (FT-NIR) analysis, totaling 90 spectra. Subsequently, principal component analysis (PCA) and variable selection algorithms, including the successive projections algorithm (SPA) and genetic algorithm (GA) coupled with linear discriminant analysis (LDA), were employed to discern distinctive patterns among the samples. The use of PCA, SPA, and GA algorithms associated with LDA achieved 100% sensitivity and specificity for the discriminations. The SPA-LDA and GA-LDA algorithms were essential in selecting the variables (infrared wavelengths) of most importance for the models, which could be attributed to binding of cell wall structures such as polysaccharides, peptides, proteins, or molecules resulting from yeasts' metabolism. These results show the high potential of combined FT-NIR and multivariate analysis techniques for the classification of Candida-like fungi, which can contribute to faster and more effective diagnosis and treatment of patients affected by these microorganisms.

Comparison of Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) for Estimating the Susceptible-Exposed-Infected-Recovered (SEIR) Model Parameter Values

Background: The most commonly used mathematical model for analyzing disease spread is the Susceptible-Exposed-Infected-Recovered (SEIR) model. Moreover, the dynamics of the SEIR model depend on several factors, such as the parameter values. Objective: This study aimed to compare two optimization methods, namely genetic algorithm (GA) and particle swarm optimization (PSO), in estimating the SEIR model parameter values, such as the infection, transition, recovery, and death rates. Methods: GA and PSO algorithms were compared to estimate parameter values of the SEIR model. The fitness value was calculated from the error between the actual data of cumulative positive COVID-19 cases and the numerical data of cases from the solution of the SEIR COVID-19 model. Furthermore, the numerical solution of the COVID-19 model was calculated using the fourth-order Runge-Kutta algorithm (RK-4), while the actual data were obtained from the cumulative dataset of positive COVID-19 cases in the province of Jakarta, Indonesia. Two datasets were then used to compare the success of each algorithm, namely, Dataset 1, representing the initial interval for the spread of COVID-19, and Dataset 2, representing an interval where there was a high increase in COVID-19 cases. Results: Four parameters were estimated, namely the infection rate, transition rate, recovery rate, and death rate, due to disease. In Dataset 1, the smallest error of GA method, namely 8.9%, occurred when the value of , while the numerical error of PSO was 7.5%. In Dataset 2, the smallest error of GA method, namely 31.21%, occurred when , while the numerical error of PSO was 3.46%. Conclusion: Based on the parameter estimation results for Datasets 1 and 2, PSO had better fitting results than GA. This showed PSO was more robust to the provided datasets and could better adapt to the trends of the COVID-19 epidemic. Keywords: Genetic algorithm, Particle swarm optimization, SEIR model, COVID-19, Parameter estimation.

Comparing Optimized Sound Absorption Coefficient of Aluminum Foam with Local Search Algorithm, Genetic Algorithm, and Particle Swarm Optimization

Introduction: The principle of passive sound control is based on the phenomenon of sound absorption by absorbers. The factors affecting sound absorption include porosity, pore size, pore opening size, thickness, and air flow resistance. Materials and methods: In this study, the authors compared the optimization results of the effective parameters on sound absorption coefficient (AC) using the three optimization methods: Guided Local Search (GLS), Genetic Algorithm (GA) and Particle Swarm Optimization (PSO). The programming was done in MATLAB software. Thicknesses of 5, 10, 20, 30 and 40 mm were chosen for optimization at frequencies of 500 to 3000 Hz. Results: In frequencies above 2 kHz (thickness 5 to 40 mm), the three optimal methods had the same performance and estimated AC of 1. At low frequencies of 2 kHz and thicknesses of 30 and 40 mm, GA and PSO methods obtained an AC of 1. Conclusion: It seems that the GA and PSO optimization algorithm are suitable methods to optimize the AC of metal foam in low and high frequencies.

Development and evaluation of parallel simulated annealing algorithm for reactor core optimization problems

Background The optimization of core loading patterns in nuclear reactors is one of the most studied optimization problems in nuclear engineering due to the enormous economical and safety benefits. Various algorithms such as Genetic Algorithms (GA), Simulated Annealing (SA), and Parallel Simulated Annealing (PSA) have been used in the past for such problems. Methods In this work, a PSA algorithm was developed and integrated into the Modularly Implemented Design Assistance Suite (MIDAS), a framework developed at North Carolina State University to solve nuclear engineering problems. The effectiveness of PSA was compared against the GA and SA algorithms available in MIDAS for a Pressurized Water Reactor first cycle core loading pattern optimization problem. Results PSA consistently generates more optimal solutions than SA and GA by having the higher average fitness, and showing less variance in its performance and thus being more robust. This provides confidence in the PSA implementation within MIDAS. The obtained loading pattern positions high reactive fuel in peripheral locations and low reactive fuel towards the centers in a strategy resembling both Out-In-Checkboard and L3P loading pattern approaches. Conclusions Future studies will involve applying the PSA algorithm to other optimization studies in larger combinatorial spaces, such as in multi-cycle optimization problems.

Metaheuristics Approach for Hyperparameter Tuning of Convolutional Neural Network

Deep learning is an artificial intelligence technique that has been used for various tasks. Deep learning performance is determined by its hyperparameter, architecture, and training (connection weight and bias). Finding the right combination of these aspects is very challenging. Convolution neural networks (CNN) is a deep learning method that is commonly used for image classification. It has many hyperparameters; therefore, tuning its hyperparameter is difficult. In this research, a metaheuristic approach is proposed to optimize the hyperparameter of convolution neural networks. Three metaheuristic methods are used in this research: ant colony optimization (ACO), genetic algorithm (GA), and Harmony Search (HS). The metaheuristics methods are used to find the best combination of 8 hyperparameters with 8 options each which creates 1.6. 107 of solution space. The solution space is too large to explore using manual tuning. The Metaheuristics method will bring benefits in terms of finding solutions in the search space more effectively and efficiently. The performance of the metaheuristic methods is evaluated using MNIST datasets. The experiment results show that the accuracy of ACO, GA and HS are 99,7%, 97.7% and 89,9% respectively. The computational times for the ACO, GA and HS algorithms are 27.9 s, 22.3 s, and 56.4 s, respectively. It shows that ACO performs the best among the three algorithms in terms of accuracy, however, its computational time is slightly longer than GA. The results of the experiment reveal that the metaheuristic approach is promising for the hyperparameter tuning of CNN. Future research can be directed toward solving larger problems or improving the metaheuristics operator to improve its performance.

Friction compensation control method for a typical excavator system based on the accurate friction model

Traditional identification methods make it difficult to accurately determine the parameters of the friction model. Based on the LuGre friction model, a hybrid optimization algorithm (HO) is presented to address such aspects. This algorithm combines the genetic algorithm (GA) and particle swarm optimization algorithm (PSO) to obtain the globally optimal solutions. It incorporates three main improvements: introducing selection and crossover operations into the PSO algorithm, utilizing a variable inertia weight coefficient, and applying variable learning coefficients for enhanced performance. Compared to the static optimal solutions of the GA and PSO algorithms, the HO algorithm can improve the optimal solution by 40.68 % and 35.89 % respectively. The HO algorithm can achieve the highest model accuracy: the maximum identified friction error with the HO algorithm is only 1.70 kN, compared to 3.69 kN with the GA method and 8.52 kN with the PSO algorithm. Furthermore, a novel adaptive friction compensation controller (AC) is introduced based on the identification results to improve the stability and accuracy of the electro-hydraulic proportional systems of a 23-ton robotic excavator. Comparisons with existing proportional integral differential (PID) and feedforward compensation controllers (FC) highlight the superiority of the proposed adaptive controller in terms of trajectory accuracy and robustness. In sinusoidal trajectory tracking experiments, the AC method can achieve the highest tracking performance among the three controllers, and the maximum tracking error is only 11.56 mm. Compared to the FC controller, the control accuracy with the AC controller can be improved by 43.38 %. Experiments and simulations have shown that the friction compensation controller can effectively mitigate friction and other disturbances. As a result, the phenomena of crawling and flat peak phenomena can be eliminated.