Genetic Algorithm Model Research Articles

Distribution network power flow calculation based on the BPNN optimized by GA‐ADAM

AbstractPower system operation and control are based on power flow calculations. In order to solve the uncertainty of the increasing penetration of renewable energy, the voltage fluctuation at the load point increases in the distribution network, and the inaccuracy of the power flow calculation due to the insufficient power flow data collection capability of the traditional power system. In this paper, a data‐driven power flow analysis model is proposed, a back propagation neural network combined with genetic algorithm (GA) and adaptive moment estimation (ADAM) optimization algorithm model is constructed to analyze the power flow calculation method of distribution networks under stochasticity. Firstly, the power flow initial value information, topology characteristics, and power factor index are introduced to construct a training set, and the mapping relationship between bus voltage and power is fully explored by training the regression model. Second, the GA‐ADAM algorithm is used to optimize the initial values and weight parameters of the model. Finally, it is verified based on IEEE‐33 bus distribution model, and the model is used for power flow calculation, and compared with other methods through each relevant error evaluation indicators. The results show that the model constructed in this paper has small error indicators and high accuracy, which improves the efficiency and accuracy of power flow calculation.

Planning of prescriptive maintenance types for generator with fuzzy logic-based genetic algorithm in a hydroelectric power plant

The move to predictive and prescriptive maintenance underlines the need for cost-effective methods to monitoring degradation and predicting failures. It is in hydroelectric power plants (HPPs) where the most powerful and unique techniques of this evolution were observed. The final premise of this transition is the need to improve the decision-making process for planning activities in the Prescriptive Maintenance Planning (MP). This encourages the use of fuzzy set theory in process management. Although there are many studies on MP in the literature, these studies do not fully reflect the real life as they ignore potential schedule delays and the indirect impact of sustainable strategy on revenues. Therefore, in the first stage of this study, it is aimed to weight the maintenance types (MTs) on a plant basis for the generator, which is the most critical system of the pumped storage HPP, and in the second stage, it is aimed to create MPs that minimize delays due to system shutdowns with weighted MTs. For this purpose, in the first stage, the MTs for the HPP were weighted by AHP-TOPSIS combination augmented with Pythagorean fuzzy numbers, and in the second stage, a Genetic Algorithm model was designed for the problem of coordinating maintenance durations and deadlines in addition to these weights. Averaged over 50 iterations, the average number of delayed MPs is 142.55, indicating that the algorithm is successful in reducing delayed MPs, while the average weighted delay value is 172.85, indicating that more work is needed in terms of weighted delay.

Multi-objective diesel engine emission management and control technology based on SVM and NSGA-II

To improve the emission control effect of diesel engines, a pre-emission model for diesel engines using biofuels is constructed using support vector machine. Considering that traditional support vector machine training relies on initial parameters and is prone to local convergence, a genetic algorithm model is adopted to construct an improved diesel engine pre-emission model. At the same time, with the optimization goal of diesel engine emissions of particles and gases, the improved diesel engine pre-emission model is used to obtain emission data as training parameters. The second generation non-dominated genetic algorithm is introduced to construct a multi-objective diesel engine emission control model, achieving optimal control of diesel engine emissions. In the experimental analysis of diesel engine pre emission models, the proposed pre emission model has a better prediction effect in nitrogen oxides, with an average prediction accuracy of 0.986, which is better than the Particle Swarm Optimization-Support Vector Machine model's 0.906, and is more accurate in detecting emissions. Simultaneously predicting particulate matter, the average prediction accuracy of the proposed model is 0.978, while the prediction accuracy of the Particle Swarm Optimization-Support Vector Machine model is 0.902, indicating that the overall prediction effect of the proposed method is better. In the experimental analysis of multi-objective optimization models, the proposed multi-objective optimization model performs best in diesel emission optimization. The proposed model can achieve convergence in the shortest time, and the minimum nitrogen oxide emissions are 210.7 ppm. Compared with Non-dominated Sorting Genetic Algorithm II and Multi-Objective Particle Swarm Optimization models, the particle emission optimization effect is improved by 23.5 % and 18.6 %. From this, it can be seen that the proposed method has excellent performance in diesel engine emission monitoring and pollutant optimization, surpassing related technologies in the same period. The research technology provides important technical references for the monitoring and emission control of harmful substances in diesel engines.

Software fault prediction with imbalanced datasets using SMOTE-Tomek sampling technique and Genetic Algorithm models

Software fault prediction with imbalanced datasets using SMOTE-Tomek sampling technique and Genetic Algorithm models

Reducing passenger waiting time and energy consumption for metro timetabling: A robust optimization approach

Urban subway scheduling is a quite complicated task as any small fluctuation could disturb the entire system. Subway companies need an efficient scheduling system that best serves passengers and optimizes the utilization of resources. They need, therefore, to include real-world problems in their schedule through non-deterministic modeling. A robust optimization approach is applied in real-world data to propose an optimum train schedule that minimizes passenger waiting time and maximizes electrical energy converted from kinetic energy. Three high, normal, and low travel demand scenarios are considered to optimize train running-time duration in all three interstation movement categories, namely acceleration, coasting, and braking. To find the optimal solution, a genetic algorithm (GA) model has been applied to real data from a subway line in Tehran Metro.

Modeling and optimization of poly(lactic acid)/poly(ℇ-caprolactone)/Nigella sativa extract nanofibers production for skin wounds healing by artificial neural network and response surface methodology models

Electrospun fibrous scaffolds have great potential for the effective treatment of wounds. Novel blend scaffolds were fabricated from poly(ℇ- caprolactone) (PCL)/poly (lactic acid) (PLA) with Nigella sativa (NS) extract in different concentrations of 10 %, 15 %, 20 %, and 25 % by one nozzle electrospinning. RSM and ANN models were used to determine optimal nanofiber. The results showed that the ANN model had average goodness values of almost 1.992 which was higher than the RSM model with an amount of 1.823. The best sample was determined with the combination of parameters such as PLA/PCL (70:29) concentration, voltage 17 kV, and flow rate 0.2 ml/h in diameter of nanofiber 410 nm by Genetic Algorithm (GA) model with cost value 0.0216 that was lower than cost value (0.0927) of ANN model. The effect of NS extract on nanofibers properties showed that loading high concentrations of NS extract in PLA/PCL polymer solutions caused a decrease in nanofibers diameter, hydrophilicity, and tensile strength. Overall, PLA/PCL/NS 25 % nanofiber was selected as an optimal web with an average diameter of 370 ± 68 nm with a young modulus 5.94 MPa. This scaffold also exhibited the highest antibacterial activity, cell attachment, and cell viability based on the MTT assay.

A green approach for multivariate consecutive extraction of essential oils and flavonoids from Citrus aurantium L. var. amara Engl.: Process optimization and mechanistic insights based on machine learning methods

In this study, a green approach for the multivariate consecutive extraction of essential oils and total flavonoids from Daidaihua was proposed, which combined classical steam distillation with alkali extraction and acid precipitation. The process parameters were optimized through single-factor experiments and Box-Behnken design, and the model for evaluating and predicting the process was constructed by combining artificial neural network and genetic algorithm. The artificial neural network and genetic algorithm model demonstrated satisfactory predictive capabilities, and the following optimal parameters resulted: solid-liquid ratio 1:15.3 (g:mL), distillation time 94.1 h, NaOH concentration 5.4%, extraction temperature 46.7 ℃and extraction time 52.9 min. Under these conditions, the extraction yields of essential oils and total flavonoids reached 0.81 ± 0.07% and 5.52 ± 0.18%, respectively. Significantly, the mass transfer mechanism of the multivariate consecutive extraction process was clearly elucidated using the mass transfer theory model and physical characterization. The essential oils were vaporized from the Daidaihua cells along with water vapor, promoting the destruction of cell structure, reducing the extraction resistance of flavonoids, and speeding up the diffusion of sodium hydroxide solution rapidly into the cell interior and dissolving the flavonoids, improving the efficiency of the mass transfer of flavonoids from the inside to the outside of the cell, consequently the flavonoids extraction yield and efficiency were significantly improved. Obviously, the extraction process dovetailed neatly with environmentally friendly and low-carbon needs. Meanwhile, the composition of the essential oils and total flavonoids was identified by GC-MS and HPLC-MS analysis, respectively. Overall, this research presented a green and clean multivariate consecutive extraction process of Daidaihua, which could contribute to promoting the upgrading and sustainable development of the plant extraction industry.

Thermodynamic performance of a cryogenic energy storage system based on natural gas liquefaction

Cryogenic energy storage (CES) is a viable method for grid-scale electrical energy storage. Considering the high energy density and mature application of liquefied natural gas (LNG), we proposed an LNG cryogenic energy storage (LNGES) system. A steady-state process model of the LNGES system was established using Aspen HYSYS. The effects of the natural gas composition and key operating parameters such as the charging pressure, discharging pressure, throttling temperature, and liquid storage pressure on the system performance were investigated. A multi-parameter genetic algorithm model built using the MATLAB software was used to optimize the LNGES system to optimize the round-trip efficiency (RTE). Then, an exergy analysis of the optimal configuration was conducted. The results suggested that the LNGES system could achieve optimal RTE and exergy efficiency values of 60.14% and 71.64%, respectively. Exergy destruction mainly occurred during the compression, throttling, expansion, and heat exchange. The proposed LNGES system could be a promising candidate for the large-scale application of CES technology in power grids and gas networks.

Adsorptive elimination of methylene blue dye from aqueous solution by chitosan‐nSiO2 nanocomposite: Adsorption and desorption study, scale‐up design, statistical, and genetic algorithm modeling

AbstractHeavy metals and organic dyes in water in concentrations exceeding the tolerable limit are unsafe for aquatic life and humans. Chitosan nanocomposites show a potential adsorption capacity to organic dyes. The present study encompasses the preparation of chitosan‐nSiO2 nanocomposites (CSNC) with different weight ratios of chitosan to nSiO2 and its application in the methylene blue (MB) dye adsorption. The nanocomposites were characterized with the help of SEM, BET, FTIR, XRD, and TGA. The adsorption experimentation was executed in batch mode under varying experimental conditions. Several isotherm and kinetic models were analyzed with the experimental data, and thermodynamic conditions required for adsorption were also determined. Maximum Langmuir adsorption capacities (qL) of the adsorbents for MB varied in the range of 21.32–31.34 mg g−1. The pseudo‐second‐order model was the best‐fitted kinetic model concerning all three adsorbents. The statistical modeling using multiple polynomial regression (MPR) for CWS CSNC1‐1 and CSNC2‐1 yielded the equations with R‐square ranging from 0.984 to 0.996. For GA modeling, it is more than 0.999. So, the efficient employment of statistical modeling, and genetic algorithms has also been achieved.

Heat transfer enhancement of serpentine channels with twisted tape insert by computational fluid dynamics and artificial intelligence

AbstractImproving the thermal–hydraulic performance of heat exchangers is a critical challenge. In this study, we investigated the impact of twisted tape insert geometrical variables, namely pitch length (P) and diameter (W), on the Nusselt number (Nu) and friction factor (f) within serpentine channels. To accomplish this, we employed computational fluid dynamics (CFD) as our simulation technique, with water as the working fluid. The validity of the CFD data was confirmed, leading to the development of artificial intelligence (AI) subsets: artificial neural network (ANN) and genetic algorithm (GA). In the prediction models, the following input variables were considered: Reynolds number (Re), (P/Dh), and (W/Dh). The ANN and GA models achieved a mean relative error (MRE) of 0.125% and 1.326% for Nu, and 0.139% and 4.104% for f, respectively. These results demonstrate the high accuracy of the correlations, with the ANN model showcasing superior performance.

Design of an efficient convolutional buck-boost converter for hybrid bioinspired parameter tuning

Power-electronic systems with voltage boosts use buck-boost converters. These converters suppress current and invert voltage to improve voltage swing. Power-electronic systems with voltage boosts use buck-boost converters that suppress current and invert voltage to improve voltage swings. Researchers propose many converter models, but their total harmonic distortion (THD) limits their scalability. Harmonics from additional current components increase THD. The model filters excessive currents using inductor-based storage, capacitive filters, and resistive circuits. However, these models are unstable, reducing their performance in large converter circuits. This text proposes a novel convolutional neural network (CNN) with a hybrid bioinspired model based on genetic algorithm (GA) and particle swarm optimization (PSO) to overcome this limitation. Estimating internal buck and boost parameters efficiently reduces reverse currents. These parameters include inductor current ripple, recommended inductance, internal switch current limit, and switching frequency. The model finds low-power, high-efficiency buck-boost configurations based on these values. Incremental learning operations tuned the GA model, which was applied to many buck-boost configurations. The proposed model had a 5.9% lower delay, 16.2% lower harmonics, and 4.6% better power efficiency than state-of-the-art buck-boost models.

Multi-objective optimization of breakthrough times for hydrogen purification through layered bed pressure swing adsorption based on genetic algorithm and artificial neural network model

Hydrogen purification from steam methane reforming (SMR) by pressure swing adsorption (PSA) technology is a common method to obtain high-purity hydrogen. The breakthrough time can well reflect the adsorption dynamics and help PSA cycle design. In order to avoid time-consuming and labor-intensive breakthrough curve experiments, it is very necessary to develop a fast and accurate surrogate model. In this study, a genetic algorithm (GA) and artificial neural network (ANN) are combined to predict and optimize breakthrough times of adsorbates in activated carbon/zeolite layered beds. Using the Latin hypercube sampling strategy, training data sets of GA-optimized ANN (GA-ANN) obtains from the physical model of adsorption, heat and mass transfer model. The genetic algorithm (GA) optimizes the weights and biases of ANN for better performance. The ANN topology has 4 input variables (superficial velocity, activated carbon height, adsorption pressure and feed temperature) and 3 output variables (breakthrough times of CH4, CO and CO2). The number of neurons in the hidden layer for the GA-ANN model was optimized as 7 to predict and optimize the maximum breakthrough time of SMR with a four-component (H2/CH4/CO/CO2) system. The sensitivity analysis displays that relative importance to the breakthrough time is in the order of superficial velocity (40.88%) > adsorption pressure (24.55%) > activated carbon height (21.04%) > feed temperature (13.53%). To obtain high-purity hydrogen, the GA-ANN model combined with multi-objective GA optimization is used to maximize the breakthrough times of CH4 and CO. The GA-ANN surrogate model proposed in this paper can not only accurately and quickly achieve the purpose of predicting the system breakthrough time with a high correlation coefficient (R = 0.9937) but also obtains the optimal operating conditions of PSA hydrogen purification.

Stackelberg game-based three-stage optimal pricing and planning strategy for hybrid shared energy storage

Inspired from sharing economy and advanced energy storage technologies, hybrid shared energy storage (HSES), as an innovative business model, can provide flexible storage leasing services to new energy stations (NESs) and bring additional profits to the energy storage owner. Under this business model, pricing and planning issues are the main focus of the HSES operator to increase revenues but are rarely considered in current studies. Therefore, a Stackelberg game-based three-stage optimal pricing and planning strategy of HSES is formulated for the operator. First, an HSES model considering two leasing options is developed to provide two kinds of short-term use rights of energy storage resources for NESs. Then, the interactions between selfish NESs and the HSES operator are characterized as a Stackelberg game, and a bi-level pricing and planning strategy optimization model is developed to help the HSES operator make optimal decisions. Finally, considering different characteristics in each stage of the Stackelberg game, a three-stage solution method based on the genetic algorithm (GA) and mixed-integer linear programming (MILP) models is proposed to solve the optimization problem. Case studies on six NESs in a certain region are taken to verify the effectiveness of the proposed strategy. Simulation results show that the HSES operator can obtain maximum profit under the proposed pricing and planning strategy. In addition, the proposed HSES leasing model can provide additional benefits to both the operator and NESs.

Modelo matemático basado en algoritmos genéticos para optimizar las utilidades en una empresa de transporte interprovincial de pasajeros

The main objective of this research was to improve profits in an interprovincial passenger transport company, having designed a mathematical programming model and a genetic algorithm model with the purpose of reaching solutions for a future optimization of the company's profits. The type of research was applied, the technique that was applied was the survey and the sample was made up of all the variables of income, costs and profits of an interprovincial passenger transport company in the Peruvian territory. The results confirmed that the implementation of a mathematical model based on genetic algorithms contributes significantly to the optimization of profits, since the profit associated with the economic bus was improved, to the sleeper bus, to the mixed bus and to the super sleeper bus, achieving an optimal occupancy rate of 58.33%, 85.00%, 58.33% and 95.00%, respectively. The implemented model was based on a chromosome size of 19 bits, the population size was 10 chromosomes, the selection method is probabilistic with 4 chromosomes, crossing with four chromosomes, type of multipoint crossing and mutation process a from the second iteration of probabilistic type.

Credit Risk Research on Chinese Real Estate Enterprises Based on Modified KMV Model

This paper first analyses the current situation of credit risk in China’s real estate industry, and then compares the traditional and modern credit risk measurement models. On this basis, the KMV model is selected, and the artificial intelligence model Genetic Algorithm (GA) and GARCH model are introduced to improve the accuracy of the KMV model. Secondly, the annual financial data and stock trading data of 24 real estate listed companies for 2018 – 2022 are selected for empirical research. By analyzing the total default distance of the 24 companies and the actual economic development of China, it is proved that the results of the GA-GARCH-KMV model are 8% more correct than the classical KMV model, which indicates that the model has better applicability.

Application of GA-WELM Model Based on Stratified Cross-Validation in Intrusion Detection

Aiming at the problem of poor detection performance under the environment of imbalanced type distribution, an intrusion detection model of genetic algorithm to optimize weighted extreme learning machine based on stratified cross-validation (SCV-GA-WELM) is proposed. In order to solve the problem of imbalanced data types in cross-validation subsets, SCV is used to ensure that the data distribution in all subsets is consistent, thus avoiding model over-fitting. The traditional fitness function cannot solve the problem of small sample classification well. By designing a weighted fitness function and giving high weight to small sample data, the performance of the model can be effectively improved in the environment of imbalanced type distribution. The experimental results show that this model is superior to other intrusion detection models in recall and McNemar hypothesis test. In addition, the recall of the model for small sample data is higher, reaching 91.5% and 95.1%, respectively. This shows that it can effectively detect intrusions in an environment with imbalanced type distribution. Therefore, the model has practical application value in the field of intrusion detection, and can be used to improve the performance of intrusion detection systems in the actual environment. This method has a wide application prospect, such as network security, industrial control system, and power system.

Intelligent and Resource-Conserving Service Function Chain (SFC) Embedding

Network Function Virtualization (NFV) opens us great opportunities for network processing with higher resource efficiency and flexibility. In this respect, there is an increasing need for intelligent orchestration mechanisms, such that NFV can exploit its potential and live up to its promise. Genetic algorithms have emerged as a promising alternative to the proliferation of heuristic and exact methods for the Service Function Chain (SFC) embedding problem. To this end, we design and evaluate a genetic algorithm (GA), which computes efficient embeddings with runtimes on par with approximate methods. We present a GA model as state-space search in order to clarify the design choices of a GA. Our proposed GA utilizes a heuristic for the generation of the initial population, with the aim of directing the search towards the solution. Given the sensitivity of GAs on their various parameters, we introduce a parameter adjustment framework for GA fine-tuning. A comparative evaluation among a range of GA variants with diverse features sheds light on the impact of these features on SFC embedding efficiency. The GA variant that stands out is further benchmarked against a baseline greedy algorithm and a state-of-the-art heuristic. Our evaluation results indicate that the GA yields notable gains in terms of request acceptance and resource efficiency.

WOA-AGA Algorithm Design for Robot Path Planning

Currently, mobile robot has great application value in industrial production. It can play a unique advantage in improving industrial production efficiency and saving industrial production costs. Path planning plays an important role in the performance of mobile robots. Therefore, to improve the path planning efficiency of mobile robots in complex environments, a path planning model combining genetic algorithm (GA) and whale optimization algorithm (WOA), namely WOAAGA model, is proposed. In the model, the traditional GA model is introduced into the difference degree function. WOA makes up for the local optimization problem and the low proficiency of AGA algorithm. WOA-AGA effectively solves the problems of local optimization, long convergence time and unstable optimization results. The experiment is simulated in dynamic and static environment: AGA algorithm has 1.87% higher efficiency than GA algorithm; Compared with AGA algorithm, the overall operation efficiency of WOA-AGA algorithm is increased by 3.87%. Finally, two types of complex scenes are selected for path planning in the experiment. The results indicate that WOAAGA algorithm can obtain shorter and more reasonable optimal path than other similar algorithms. From the perspective of improving the path planning effect of mobile robots, this study aims to obtain the best path through the reasonable application of WOA-AGA model to improve industrial production efficiency.

Parameters prediction in additively manufactured Al-Cu alloy using back propagation neural network

The relationship between tensile strength, wire feeding speed and travel speed is built based on Back Propagation (BP) neural network during the wire arc additive manufacturing (WAAM) process. The introduction of a genetic algorithm for optimising the BP neural network (GA-BP) and incorporation of additional parameter combinations through the forward model markedly enhance the prediction accuracy of the process parameter reverse model. The BP neural network with a genetic algorithm model exhibits excellent training results, and the sample population regression reaches 0.97. An error value of the optimised model is only 3.10% for wire feeding speed prediction, only 1.55% for travel speed prediction. The GA-BP reverse model optimises WAAM process parameters and achieves a tensile strength exceeding 230 MPa.

Predicting aqueous sorption of organic pollutants on microplastics with machine learning

Microplastics (MPs) are ubiquitously distributed in freshwater systems and they can determine the environmental fate of organic pollutants (OPs) via sorption interaction. However, the diverse physicochemical properties of MPs and the wide range of OP species make a deeper understanding of sorption mechanisms challenging. Traditional isotherm-based sorption models are limited in their universality since they normally only consider the nature and characteristics of either sorbents or sorbates individually. Therefore, only specific equilibrium concentrations or specific sorption isotherms can be used to predict sorption. To systematically evaluate and predict OP sorption under the influence of both MPs and OPs properties, we collected 475 sorption data from peer-reviewed publications and developed a poly-parameter-linear-free-energy-relationship-embedded machine learning method to analyze the collected sorption datasets. Models of different algorithms were compared, and the genetic algorithm and support vector machine hybrid model displayed the best prediction performance (R2 of 0.93 and root-mean-square-error of 0.07). Finally, comparison results of three feature importance analysis tools (forward step wise method, Shapley method, and global sensitivity analysis) showed that chemical properties of MPs, excess molar refraction, and hydrogen-bonding interaction of OPs contribute the most to sorption, reflecting the dominant sorption mechanisms of hydrophobic partitioning, hydrogen bond formation, and π-π interaction, respectively. This study presents a novel sorbate-sorbent-based ML model with a wide applicability to expand our capacity in understanding the complicated process and mechanism of OP sorption on MPs.