Wyniki wyszukiwania - BazTech

1

Robust dual color images watermarking scheme with hyperchaotic encryption based on quaternion DFrAT and genetic algorithm

Luo Hui-Xin, Gong Li-Hua, Chen Su-Hua

Optica Applicata

|

2023

|

Vol. 53, nr 2

261--280

EN

A robust dual color images watermarking algorithm is designed based on quaternion discrete fractional angular transform (QDFrAT) and genetic algorithm. To guarantee the watermark security, the original color watermark image is encrypted with a 4D hyperchaotic system. A pure quaternion matrix is acquired by performing the discrete wavelet transform (DWT), the block division and the discrete cosine transform on the original color cover image. The quaternion matrix is operated by the QDFrAT to improve the robustness and the security of the watermarking scheme with the optimal transform angle and the fractional order. Then the singular value matrix is obtained by the quaternion singular value decomposition (QSVD) to further enhance the scheme’s stability. The encryption watermark is also processed by DWT and QSVD. Afterward, the singular value matrix of the encryption watermark is embedded into the singular value matrix of the host image by the optimal scaling factor. Moreover, the values to balance imperceptibility and robustness are optimized with a genetic algorithm. It is shown that the proposed color image watermarking scheme performs well in imperceptibility, security, robustness and embedding capacity.

2

Minimizing the Makespan and Total Tardiness in Hybrid Flow Shop Scheduling with Sequence-Dependent Setup Times

Mousavi Seyyed Mostafa, Shahnazari-Shahrezaei Parisa

Management and Production Engineering Review

|

2023

|

Vol. 14, No. 1

13--24

EN

The paper considers the production scheduling problem in a hybrid flow shop environment with sequence-dependent setup times and the objectives of minimizing both the makespan and the total tardiness. The multi-objective genetic algorithm is applied to solve this problem, which belongs to the non-deterministic polynomial-time (NP)-hard class. In the structure of the proposed algorithm, the initial population, neighborhood search structures and dispatching rules are studied to achieve more efficient solutions. The performance of the proposed algorithm compared to the efficient algorithm available in literature (known as NSGA-II) is expressed in terms of the data envelopment analysis method. The computational results confirm that the set of efficient solutions of the proposed algorithm is more efficient than the other algorithm.

3

Synthesis of Reconfigurable Multiple Shaped Beams of a Concentric Circular Ring Array Antenna Using Evolutionary Algorithms

Dubey Sanjay Kumar, Mandal Debasis

Journal of Telecommunications and Information Technology

|

2023

|

nr 1

8--17

EN

The approach described in this paper uses evolutionary algorithms to create multiple-beam patterns for a concentric circular ring array (CCRA) of isotropic antennas using a common set of array excitation amplitudes. The flat top, cosec2, and pencil beam patterns are examples of multiple-beam patterns. All of these designs have an upward angle of θ = 0◦. All the patterns are further created in three azimuth planes (φ = 0◦, 5◦, and 10◦). To create the necessary patterns, non-uniform excitations are used in combination with evenly spaced isotropic components. For the flat top and cosecant-squared patterns, the best combination of common components, amplitude and various phases is applied, whereas the pencil beam pattern is produced using the common amplitude only. Differential evolutionary algorithm (DE), genetic algorithm (GA), and firefly algorithm (FA) are used to generate the best 4-bit discrete magnitudes and 5-bit discrete phases. These discrete excitations aid in lowering the feed network design complexity and the dynamic range ratio (DRR). A variety of randomly selected azimuth planes are used to verify the excitations as well. With small modifications in the desired parameters, the patterns are formed using the same excitation. The results proved both the efficacy of the suggested strategy and the dominance of DE over GA as well as FA.

4

Beam Pattern Optimization Via Unequal Ascending Clusters

Abdulqader Ahmed Jameel, Mohammed Jafar Ramadhan, Mohammad Ali Yessar E.

Journal of Telecommunications and Information Technology

|

2023

|

nr 1

1--7

EN

In this paper, two different architectures based on completely and sectionally clustered arrays are proposed to improve the array patterns. In the wholly clustered arrays, all elements of the ordinary array are divided into multiple unequal ascending clusters. In the sectionally clustered arrays, two types of architectures are proposed by dividing a part of the array into clusters based on the position of specific elements. In the first architecture of sectionally clustered arrays, only those elements that are located on the sides of the array are grouped into unequal ascending clusters, and other elements located in the center are left as individual and unoptimized items (i.e. uniform excitation). In the second architecture, only some of the elements close the center are grouped into unequal ascending clusters, and the side elements were left individually and without optimization. The research proves that the sectionally clustered architecture has many advantages compared to the completely clustered structure, in terms of the complexity of the solution. Simulation results show that PSLL in the side clustered array can be reduced to more than −28 dB for an array of 40 elements. The PSLL was −17 dB in the case of a centrally clustered array, whereas the complexity percentage in the wholly clustered array method was 12.5 %, while the same parameter for the partially clustered array method equaled 10%.

5

FSPL: A meta-learning approach for a filter and embedded feature selection pipeline

Lazebnik Teddy, Rosenfeld Avi

International Journal of Applied Mathematics and Computer Science

|

2023

|

Vol. 33, no. 1

103--115

EN

There are two main approaches to tackle the challenge of finding the best filter or embedded feature selection (FS) algorithm: searching for the one best FS algorithm and creating an ensemble of all available FS algorithms. However, in practice, these two processes usually occur as part of a larger machine learning pipeline and not separately. We posit that, due to the influence of the filter FS on the embedded FS, one should aim to optimize both of them as a single FS pipeline rather than separately. We propose a meta-learning approach that automatically finds the best filter and embedded FS pipeline for a given dataset called FSPL. We demonstrate the performance of FSPL on n = 90 datasets, obtaining 0.496 accuracy for the optimal FS pipeline, revealing an improvement of up to 5.98 percent in the model’s accuracy compared to the second-best meta-learning method.

6

A genetic algorithm based optimized convolutional neural network for face recognition

Karlupia Namrata, Mahajan Palak, Abrol Pawanesh, Lehana Parveen K.

International Journal of Applied Mathematics and Computer Science

|

2023

|

Vol. 33, no. 1

21--31

EN

Face recognition (FR) is one of the most active research areas in the field of computer vision. Convolutional neural networks (CNNs) have been extensively used in this field due to their good efficiency. Thus, it is important to find the best CNN parameters for its best performance. Hyperparameter optimization is one of the various techniques for increasing the performance of CNN models. Since manual tuning of hyperparameters is a tedious and time-consuming task, population based metaheuristic techniques can be used for the automatic hyperparameter optimization of CNNs. Automatic tuning of parameters reduces manual efforts and improves the efficiency of the CNN model. In the proposed work, genetic algorithm (GA) based hyperparameter optimization of CNNs is applied for face recognition. GAs are used for the optimization of various hyperparameters like filter size as well as the number of filters and of hidden layers. For analysis, a benchmark dataset for FR with ninety subjects is used. The experimental results indicate that the proposed GA-CNN model generates an improved model accuracy in comparison with existing CNN models. In each iteration, the GA minimizes the objective function by selecting the best combination set of CNN hyperparameters. An improved accuracy of 94.5% is obtained for FR.

7

Modelling Microcystis Cell Density in a Mediterranean Shallow Lake of Northeast Algeria (Oubeira Lake), Using Evolutionary and Classic Programming

Arif Salah, Djellal Adel, Djebbari Nawel, Belhaoues Saber, Touati Hassen, Guellati Fatma Zohra, Bensouilah Mourad

Geomatics and Environmental Engineering

|

2023

|

Vol. 17, no. 2

31--68

EN

Caused by excess levels of nutrients and increased temperatures, freshwater cyanobacterial blooms have become a serious global issue. However, with the development of artificial intelligence and extreme learning machine methods, the forecasting of cyanobacteria blooms has become more feasible. We explored the use of multiple techniques, including both statistical [Multiple Regression Model (MLR) and Support Vector Machine (SVM)] and evolutionary [Particle Swarm Optimization (PSO), Genetic Algorithm (GA), and Bird Swarm Algorithm (BSA)], to approximate models for the prediction of Microcystis density. The data set was collected from Oubeira Lake, a natural shallow Mediterranean lake in the northeast of Algeria. From the correlation analysis of ten water variables monitored, six potential factors including temperature, ammonium, nitrate, and ortho-phosphate were selected. The performance indices showed; MLR and PSO provided the best results. PSO gave the best fitness but all techniques performed well. BSA had better fitness but was very slow across generations. PSO was faster than the other techniques and at generation 20 it passed BSA. GA passed BSA a little further, at generation 50. The major contributions of our work not only focus on the modelling process itself, but also take into consideration the main factors affecting Microcystis blooms, by incorporating them in all applied models.

8

Hybrid EEG-fNIRS brain-computer interface based on the non-linear features extraction and stacking ensemble learning

Maher Asmaa, Mian Qaisar Saeed, Salankar N., Jiang Feng, Tadeusiewicz Ryszard, Pławiak Paweł, Abd El-Latif Ahmed A., Hammad Mohamed

Biocybernetics and Biomedical Engineering

|

2023

|

Vol. 43, no. 2

463--475

EN

The Brain-computer interface (BCI) is used to enhance the human capabilities. The hybridBCI (hBCI) is a novel concept for subtly hybridizing multiple monitoring schemes to maximize the advantages of each while minimizing the drawbacks of individual methods. Recently, researchers have started focusing on the Electroencephalogram (EEG) and ‘‘Functional Near-Infrared Spectroscopy” (fNIRS) based hBCI. The main reason is due to the development of artificial intelligence (AI) algorithms such as machine learning approaches to better process the brain signals. An original EEG-fNIRS based hBCI system is devised by using the non-linear features mining and ensemble learning (EL) approach. We first diminish the noise and artifacts from the input EEG-fNIRS signals using digital filtering. After that, we use the signals for non-linear features mining. These features are ‘‘Fractal Dimension” (FD), ‘‘Higher Order Spectra” (HOS), ‘‘Recurrence Quantification Analysis” (RQA) features, and Entropy features. Onward, the Genetic Algorithm (GA) is employed for Features Selection (FS). Lastly, we employ a novel Machine Learning (ML) technique using several algorithms namely, the ‘‘Naïve Bayes” (NB), ‘‘Support Vector Machine” (SVM), ‘‘Random Forest” (RF), and ‘‘K-Nearest Neighbor” (KNN). These classifiers are combined as an ensemble for recognizing the intended brain activities. The applicability is tested by using a publicly available multi-subject and multiclass EEG-fNIRS dataset. Our method has reached the highest accuracy, F1-score, and sensitivity of 95.48%, 97.67% and 97.83% respectively.

9

TOD mode: study on land planning and land use optimization for urban rail transit station area

Shen Hongtian, Ma Guimin

Archives of Transport

|

2023

|

Vol. 66, iss. 2

33--43

EN

The rational planning of land around rail transit stations in cities can effectively improve the convenience of transportation and economic development of cities. This paper briefly introduced the transit-oriented development (TOD) mode of urban planning. We constructed a hierarchical structure for evaluating the quality of land planning of urban rail transit stations through the analytic hierarchy process (AHP) method. The structure started from three large aspects, i.e., traffic volume, regional environmental quality, and regional economic efficiency, and every large aspect was divided into three small aspects. Then, an optimization model was established for land planning of rail transit stations. The land planning scheme was optimized by a genetic algorithm (GA). To enhance the optimization performance of the GA, it was improved by coevolution, i.e., plural populations iterated independently, and every population replaced the poor chromosomes in the other populations with its excellent chromosomes in the previous process. Finally, the Jinzhonghe street station in Hebei District, Tianjin city, was taken as a subject for analysis. The results suggested that the improved GA obtained a set of non-inferior Pareto solutions when solving a multi-objective optimization problem. The distribution of solutions in the set also indicated that any two objectives among traffic volume, environmental quality, and economic efficiency was improved at the cost of the remaining objectives. The land planning schemes optimized by the particle swarm optimization (PSO) algorithm, the traditional GA, and the improved GA, respectively, were superior than the initial scheme, and the optimized scheme of the improved GA was more in line with the characteristics of the TOD mode than the traditional one and the PSO algorithm, and the fitness value was also higher. In conclusion, the GA can be used to optimize the planning design of land in rail transit areas under the TOD mode, and the optimization performance of the GA can be improved by means of coevolution.

10

Developing numerical equality to regional intensity-duration-frequency curves using evolutionary algorithms and multi-gene genetic programming

Citakoglu Hatice, Demir Vahdettin

Acta Geophysica

|

2023

|

Vol. 71, no 1

469--488

EN

This study aims to carry out regional intensity−duration−frequency (IDF) equality using the relationship with IDF obtained from point frequency analysis. Eleven empirical equations used in the literature for seven climate regions of Turkey were calibrated using particle swarm optimization (PSO) and genetic algorithm (GA) optimization techniques and the obtained results were compared. In addition, in this study, new regional IDF equations were obtained for each region utilizing Multi-Gene Genetic Programming (MGGP) method. Finally, Kruskal–Wallis (KW) test was applied to the IDF values obtained from the methods and the observed values. As a result of the study, it was observed that the coefficients of 11 empirical equations calibrated with PSO, and GA techniques were different from each other. The mean absolute error (MAE), root mean square error (RMSE), mean absolute relative error (MARE), coefficient of determination (R2 ), and Taylor diagram were used to evaluate the performances of PSO, GA, and MGGP techniques. According to the performance criteria, it has been determined that the IDF equations obtained by the MGGP method for the Eastern Anatolia, Aegean, Southeastern Anatolia, and Central Anatolia regions are more successful than the empirical equations calibrated with the PSO and GA method. The empirical IDF equations produced with PSO and the IDF equations acquired with MGGP have similar findings in the Mediterranean, Black Sea, and Marmara. In addition, the KW test results showed that the data of all models were from the same population.

11

Genetic algorithms in active vibration reduction problem

Grochowina Marcin, Tyburski Krzysztof

Vibrations in Physical Systems

|

2022

|

Vol. 33, nr 2

art. no. 2022219

EN

The design of active vibration reduction systems usually consists in selecting a control algorithm and determining the value of its settings. This article presents the results of research on the concept of using genetic algorithms to induce the settings of control systems. To test the concept, a simple pulse-excited flat bar model was selected. The vibrations were suppressed by the PID controller. Genetic algorithms with two types of crossover were tested - arithmetic and uniform. As a result, the settings for the PID controller were obtained, enabling effective reduction of vibrations in a short time.

12

Genetic algorithm application for optimizing traffic signal timing reflecting vehicle emission intensity

Hai Dinh Tuan, Manh Do Van, Nhat Nguyen Minh

Transport Problems

|

2022

|

T. 17, z. 1

5--16

EN

Urbanization has created continuous growth in transportation demand, leading to serious issues, including infrastructure overload, disrupted traffic flow, and associated vehicular emissions. As a result, resolving these problems has become one of the primary missions of governments worldwide. The optimization of the traffic signal timing system is considered a promising approach to overcoming the negative consequences of increasing vehicle volume. In metropolises, oversaturated intersections, where the traffic density and vehicle exhaust emission levels are significant, have been considered as the priority to target. Several scientists have attempted to design traffic lights with the most appropriate timing. However, the majority of previous studies have not formed a comprehensive evaluation of essential factors, especially regarding the appropriate weighting of vehicle emission parameters. By assessing the all-inclusive relationship of critical elements with an emphasis on vehicle exhaust emissions, a performance index model using a genetic algorithm (GA) is established in this paper, demonstrated by data from a case study in Taiwan.

13

Comparison of genetic algorithm and neural network approaches for the prognosis of mechanical idle running losses in agriculture tractor transmission

Ivanov Rosen, Ivanova Donka

Transport Problems

|

2022

|

T. 17, z. 3

51--59

EN

An experimental investigation of mechanical idle running losses in an agriculture tractor transmission was used to collect a wide range of data. The influence of the engine rotation speed, the number of switched-on gears, and the oil level in the transmission gearbox on the idle running losses was determined. Adequate regression models in cases of switched-on and switched-off PTO were received. A genetic algorithm was used to optimize mathematical models obtained using regression analysis. A feedforward artificial neural network was also developed to estimate the same experimental data for mechanical idle running losses in transmission. A back-propagation algorithm was used when training and testing the network. A comparison of the correlation coefficient, reduced chi-square, mean bias error, and root mean square error between the experimental data and fit values of the obtained models was made. It was concluded that the neural network represented the mechanical idle running losses in tractor transmission more accurately than other models.

14

Zastosowanie algorytmu genetycznego w procesie estymacji parametrów modeli przyrządów półprzewodnikowych

Bisewski Damian

Przegląd Elektrotechniczny

|

2022

|

R. 98, nr 9

103--106

PL

Praca dotyczy zastosowania algorytmu genetycznego w procesie estymacji wartości parametrów arbitralnie wybranego modelu tranzystora MOS. Przedstawiono budowę oraz zasadę działania algorytmu genetycznego. Pokazano wpływ wybranych parametrów sterujących działaniem algorytmu na obliczone wartości funkcji celu. Zaproponowano modyfikację operatora krzyżowania wpływającą na uzyskiwanie wyników obliczeń z większą dokładnością.

EN

The paper concerns the application of the genetic algorithm in the process of parameter estimation of an arbitrarily selected MOSFET model. The structure and the principle of operation of the genetic algorithm have been presented. The influence of selected parameters controlling the algorithm operation on the calculated values of the objective function has been shown. A modification of the crossing operator providing calculation results of greater accuracy has been proposed.

15

Genetic PID and Feedforward controllers for DC-DC chopper converter

Obeidat Mohammad A.

Przegląd Elektrotechniczny

|

2022

|

R. 98, nr 7

30--35

EN

DC voltage choppers such as buck, boost, and buck/boost are widely used in electrical power applications. Since these choppers are connected directly between DC source such as solar photovoltaic PV systems or batteries, a disturbance or dc source fluctuations may occur at the input of chopper circuits. Therefore, the control systems must be designed and developed in order to reduce such an increase or decrease in voltage. In this paper, two control strategies have been studied and analyzed to reduce system disturbance and minimize the error resulted from noise. The first strategy uses both feedback and feedforward controllers, in this strategy the controllers are designed based on linearization system. The second strategy uses genetic algorithm to tune the integrated proportional, integral, and differentiator PID feedback controller parameters directly for the nonlinear system. The results show that, the genetic PID controller has better performance than the Feedforward/Feedback controller. The mathematical model of the chopper-controlled system using both strategies and the simulation results are extracted using Matlab/Simulink 2018.

PL

Przerywacze napięcia stałego, takie jak buck, boost i buck/boost, są szeroko stosowane w zastosowaniach elektroenergetycznych. Ponieważ przerywacze te są połączone bezpośrednio między źródłami prądu stałego, takimi jak fotowoltaiczne systemy fotowoltaiczne lub akumulatory, na wejściu obwodów przerywacza mogą wystąpić zakłócenia lub wahania źródła prądu stałego. Dlatego też układy sterowania muszą być projektowane i rozwijane w celu ograniczenia takiego wzrostu lub spadku napięcia. W niniejszym artykule zbadano i przeanalizowano dwie strategie sterowania w celu zmniejszenia zakłóceń systemu i zminimalizowania błędu wynikającego z hałasu. Pierwsza strategia wykorzystuje zarówno regulatory sprzężenia zwrotnego, jak i sprzężenia do przodu, w tej strategii regulatory są projektowane w oparciu o system linearyzacji. Druga strategia wykorzystuje algorytm genetyczny do dostrojenia parametrów zintegrowanego regulatora proporcjonalnego, całkowego i różniczkowego ze sprzężeniem zwrotnym PID bezpośrednio dla systemu nieliniowego. Wyniki pokazują, że genetyczny regulator PID ma lepszą wydajność niż regulator sprzężenia zwrotnego/zwrotnego. Model matematyczny systemu sterowanego chopperem wykorzystujący obie strategie i wyniki symulacji są wyodrębniane za pomocą Matlab/Simulink 2018.

16

Automatic identification of malfunctions of large turbomachinery during transient states with genetic algorithm optimization

Barszcz Tomasz, Zabaryłło Mateusz

Metrology and Measurement Systems

|

2022

|

Vol. 29, nr 1

175--190

EN

Turbines and generators operating in the power generation industry are a major source of electrical energy worldwide. These are critical machines and their malfunctions should be detected in advance in order to avoid catastrophic failures and unplanned shutdowns. A maintenance strategy which enables to detect malfunctions at early stages of their existence plays a crucial role in facilities using such types of machinery. The best source of data applied for assessment of the technical condition are the transient data measured during start-ups and coast-downs. Most of the proposed methods using signal decomposition are applied to small machines with a rolling element bearing in steady-state operation with a shaft considered as a rigid body. The machines examined in the authors’ research operate above their first critical rotational speed interval and thus their shafts are considered to be flexible and are equipped with a hydrodynamic sliding bearing. Such an arrangement introduces significant complexity to the analysis of the machine behavior, and consequently, analyzing such data requires a highly skilled human expert. The main novelty proposed in the paper is the decomposition of transient vibration data into components responsible for particular failure modes. The method is automated and can be used for identification of turbogenerator malfunctions. Each parameter of a particular decomposed function has its physical representation and can help the maintenance staff to operate the machine properly. The parameters can also be used by the managing personnel to plan overhauls more precisely. The method has been validated on real-life data originating from a 200 MW class turbine. The real-life field data, along with the data generated by means of the commercial software utilized in GE’s engineering department for this particular class of machines, was used as the reference data set for an unbalanced response during the transients in question.

17

Scheduling of the Manufacturing Cell Work with the Use of a Genetic Algorithm on the Example of a Flexible Production System

Matuszny Marcin

Management and Production Engineering Review

|

2022

|

Vol. 13, No. 4

98--106

EN

The article is to present the application of genetic algorithm in production scheduling in a production company. In the research work the assumptions of the methodology were described and the operation of the proposed genetic algorithm was presented in details. Genetic algorithms are useful in complex large scale combinatorial optimisation tasks and in the engineering tasks with numerous limitations in the production engineering. Moreover, they are more reliable than the existing direct search algorithms. The research is focused on the effectivity improvement and on the methodology of scheduling of a manufacturing cell work. The genetic algorithm used in the work appeared to be robust and fast in finding accurate solutions. It was shown by experiment that using this method enables obtaining schedules suitable for a model. It gives a group of solutions that are at least as good as those created by the heuristic rules.

18

Genetic based algorithms to solving multi-quays berth allocation problem with setup time constraints

Sangsawang Chatnugrob, Longploypad Cholthida

LogForum

|

2022

|

Vol. 18, no. 4

505--515

EN

Background: This study focuses on efficient berth planning in multi-purpose terminal composed of multiple quays. A multi-quay berth offers infrastructure, equipment, and services for different types of cargo and vessels to meet the needs of users from various freight markets. Moreover, each berth from any quay can be dedicated for one or two different types of cargo and vessels. To improve port efficiency in terms of reducing the waiting time of ships, this study addresses the Multi-Quay Berth Allocation Problem (MQ-BAP), where discrete berthing layout is considered along with setup time constraints and practical constraints such as time windows and safety distances between ships. Sequence dependent setup times may arise due to the berth can convert from dedicated function to another function according to the variance of cargo demand. This problem was inspired by a real case of a multi-purpose port in Thailand. Methods: To solve the problem, we propose a mixed-integer programming model to find the optimal solutions for small instances. Furthermore, we adapted a metaheuristic solution approach based on Genetic algorithm (GA) to solving the MQ-BAP model in large-scale problem cases. Results: Numerical experiments are carried out on randomly generated instances for multi-purpose terminals to assess the effectiveness of the proposed model and the efficiency of the proposed algorithm. The results show that our proposed GA provides a near-optimal solution by average 4.77% from the optimal and show a higher efficiency over Particle swarm optimization (PSO) and current practice situation, which are first come first serve (FCFS) rule by 1.38% and 5.61%, respectively. Conclusions: We conclude that our proposed GA is an efficient algorithm for near-optimal MQ-BAP with setup time constraint at acceptable of computation time. The computational results reveal that the reliability of the metaheuristics to deal with large instances is very efficient in solving the problem considered.

19

Joint Optimization of Sum and Difference Patterns with a Common Weight Vector Using the Genetic Algorithm

Mohammed Jafar Ramadhan, Aljaf Duaa Alyas

Journal of Telecommunications and Information Technology

|

2022

|

nr 3

67--73

EN

A monopulse searching and tracking radar antenna array with a large number of radiating elements requires a simple and efficient design of the feeding network. In this paper, an effective and versatile method for jointly optimizing the sum and difference patterns using the genetic algorithm is proposed. Moreover, the array feeding network is simplified by attaching a single common weight to each of its elements. The optimal sum pattern with the desired constraints is first generated by independently optimizing amplitude weights of the array elements. The suboptimal difference pattern is then obtained by introducing a phase displacement π to half of the array elements under the condition of sharing some sided elements weights of the sum mode. The sharing percentage is controlled by the designer, such that the best performance can be met. The remaining uncommon weights of the difference mode represent the number of degrees of freedom which create a compromise difference pattern. Simulation results demonstrate the effectiveness of the proposed method in generating the optimal sum and suboptimal difference patterns characterized by independently, partially, and even fully common weight vectors.

20

Modeling Pollution Index Using Artificial Neural Network and Multiple Linear Regression Coupled with Genetic Algorithm

Abdulkareem Iman Ali, Abbas Abdulhussain A., Dawood Ammar Salman

Journal of Ecological Engineering

|

2022

|

Vol. 23, nr 3

236--250

EN

Shatt Al-Arab River in Basrah province, Iraq, was assessed by applying comprehensive pollution index (CPI) at fifteen sampling locations from 2011 to 2020, taking into consideration twelve physicochemical parameters which included pH, Tur., TDS, EC, TH, Na+, K+, Ca+2, Mg+2, Alk., SO4-2, and Cl-. The effectiveness of multiple linear regression (MLR) and artificial neural network (ANN) for predicting comprehensive pollution index was examined in this research. In order to determine the ideal values of the predictor parameters that lead to the lowest CPI value, the genetic algorithm coupled with multiple linear regression (GA-MLR) was used. A multi-layer feed-forward neural network with backpropagation algorithm was used in this study. The optimal ANN structure utilized in this research consisted of three layers: the input layer, one hidden layer, and one output layer. The predicted equation of the comprehensive pollution index was created using the regression technique and used as an objective function of the genetic algorithm. The minimum predicted comprehensive pollution index value recommended by the GA-MLR approach was 0.3777.