Evolutionary Algorithm Method Research Articles

A many-objective evolutionary algorithm based on reference vector guided selection and two diversity and convergence enhancement strategies

Achieving the balance between convergence and diversity is a key and challenging issue in many-objective optimization. Reference vector guided selection is an exemplary method for decomposition-based many-objective evolutionary algorithms (MaOEAs). However, there are some problems with it such as insufficient number of obtained solutions and inefficient convergence evaluation metric. Aiming at solving or alleviating these problems, this paper proposes a many-objective evolutionary algorithm based on reference vector guided selection and two diversity and convergence enhancement strategies. The proposed algorithm introduces two new strategies namely adaptive sparse region detection and convergence-only selection. The former is to adaptively detect sparse regions of current elite population, while the latter is to prevent the elimination of solutions with prominent convergence performance. Together with a newly proposed elite retention strategy, these two strategies can achieve diversity and convergence enhancement on the basis on reference vector guided selection. Besides, A new selection criterion for reference vector guided selection is proposed to better measure the convergence of solutions in high dimensionality. Experimental results on widely used test problem suites up to 15 objectives indicate that the proposed algorithm is highly competitive in comparison with seven state-of-the-art MaOEAs.

A hybrid search space reduction algorithm and Newton–Raphson based selective harmonic elimination for an asymmetric cascade H-bridge multi-level inverter

Abstract In this paper, a selective harmonic elimination (SHE) method for an asymmetric cascade H-bridge multi-level inverter (ACHB-MLI) has been performed by using hybrid search space reduction and Newton–Raphson (SSR-NR) algorithm. Three H-bridges are cascaded in a single phase configuration, while each of the bridge is fed with V dc , 3V dc and 6V dc respectively. This particular asymmetric combination of dc sources will produce 10 possible switching combinations and 21-levels in the output voltage ranging from −10V dc to +10V dc . Hence, there will be 10 SHE equations corresponding to fundamental and harmonics, and the solution will be the 10 switching angles which satisfy those SHE equations. While traditional gradient-based techniques like the Newton–Raphson method can yield the global optimal solution for the SHE equations, they necessitate an initial guess. Conversely, meta-heuristic methods often face challenges related to sub-optimal convergence. To address these issues, a novel the hybrid SSR-NR algorithm has been introduced in this paper. This method combines stochastic and deterministic elements, offering a promising solution for solving the SHE equations. In the first step, the evolutionary SSR algorithm will be used and the output of the SSR algorithm is given to the NR method as an initial guess to converge to the final solution, this way the proposed hybrid SSR-NR algorithm is able to address the drawbacks of both the algorithms such as initial guess requirement for NR method and sub-optimal convergence characteristics of evolutionary algorithms. The simulations were carried out in Matlab/simulink environment for a 21-level ACHB-MLI. The results obtained showcase the efficacy of the proposed SSR-NR algorithm in solving the SHE equations. For demonstrating the effectiveness of SSR algorithm the results were also compared with other algorithms such as particle swarm optimization and grey wolf optimizer as well. Further, the experimental results on a laboratory prototype using dSPACE DS1104 R&D controller board were also presented in the paper, to validate the proposed methodology.

Efficient Selection Methods in Evolutionary Algorithms

Evolutionary algorithms mimic some elements of the theory of evolution. The survival of individuals and the possibility of producing offspring play a huge role in the process of natural evolution. This process is called a natural selection. This mechanism is responsible for eliminating poor population members and gives the possibility of development for good ones. The evolutionary algorithm - an instance of evolution in the computer environment also requires a selection method, a computer version of natural selection. Widely used standard selection methods applied in evolutionary algorithms are usually derived from nature and prefer competition, randomness and some kind of ``fight'' among individuals. But computer environment is quite different from nature. Computer populations of individuals are usually small, they easily suffer from a premature convergence to local extremes. To avoid this drawback, computer selection methods must have different features than natural selection. In the computer selection methods randomness, fight and competition should be controlled or influenced to operate to the desired extent. Several new methods of individual selection are proposed in this work: several kinds of mixed selection, an interval selection and a taboo selection. Also advantages of passing them into the evolutionary algorithm are shown, using examples based on searching for the maximum α-clique problem and traditional TSP in comparison with traditionally considered as very efficient tournament selection, considered ineffective proportional (roulette) selection and similar classical methods.

A multi-granularity clustering based evolutionary algorithm for large-scale sparse multi-objective optimization

Sparse multi-objective optimization problems (SMOPs) frequently exist in a variety of disciplines such as machine learning, economy, and signal processing. Evolutionary algorithms have demonstrated their proficiency in optimizing complex problems in recent years, although their performance often deteriorates significantly on large-scale SMOPs. In an effort to accelerate the convergence, this paper suggests a multi-granularity variable clustering method for evolutionary algorithms. This method estimates the sparse distribution of decision variables at each generation and partitions them into a varying number of layers, each with a distinct probability of being zero. These clustering outcomes inspire the development of a crossover operator and a mutation operator, which prove adept at efficiently generating sparse solutions. Experimental evaluations on both benchmark and real-world SMOPs confirm that an evolutionary algorithm incorporating the new crossover operator and mutation operator converges more rapidly than its state-of-the-art counterparts.

Transfer learning based covariance matrix adaptation for evolutionary many-objective optimization

Evolutionary Algorithms (EA) have proven successful in solving Multi- and Many-objective Optimization Problems (MOPs/MaOPs) in numerous application areas. However, the various linkages among the decision variables may have posed a challenge for traditional evolutionary algorithms for solving such problems. In this paper, a transfer learning based covariance matrix adaptation algorithm, shortened as TL-M2M-CMA, is proposed to handle MOPs/MaOPs with complex decision variable linkage. In TL-M2M-CMA, the search population is first decomposed into a set of subpopulations and each subpopulation corresponds to a specific part of the Pareto-optimal Front (PF) of the MOP/MaOP. During the search, a covariance matrix adaptation process is utilized to learn the linkage among the decision variables, and then the covariance matrix is incorporated into the crossover and mutation operators for better efficiency. To facilitate performance comparison, a set of scalable MOP test instances with various linkage complexities is constructed for experimental studies. The performance of the proposed TL-M2M-CMA on these constructed instances is verified by comparing it to the performance of five state-of-the-art EA methods.

Evolutionary Bi-level Optimization via Multi-objective Transformation-Based Lower Level Search

Nested evolutionary algorithms (EAs) have been regarded as very promising tools for bi-level optimization. Due to the nested structure, the upper level population evaluation requires a set of complete lower level optimizations, thereby reducing the efficiency and practicability of EA methods. In this paper, a multi-objective transformation-based evolutionary algorithm (MOTEA) is proposed to perform multiple lower level optimizations in a parallel and collaborative manner. Specifically, the corresponding multiple lower level optimizations for each generation of the upper level population evaluation are transformed into locating a set of Pareto optimal solutions of a constructed multi-objective optimization problem. By utilizing the built-in implicit parallelism of evolutionary multi-objective optimization, multiple lower level problems can thus be optimized in parallel. Within one multi-objective search population, the collaboration among the parallel lower level optimization can be realized by exploiting and utilizing the implicit similarities among them for better efficiency. The effectiveness and efficiency of the proposed MOTEA are verified by comparing it with four state-of-the-art evolutionary bi-level optimization algorithms on two sets of popular bi-level optimization benchmark test problems and three application problems.

An Improved Evolutionary Algorithm in Formulating a Diet for Grouper

This paper reveals the high demand of fish products in many countries, which subsequently highlighted the high demand of grouper fish species for human consumption. This high demand leads to the insufficient supply of wild ocean grouper fish in the market, thus justifying the need for farmed or cultured grouper fish. Basically, in grouper fish farming, large amounts of trash fish are needed as the feed for grouper fish, which is the carnivorous type of fish. However, since the cost of trash fish is too high, searching for alternative ingredients for the feed through modelling of feed formulation is an option for reducing or minimizing the farming cost. This led to the search for methods in giving the best combination of feedstuff ingredients with appropriate nutrients in formulating the feed. One prospective method is the Evolutionary Algorithm (EA) that has been applied in solving similar problems of diet formulation for several types of animals including livestock, poultry and shrimp. Hence, in this paper, an improved EA method known as the SR-SD-EA is proposed highlighting three important EA operators, which are initialization, selection and mutation. A semi random initialization operator is introduced to filter some important constraints thus increase the chances of obtaining feasible formulations or solutions. Subsequently, the novel selection operator embeds the concept of standard deviation in the SR-SD-EA as part of the function in minimizing the total cost of the formulated grouper fish feed. Eventually, the enhanced boundary-based mutation is also introduced in the algorithm to ensure the crucial constraint of the ingredients’ total weight must be met. The overall structure of the SR-SD-EA is presented as a framework, where the three methodological contributions are embedded. The preliminary findings of SR-SD-EA show that the obtained cost computed based on the Best-So-Far feed formulation as the solution is comparable, while all the crucial constraints are fulfilled.

Neural Network Ensemble With Evolutionary Algorithm for Highly Imbalanced Classification

Imbalanced data is a major challenge in classification tasks. Most classification algorithms tend to be biased toward the samples in the majority class but fail to classify the samples in the minority class. Recently, ensemble learning, as a promising method, has been rapidly developed in solving highly imbalanced classification. However, the design of the base classifier for the ensemble is still an open question because the optimization problem of the base classifier is gradientless. In this study, the evolutionary algorithm (EA) technique is adopted to solve a wide range of optimization design problems in highly imbalanced classification without gradient information. A novel EA-based classifier optimization design method is proposed to optimize the design of multiple base classifiers automatically for the ensemble. In particular, an EA method with a neural network (NN) as the base classifier termed NN ensemble with EA (NNEAE) is developed for highly imbalanced classification. To verify the performance of NNEAE, extensive experiments are designed for testing. Results illustrate that NNEAE outperforms other compared methods.

Effective Moving Object Tracking Using Adaptive Background Subtraction with Advanced Probability Evolutionary Algorithm

Tracking moving objects is a very difficult task for many computers. The task of tracking is a moving object. There are many virtual applications such as video surveillance and object recognition. Many ideas have been suggested for moving objects tracking and detection. One of the main problems solving tracking is finding identical objects in different frames. In this article, this application mainly focuses on the detection and tracking of moving objects using adaptive Background subtraction with adaptive probability evolutionary algorithm (AREA) method. Normal probability evolutionary algorithm will track only human motion but our algorithm tracks all the moving objects very accurately. Our proposed system shows the tracked objects with the red mark border.

Multi-Objective Multi-Satellite Imaging Mission Planning Algorithm for Regional Mapping Based on Deep Reinforcement Learning

Satellite imaging mission planning is used to optimize satellites to obtain target images efficiently. Many evolutionary algorithms (EAs) have been proposed for satellite mission planning. EAs typically require evolutionary parameters, such as the crossover and mutation rates. The performance of EAs is considerably affected by parameter setting. However, most parameter configuration methods of the current EAs are artificially set and lack the overall consideration of multiple parameters. Thus, parameter configuration becomes suboptimal and EAs cannot be effectively utilized. To obtain satisfactory optimization results, the EA comp ensates by extending the evolutionary generation or improving the evolutionary strategy, but it significantly increases the computational consumption. In this study, a multi-objective learning evolutionary algorithm (MOLEA) was proposed to solve the optimal configuration problem of multiple evolutionary parameters and used to solve effective imaging satellite task planning for region mapping. In the MOLEA, population state encoding provided comprehensive population information on the configuration of evolutionary parameters. The evolutionary parameters of each generation were configured autonomously through deep reinforcement learning (DRL), enabling each generation of parameters to gain the best evolutionary benefits for future evolution. Furthermore, the HV of the multi-objective evolutionary algorithm (MOEA) was used to guide reinforcement learning. The superiority of the proposed MOLEA was verified by comparing the optimization performance, stability, and running time of the MOLEA with existing multi-objective optimization algorithms by using four satellites to image two regions of Hubei and Congo (K). The experimental results showed that the optimization performance of the MOLEA was significantly improved, and better imaging satellite task planning solutions were obtained.

Sharing hydrogen storage capacity planning for multi-microgrid investors with limited rationality: A differential evolution game approach

Bottom on the ripple of the multiplication of sharing economy, hydrogen energy storage (HES) shared calls for novel solutions to ameliorate the cleanness and economy of micro-grids under the dual carbon target in China. The traditional investment method may become impracticable due to conflicting interests, insufficient utilization, and poor interoperability. The practical and prospective design of HES relies on innovation at the joint investment mechanism under incomplete synchronized and the interested equilibrium under limited rationality. This paper studies the capacity planning method for HES by multiple micro-grids from the sharing perspective. The evolution game model is utilized for conflict of interest, which exactly accounts for limited rationality. The dynamic competitive as well as time-varying demand for HES among multiple micro-grids are established by a small-word complex network theory. The dual dynamic evolutionary game of investment decision in multi-micro-grid and capacity planning for sharing HES is realized by a proposed method of differential evolutionary game algorithm based on experience-weighted attraction learning. The numerical simulation is carried out to verify the effectiveness of the proposed method and interpret the development path for sharing HES. The result shows that, the sizing method with limited rationality reduces the cost by 4.75% compared with the traditional game under complete rationality and 9.15% compared with the multi-objective optimization. The willingness of the park to participate in joint investment in shared HES is influenced by the long-term regulatory demand for energy storage and the decisions of other parks in the neighbors.

Function expansion based topology optimization of NRD guide device using hybrid method of harmony search and gradient method

In recent years, optimization technique has attracted a lot of attention to develop high performance electromagnetic devices and various kinds of design strategies have been reported. Among them, topology optimization method has potential to develop innovative devices. In this paper, the function expansion based topology optimization method, which has been developed for design of photonic devices, is extended to be applied to design of nonradiative dielectric waveguide (NRD guide) devices. In the optimal design of NRD guide devices, solution search sometimes tends to fall into local minimum compared with the design of photonic devices. Therefore, we develop a novel hybrid optimization method combining harmony search which is one of the evolutionary algorithm and gradient method using adjoint variable method for robust and efficient design of NRD guide devices.

Mutation testing in test suite generation using separate bacterial memetic evolutionary algorithm in IoT

The security of the Internet of Things (IoT) is becoming more important as it is used more frequently and widely in our daily lives. Better security analysis tools are necessary for IoT applications in order to assure information security. Therefore, a framework for evaluating taint-flow analysis tools in the context of IoT applications is presented. In various strategies, the mutation testing has been found in detecting the faults in the test case. Fault-based checking is considered as mutation testing because the flaws are identified actively in the code to ensure that the programmed functions correctly. If the flaws in the programmed are not identified, it indicates that program requires debugging for effective functioning. Testing has made a significant contribution to a set of approaches, tools, improvements, and perfect results; however, it is an expensive test leading to increased bug-finding practices. Automated tools and minimization methods were launched to decrease the expenses of mutation testing. The Separate Bacterial Memetic Evolutionary Algorithm (SBMEA) was examined in this study effort as a worldwide seeker. The suggested algorithm strengthens the reduction cost of mutation testing in all possible ways. It's able to achieve the minimum test case and mutants score for computational and cost, here the enhanced of Memetic algorithm were proposed. The present system improved the surrogate enhancement by lowering the expense and eliminating mutants in the snippets. The Separate Bacterial Memetic Evolutionary Algorithm (SBMEA) method of Memetic algorithm was performed the high mutation score in all testing of java snippets. Compared to proposed algorithm the execution time is saved in 6.95 s better than existing algorithm. It consists of evolution of strategy and program solver and increase the computational efficiency of the solution method. Here the Separate Bacterial Memetic Evolutionary Algorithm (SBMEA) method of Memetic algorithm is capable of lowering computational costs and eliminating all mutants, its best for reduction cost and for mutation testing.

Production Scheduling Analysis To Minimize Inject–Blow Production Makespan on PT.XYZ Using Differential Evolution Algorithm Method

This study discusses the problem of job-shop scheduling at one of the well-known mineral water companies in Indonesia. This company produces a number of products in several types with different routes from each other. However, in reality, the effectiveness of production lies in the range of 54.8%–69.2%, so it is still far from running time production where the expected target is 80% for inject and 90% for blow. Where production scheduling is a complex problem so it takes the right method to get the optimal solution for these problems. The research method used is the differential evolution algorithm method, where this method aims to minimize the makespan value (total completion time of all jobs) on the job-shop production route so as to obtain optimal scheduling results and improve the existing system. Scheduling obtained through the differential evolution algorithm method produces a makespan value of x minutes, while the company's schedule has a makespan value of y minutes. So that the proposed scheduling results in a decrease of x% compared to company scheduling.

Performance analysis and multi-objective optimization of bionic dendritic furcal energy-absorbing structures for trains

Rail vehicles offer the characteristics of large carrying capacity and high running speed. Train accidents usually result in substantial personal casualties and economic losses. Energy-absorbing structures can efficiently and quickly absorb and disperse the energy generated during collisions. In this paper, the performances with different cross-sections and partings of the bionic dendritic furcal structure (BDFS) inspired by the furcal branch and bifurcation structure of neurons are analyzed and compared by numerical simulation. The simplified super folding element theory and experiments verify that the finite element models are effective. A hybrid multi-stage optimization decision system, i.e., multi-criteria decision-making (MCDM)- multi-objective optimization (MOP)- MCDM, which integrates the VlseKriterijumska Optimizacija I Kompromisno Resenje (VIKOR) method, the multiple objective particle swarm optimization-crowding distance (MOPSO-CD) evolutionary algorithm and repetitive VIKOR method, is proposed to find the optimal structure and obtain the optimal parameter alternative of the structure. Comparisons and analyses of experiments and simulations are carried out to verify the effectiveness of the proposed method. The results show that BDFS with six furcal ribs and 3rd-order parting display the best comprehensive crashworthiness under the working condition of 10 m/s specified in EN15227 standard. When the single target requirement is considered, specific energy absorption (SEA) can be increased by 16.18%, peak crushing force (PCF) can be decreased by 67.07%, and undulation of the load-carrying capacity (ULC) can be decreased by 14.02%. This study provides an effective tool for the analysis and optimization of energy-absorbing structures.

Application of inclusive multiple model for the prediction of saffron water footprint

Applying new approaches in the management of water resources is a vital issue, especially in arid and semi-arid regions. The water footprint is a key index in water management. Therefore, it is necessary to predict its changes for future durations. The soft computing model is one of the most widely used models in predicting and estimating agroclimatic variables. The purpose of this study is to predict the green and blue water footprints of saffron product using the soft computing model. In order to select the most effective variables in prediction water footprints, the individual input was eliminated one by one and the effect of each on the residual mean square error (RMSE) was measured. In the first stage, the Group Method of Data Handling (GMDH) and evolutionary algorithms have been applied. In the next stage, the output of individual models was incorporated into the Inclusive Multiple Model (IMM) as the input variables in order to predict the blue and green water footprints of saffron product in three homogenous agroclimatic regions. Finally, the uncertainty of the model caused by the input and parameters was evaluated. The contributions of this research are introducing optimized GMDH and new ensemble models for predicting BWF, and GWF, uncertainty analysis and investigating effective inputs on the GWF and BWF. The results indicated that the most important variables affecting green and blue water footprints are plant transpiration, evapotranspiration, and yield, since removing these variables significantly increased the RMSE (range=11–25). Among the GMDH models, the best performance belonged to NMRA (Naked Mole Ranked Algorithm) due to the fast convergence and high accuracy of the outputs. In this regard, the IMM has a better performance (FSD=0.76, NSE=0.95, MAE) = 8, PBIAS= 8) than the alternatives due to applying the outputs of several individual models and the lowest uncertainty based on the parameters and inputs of the model (p = 0.98, r = 0.08).

Идентификация угроз средствами обеспечения безопасности и выбор методами машинного обучения

The following tasks are considered: identifying the sources and nature of threats using the methods of naive Bayes and Bayesian trust networks to ensure security in the development, production and operation of on-board systems; selection of the optimal set of operations and measures of an organizational and technological na-ture, taking into account the expected costs or maximum security with restrictions on allowable costs. It is proposed to solve them by methods of machine learning and evolutionary optimization algorithms.

A Weight Vector Adjustment Method for Decomposition-Based Multi-Objective Evolutionary Algorithms

Multi-objective evolutionary algorithm based on decomposition (MOEA/D) is effective to solve most multi-objective optimization problems (MOPs) in the past 20 years. However, the algorithm MOEA/D with constant weight vectors has bad performance in solving several MOPs with discontinuous Pareto front (PF). This paper analyses the limitations of the constant weight vectors in MOEA/D and explains the necessity of adjusting the weight vectors in the processing. This paper proposes a weight vector adjustment method for MOEA/D (MOEA/D-WVA). It deletes the weight vectors which have bad search direction, and adds some new weight vectors in the processing. Experimental studies are conducted on several MOPs with discontinuous PF to compare the MOEA/D-WVA with other state-of-the-art multi-objective optimization algorithms in solving those MOPs with complex PF. The results show MOEA/D-WVA performs better than other algorithms on those MOPs with discontinuous PF.

Design of grayscale digital light processing 3D printing block by machine learning and evolutionary algorithm

Grayscale digital light processing (DLP) 3D printing method, using grayscale light patterns to adjust the microscale material properties, is a revolutionary technology for future advanced manufacturing. However, the design and optimization method of the grayscale distribution remains elusive. In this paper, we propose an optimization method for grayscale DLP printed rectangular blocks based on a machine learning and evolutionary algorithm. We use an automated finite element model-based evaluation to predict the deformation shapes with arbitrary grayscale distribution, considering the nonlinear mechanical performances. A machine learning model based on recurrent neural networks is trained using the database formed by the finite element method model to enable fast and precise prediction of the deformed shapes. Optimal designs of the deformed shapes are efficiently realized via integrating the machine learning model and evolution algorithm method. Grayscale distributions are optimized to form desired deformations and validated by experiments. This work paves the way for designing and optimizing grayscale digital light processing 3D printed structures.

Multi-objective optimization through artificial intelligence for designing of an Agave angustifolia leaf shredder

A neural network and a genetic algorithm were used in a hybrid method to get the optimal design parameters of an Agave angustifolia Haw. green leaf shredder. First, a prototype of an experimental machine was built using the design parameters recommended by the literature and calculated using linear equations. Then, the shredder prototype was subjected to experiments. The defibration data with different blade adjustments were obtained with experimental values. The data was configured and trained with an artificial neural network to establish a correlation between the defibration quality and the design parameters. The multi-objective optimization method based on genetic algorithms determined the optimal design parameters of the shredder’s functional mechanical elements. The best point was obtained from the least number of broken fibers (2.83%) and the most waste (73.15%). The method used proved suitable to optimize the design parameters; this was based on actual data obtained by experiments performed with the prototype and then modeled through artificial intelligence methods such as neural networks to determine an optimal solution using evolutionary genetic algorithm methods.