Search Agents Research Articles

Extracting tactics learned from self-play in general games

Local, spatial state-action features can be used to effectively train linear policies from self-play in a wide variety of board games. Such policies can play games directly, or be used to bias tree search agents. However, the resulting feature sets can be large, with a significant amount of overlap and redundancies between features. This is a problem for two reasons. Firstly, large feature sets can be computationally expensive, which reduces the playing strength of agents based on them. Secondly, redundancies and correlations between features impair the ability for humans to analyse, interpret, or understand tactics learned by the policies. We look towards decision trees for their ability to perform feature selection, and serve as interpretable models. Previous work on distilling policies into decision trees uses states as inputs, and distributions over the complete action space as outputs. In contrast, we propose and evaluate a variety of decision tree types, which take state-action pairs as inputs, and provide various different types of outputs on a per-action basis. An empirical evaluation over 43 different board games is presented, and two of those games are used as case studies where we attempt to interpret the discovered features.

Epileptic Seizure Prediction Based on Hybrid Seek Optimization Tuned Ensemble Classifier Using EEG Signals.

Visual analysis of an electroencephalogram (EEG) by medical professionals is highly time-consuming and the information is difficult to process. To overcome these limitations, several automated seizure detection strategies have been introduced by combining signal processing and machine learning. This paper proposes a hybrid optimization-controlled ensemble classifier comprising the AdaBoost classifier, random forest (RF) classifier, and the decision tree (DT) classifier for the automatic analysis of an EEG signal dataset to predict an epileptic seizure. The EEG signal is pre-processed initially to make it suitable for feature selection. The feature selection process receives the alpha, beta, delta, theta, and gamma wave data from the EEG, where the significant features, such as statistical features, wavelet features, and entropy-based features, are extracted by the proposed hybrid seek optimization algorithm. These extracted features are fed forward to the proposed ensemble classifier that produces the predicted output. By the combination of corvid and gregarious search agent characteristics, the proposed hybrid seek optimization technique has been developed, and is used to evaluate the fusion parameters of the ensemble classifier. The suggested technique's accuracy, sensitivity, and specificity are determined to be 96.6120%, 94.6736%, and 91.3684%, respectively, for the CHB-MIT database. This demonstrates the effectiveness of the suggested technique for early seizure prediction. The accuracy, sensitivity, and specificity of the proposed technique are 95.3090%, 93.1766%, and 90.0654%, respectively, for the Siena Scalp database, again demonstrating its efficacy in the early seizure prediction process.

An Innovative Hunter-Prey-Based Optimization for Electrically Based Single-, Double-, and Triple-Diode Models of Solar Photovoltaic Systems

The derivation of PV model parameters is crucial for the optimization, control, and simulation of PV systems. Although many parameter extraction algorithms have been developed to address this issue, they might have some limitations. This work presents an efficient hybrid optimization approach for reliably and effectively extracting PV parameters based on the hunter–prey optimizer (HPO) technique. The proposed HPO technique is a new population-based optimizer inspired by the behavior of prey and predator animals. In the proposed HPO mechanism, the predator attacks the prey that leaves the prey population. Accordingly, the position of a hunter is adjusted toward this distant prey, while the position of the prey is adjusted towards a secure place. The search agent’s position, which represents the best fitness function value, is considered a secure place. The proposed HPO technique worked as suggested when parameters are extracted from several PV models, including single-, double-, and triple-diode models. Moreover, a statistical error analysis was used to demonstrate the superiority of the proposed method. The proposed HPO technique outperformed other recently reported techniques in terms of convergence speed, dependability, and accuracy, according to simulation data.

Diversity-Based Evolutionary Population Dynamics: A New Operator for Grey Wolf Optimizer

Evolutionary Population Dynamics (EPD) refers to eliminating poor individuals in nature, which is the opposite of survival of the fittest. Although this method can improve the median of the whole population of the meta-heuristic algorithms, it suffers from poor exploration capability to handle high-dimensional problems. This paper proposes a novel EPD operator to improve the search process. In other words, as the primary EPD mainly improves the fitness of the worst individuals in the population, and hence we name it the Fitness-Based EPD (FB-EPD), our proposed EPD mainly improves the diversity of the best individuals, and hence we name it the Diversity-Based EPD (DB-EPD). The proposed method is applied to the Grey Wolf Optimizer (GWO) and named DB-GWO-EPD. In this algorithm, the three most diversified individuals are first identified at each iteration, and then half of the best-fitted individuals are forced to be eliminated and repositioned around these diversified agents with equal probability. This process can free the merged best individuals located in a closed populated region and transfer them to the diversified and, thus, less-densely populated regions in the search space. This approach is frequently employed to make the search agents explore the whole search space. The proposed DB-GWO-EPD is tested on 13 high-dimensional and shifted classical benchmark functions as well as 29 test problems included in the CEC2017 test suite, and four constrained engineering problems. The results obtained by the proposal upon implemented on the classical test problems are compared to GWO, FB-GWO-EPD, and four other popular and newly proposed optimization algorithms, including Aquila Optimizer (AO), Flow Direction Algorithm (FDA), Arithmetic Optimization Algorithm (AOA), and Gradient-based Optimizer (GBO). The experiments demonstrate the significant superiority of the proposed algorithm when applied to a majority of the test functions, recommending the application of the proposed EPD operator to any other meta-heuristic whenever decided to ameliorate their performance.

Underwater Image Classification Algorithm Based on Convolutional Neural Network and Optimized Extreme Learning Machine

In order to deal with the target recognition in the complex underwater environment, we carried out experimental research. This includes filtering noise in the feature extraction stage of underwater images rich in noise, or with complex backgrounds, and improving the accuracy of target classification in the recognition process. This paper discusses our contribution to improving the accuracy of underwater target classification. This paper proposes an underwater target classification algorithm based on the improved flow direction algorithm (FDA) and search agent strategy, which can simultaneously optimize the weight parameters, bias parameters, and super parameters of the extreme learning machine (ELM). As a new underwater target classifier, it replaces the full connection layer in the traditional classification network to build a classification network. In the first stage of the network, the DenseNet201 network pre-trained by ImageNet is used to extract features and reduce dimensions of underwater images. In the second stage, the optimized ELM classifier is trained and predicted. In order to weaken the uncertainty caused by the random input weight and offset of the introduced ELM, the fuzzy logic, chaos initialization, and multi population strategy-based flow direction algorithm (FCMFDA) is used to adjust the input weight and offset of the ELM and optimize the super parameters with the search agent strategy at the same time. We tested and verified the FCMFDA-ELM classifier on Fish4Knowledge and underwater robot professional competition 2018 (URPC 2018) datasets, and achieved 99.4% and 97.5% accuracy, respectively. The experimental analysis shows that the FCMFDA-ELM underwater image classifier proposed in this paper has a greater improvement in classification accuracy, stronger stability, and faster convergence. Finally, it can be embedded in the recognition process of underwater targets to improve the recognition performance and efficiency.

Quantum-Inspired Distributed Memetic Algorithm

This paper proposed a novel distributed memetic evolutionary model, where four modules distributed exploration, intensified exploitation, knowledge transfer, and evolutionary restart are coevolved to maximize their strengths and achieve superior global optimality. Distributed exploration evolves three independent populations by heterogenous operators. Intensified exploitation evolves an external elite archive in parallel with exploration to balance global and local searches. Knowledge transfer is based on a point-ring communication topology to share successful experiences among distinct search agents. Evolutionary restart adopts an adaptive perturbation strategy to control search diversity reasonably. Quantum computation is a newly emerging technique, which has powerful computing power and parallelized ability. Therefore, this paper further fuses quantum mechanisms into the proposed evolutionary model to build a new evolutionary algorithm, referred to as quantum-inspired distributed memetic algorithm (QDMA). In QDMA, individuals are represented by the quantum characteristics and evolved by the quantum-inspired evolutionary optimizers in the quantum hyperspace. The QDMA integrates the superiorities of distributed, memetic, and quantum evolution. Computational experiments are carried out to evaluate the superior performance of QDMA. The results demonstrate the effectiveness of special designs and show that QDMA has greater superiority compared to the compared state-of-the-art algorithms based on Wilcoxon's rank-sum test. The superiority is attributed not only to good cooperative coevolution of distributed memetic evolutionary model, but also to superior designs of each special component.

A comparative study between deterministic and two meta-heuristic algorithms for solar PV MPPT control under partial shading conditions

• Under partial shading conditions, adequate tuning of firefly and particle swarm optimization maximum power point tracking algorithms converge to the global maximum power point irrespectively of the starting voltage value. • Under partial shading conditions, there is no guarantee that perturb and observe algorithm will converge to the global maximum power point even for different starting voltage values. • The sensitivities of the parameters of firefly algorithm in improving the performance of the algorithm are investigated to find the best possible control parameters. • Firefly algorithm based maximum power point tracking performed slightly well than particle swarm optimization based maximum power point tracking for most irradiance patterns investigated in this paper. During partial shading conditions (PSC), the PV modules in a solar PV array become reverse biased and act as a load, resulting in hotspot issues. This can significantly decrease the efficiency of the solar PV. The general way to deal with PSC is to connect bypass diodes across the non-shaded PV module. However, this will alter the uniform characteristics of the PV array, resulting in multiple power peaks i.e., a multi modal landscape. The conventional gradient-based Perturb and Observe (PnO) algorithm generally used for maximum power point tracking (MPPT) in solar PV is ineffective in finding the maximum power during PSC because it is prone to converge to local optimum due to its nature of searching in a multimodal landscape. However, in this article an attempt is made to show how the PnO behaves in a multimodal landscape considering different starting points. Robust stochastic algorithms that are based on a population of search agents are used to guarantee convergence to the global MPP. In this paper, the maximum power point under PSC was tracked using two meta-heuristic algorithms namely, particle swarm optimization (PSO) and the firefly algorithm (FA). The performances of these algorithms in tracking the maximum power point are evaluated. The efficiency, standard deviation (STD), and root mean square error (RMSE) of the PSO and FA algorithms are compared to those of the PnO algorithm. The PSO was found to have the lowest RMSE, and the FA had the lowest STD. The efficiency of the PSO and FA were relatively the same. Simulation results show that PSO and FA-based MPPT algorithms can efficiently track the global maximum power point (GMPP) irrespectively of the starting points. On the other hand, PnO is shown to be unreliable under PSC even with different starting points.

Web page prediction using adaptive deer hunting with chicken swarm optimization based neural network model

The world wide web acts as the dominant tool for data transmission due to access such as data retrieving and data transactions. The retrieval of data from the web is a complex procedure due to the large volume of web domain. The basic uses of the websites are described through web usage mining, which mines the weblog records to identify the pattern of accessing the web pages through the user. The web page prediction assists the web users in finding the plot and obtains the information as to their requirements. Several effective algorithms have been developed to mine association rules that make the predictive model more appropriate for web prediction. They can be commonly revised to ensure the changing feature of web access patterns. The Apriori algorithm involves extracting the recurrent itemset and interrelation rule that learns the relational data is commonly utilized for web page prediction. The Apriori algorithm remains the standard model for deriving the patterns and rules from the datasets in co-operative rule extraction. The Apriori algorithm thus generates large mines associated rules for web page prediction. Hence, to select the best rule, the proposed deer hunting rooster-based chicken swarm optimization algorithm is used by integrating the cockerel search agents’ dominating social search creatures’ hunting habits and their traits of looking for food. Further, the neural network (NN) is employed in this research for the prediction of web pages with minimum error. The trained NN is a technique of unsupervised learning that analyzes a dataset of input to produce the desired result, in which the effectiveness of the NN is enhanced by optimal tuning of weight by the adaptive deer hunting rooster-based chicken swarm optimization algorithm. The experimental analysis illustrates that the proposed adaptive deer hunting rooster-based chicken swarm optimization frameworks inherit lower error measures such as mean deviation = 139.89 and symmetric mean absolute percentage error[Formula: see text]0.45579 for the FIFA dataset. The proposed web page prediction models’ L2 norm and infinity norm are 58.017 and 14, respectively, for the MSNBC_SPMF dataset.

IEGQO-AOA: Information-Exchanged Gaussian Arithmetic Optimization Algorithm with Quasi-opposition learning

Arithmetic optimization algorithm (AOA) is a math optimizer proposed to solve optimization challenges. Its capability to find the global solution comes from the behavior of four arithmetic operators: multiplication, division, subtraction and addition. Local minima stagnation and sluggish convergence are the major concerns of AOA. To handle these issues, three effective modifications are proposed. Information exchange is introduced among the search agents first. Then, promising solutions around the best and current solutions are visited by a plausible way based on the Gaussian distribution. Finally, quasi-opposition of the best solution is obtained to have a higher chance of approaching the global solution. The proposed approach is named as Information-Exchanged Gaussian AOA with Quasi-Opposition learning (IEGQO-AOA). 23 standard benchmark functions, 10 CEC2020 test functions and 1 real-life engineering design problem are solved by the proposed IEGQO-AOA and its competing peers such as the original and modified versions of AOA, dwarf mongoose optimization, reptile search algorithm, aquila optimizer, bat algorithm, sine cosine algorithm, original and enhanced version of salp swarm algorithm, dragonfly search algorithm, LSHADE-EpSin, stochastic fractal search, improved jaya and moth-flame optimization, perturbed stochastic fractal search and nelder-mead simplex orthogonal learning moth-flame optimization algorithm. Comparative results based on the statistical tests ratify the potential of IEGQO-AOA in solving problems concerning accuracy and convergence without compromising on the algorithm’s simplicity much.

Raven finch optimized deep convolutional neural network model for intra‐frame video forgery detection

SUMMARYDue to the tremendous growth of video editing software, it has become extremely simple to introduce malicious content by manipulating multimedia data. This may include modification of videos either by adding or deleting selective frames with malicious intentions. Hence, it is essential to find the forged frames of the videos by introducing efficient and reliable video forensic methods. This article presents an automatic intra‐frame video forgery detection strategy based on a hybrid optimization tuned deep‐convolutional neural network (deep‐CNN) classifier. The significance of the proposed method lies in developing the proposed raven‐finch optimization algorithm that tunes the weights of the deep‐CNN to exhibit enhanced detection accuracy. The proposed raven‐finch optimization algorithm combines raven search agents and finches search agents, possessing the benefits of both search agents. The features of the input video frames act as the input to the deep‐CNN classifier that detects forgery. The performance of the proposed raven‐finch‐based deep CNN method is analyzed in terms of the performance indices, such as accuracy, sensitivity, and specificity. It is attained to be 97.56%, 95.48%, and 96.38%, respectively, which shows the superiority of the proposed method for intra‐frame forgery detection.

Evolving convolution neural network by optimal regularization random vector functional link for computational color constancy

To automatically eliminate the influence of external illumination on the acquired image, restore the true color information of the object, and provide true and stable color features for computer vision tasks, our work is based on the convolutional neural network (CNN) VGG19 for feature extraction of image information. Moreover, it proposes a model based on atom search optimization improved by chaotic-logistic maps to optimize the regularization random vector functional link (RRVFL). The optimized RRVFL replaces CNN’s regression layer to estimate the image illumination and then restore the image. First, we used chaotic-logistic maps to optimize the initial value of search agent of the atom search optimization (CASO) algorithm and then used the search agent segment mapping method proposed in our study to simultaneously optimize the number of nodes in hidden layer, regularization factors, input weight, and the bias of RRVFL hidden layer. This avoids the problem of serious fluctuation of prediction results and low accuracy caused by the randomness of the RRVFL parameters. After the predicted illumination information was obtained with the optimized RRVFL, the image was restored using the diagonal transformation method. Comparative experiments showed that the average angular errors of the illumination estimation of the V19-CASO-RRVFL algorithm proposed in our study were 1.73, 0.30, 0.3406, and 0.0976 lower than those of the gray-edge-2, GE-CASO-RRVFL, DS-Net, and color constancy on deep residual learning algorithms, respectively.

Distributed Co-Evolutionary Memetic Algorithm for Distributed Hybrid Differentiation Flowshop Scheduling Problem

This article deals with a practical distributed hybrid differentiation flowshop scheduling problem (DHDFSP) for the first time, where manufacturing products to minimize makespan criterion goes through three consecutive stages: 1) job fabrication in first-stage distributed flowshop factories; 2) job-to-product assembly based on specified assembly plan on a second-stage single machine; and 3) product differentiation according to customization on one of the third-stage dedicated machines. Considering the characteristics of multistage and diversified processing technologies of the problem, building new powerful evolutionary algorithm (EA) for DHDFSP is expected. To achieve this, we propose a general EA framework called distributed co-evolutionary memetic algorithm (DCMA). It includes four basic modules: 1) dual population (POP)-based global exploration; 2) elite archive (EAR)-oriented local exploitation; 3) elite knowledge transfer (EKT) among POPs and EAR; and 4) adaptive POP restart. EKT is a general model for information fusion among search agents due to its problem independence. In execution, four modules cooperate with each other and search agents co-evolve in a distributed way. This DCMA evolutionary framework provides some guidance in algorithm construction of different optimization problems. Furthermore, we design each module based on problem knowledge and follow the DCMA framework to propose a specific DCMA metaheuristic for coping with DHDFSP. Computational experiments validate the effectiveness of the DCMA evolutionary framework and its special designs, and show that the proposed DCMA metaheuristic outperforms the compared algorithms.

Traffic flow forecasting using natural selection based hybrid Bald Eagle Search-Grey Wolf optimization algorithm.

In a fast-moving world, transportation consumes most of the time and resources. Traffic prediction has become a thrust application for machine learning algorithms to overcome the hurdles faced by congestion. Its accuracy determines the selection and existence of machine learning algorithms. The accuracy of such an algorithm is improved better by the proper tuning of the parameters. Support Vector Regression (SVR) is a well-known prediction mechanism. This paper exploits the Hybrid Grey Wolf Optimization–Bald Eagle Search (GWO-BES) algorithm for tuning SVR parameters, wherein the GWO selection methods are of natural selection. SVR-GWO-BES with natural selection has error performance increases by 48% in Mean Absolute Percentage Error and Root Mean Square Error, with the help of Caltrans Performance Measurement System (PeMS) open-source data and Chennai city traffic data for traffic forecasting. It is also shown that the increasing population of search agents increases the performance.

Optimal architecture artificial neural network model design with exploitative alpha gray wolf optimization for soft calibration of CO concentration measurements in electronic nose applications

The low-cost and small size solid-state sensor arrays are suitable to implement a wide-area electronic nose (e-nose) for real-time air quality monitoring. However, accuracy of these low-cost sensors is not adequate for precise measurements of pollutant concentrations. Artificial neural network (ANN) estimation models are used for the soft calibration of low-cost sensor array measurements and significantly improve the accuracy of low-cost multi-sensor measurements. However, optimality of neural architecture affects the performance of ANN estimation models, and optimization of the ANN architecture for a training data set is essential to improve data-driven modeling performance of ANNs to reach optimal neural complexity and improved generalization. In this study, an optimal architecture ANN estimator design scheme is suggested to improve the estimation performance of ANN models for e-nose applications. To this end, a gray wolf optimization (GWO) algorithm is modified, and an exploitative alpha gray wolf optimization (EA-GWO) algorithm is suggested. This modification enhances local exploitation skill of the best alpha gray wolf search agent, and thus allows the fine-tuning of ANN architectures by minimizing a multi-objective cost function that implements mean error search policy. Experimental study demonstrates the effectiveness of optimal architecture ANN models to estimate CO concentration from the low-cost multi-sensor data.

WOASCALF: A new hybrid whale optimization algorithm based on sine cosine algorithm and levy flight to solve global optimization problems

In recent years, researchers have been focused on solving optimization problems in order to determine the global optimum. Increasing the dimension of a problem increases its computational cost and complexity as well. In order to solve these types of problems, metaheuristic algorithms are used. The whale optimization algorithm (WOA) is one of the most well-known algorithms based on whale hunting behavior. In this paper, the WOA algorithm is combined with the Sine Cosine Algorithm (SCA), which is based on the principle of trigonometric sine-cosine. The WOA algorithm has superior performance in the exploration phase in contrast with the exploitation phase, whereas the SCA algorithm has weaknesses in the exploitation phase. The levy flight distribution has been used in the hybrid WOA and SCA algorithm to improve these deficiencies. This study introduced a novel hybrid algorithm named WOASCALF. In this algorithm, the search agents' position updates are based on a hybridization of the WOA, SCA, and levy flight. Each of these metaheuristic algorithms has reasonable performance, however, the Levy distribution caused small and large distance leaps in each phase of the algorithm. Thus, it is possible for the appropriate search agent to move in different directions of the search space. The performance of the WOASCALF has been evaluated by the 23 well-known benchmark functions and three real-world engineering problems. The result analysis demonstrates that the exploration ability of WOASCALF has strong superiority over other compared algorithms.

Dragonfly Algorithm-Based Optimization for Selective Harmonics Elimination in Cascaded H-Bridge Multilevel Inverters with Statistical Comparison

Harmonics worsen the quality of electrical signals, hence, there is a need to eliminate them. The test objects under discussion are single-phase versions of cascaded H-bridge (CHB) multilevel inverters (MLIs) whose switching angles are optimized to eliminate specific harmonics. The Dragonfly Algorithm (DA) is used to eradicate low-order harmonics, and its statistical performance is compared to that of many other optimization techniques, including Particle Swarm Optimization (PSO), Accelerated Particle Swarm Optimization (APSO), Differential Evolution (DE), and Grey Wolf Optimization (GWO). Various scenarios of the algorithms’ search agent population for inverters with seven, nine, and eleven levels of output voltages are comprehensively addressed in this research. No algorithm shows total dominance in every scenario. The DA is least impacted by the change in dimensions of the narrated problem.

A study of efficacy and safety of newer chemical peels for acne vulgaris grade 1 to grade 3

Acne is a common and complex and multifactorial skin disease that affects individuals of all ages. It commonly presents as Seborrhoea and formation of comedones, papules and pustules. Aim of our study is to evaluate various new chemical peeling agent for search of treatment which stand out with maximum efficacy and least side effect.A total 120 patients were included in our study. Grading of acne was done according to Pillsbury classification and divided into five Treatment groups: Group A: Azelaic peel (20% Azelaic acid peel) Group B: Black peel (8% Acetic acid peel) Group C: Kojic acid peel (12% kojic acid peel) Group D: Lactic acid peel (90% lactic acid peel) Group E: Mandelic acid peel (35% mandelic acid peel). Patients in each arm were subjected to respective chemical peels at an interval of 15 days for a total of 6 sessions. Last visit was done at 12th week. Evaluation done by dermatological examination for acne including counting the number of comedones, papules and pustules.Patients with improvement at the end of 12 weeks the overall reduction rate was 60.4%, 72.4%, 24.1%, 35.2% and 50.1% with Azelaic peel, Black peel, Kojic acid peel, Lactic acid peel and Mandelic acid peel respectively.Our study thus concluded that the efficacy of Black peel is highest in the treatment of acne vulgaris grade I to III in Indian patients. The resolution was also observed good with Azelaic acid and Mandelic acid while response was not up to the mark by Lactic acid and Kojic acid peel. However, immediate erythema and burning sensations were more common side effects with black peel. Though the side effects observed with these peels reduced by itself, application of emollients, mild steroids and sun screeners hastened the recovery.

Enhanced Gaussian bare-bones grasshopper optimization: Mitigating the performance concerns for feature selection

As a recent meta-heuristic algorithm, the uniqueness of the grasshopper optimization algorithm (GOA) is to imitate the biological features of grasshoppers for single-objective optimization cases. Despite its advanced optimization ability, the basic GOA has a set of shortcomings that pose challenges in numerous practical scenarios. The GOA core limit is its early convergence to the local optimum and suffering from slow convergence. To mitigate these concerns, this study adopts the elite opposition-based learning and bare-bones Gaussian strategy to extend GOA's global and local search capabilities and effectively balance the exploration and exploitation inclinations. Specifically, elite opposition-based learning can help find better solutions at the early stage of exploration, while the bare-bones Gaussian strategy has an excellent ability to update the search agents. To evaluate the robustness of the proposed Enhanced GOA (EGOA) based on global constrained and unconstrained optimization problems, a straight comparison was made between the proposed EGOA and other meta-heuristics on 30 IEEE CEC2017 benchmark tasks. Moreover, we applied it experimentally to structural design problems and its binary version to the feature selection cases. Findings demonstrate the effectiveness of EGOA and its binary version as an acceptable tool for optimization and feature selection purposes.

Solving dynamic multi-objective problems using polynomial fitting-based prediction algorithm

Recently, dynamic multi-objective optimization has received growing attention due to its popularity in real-world applications. Inspired by polynomial fitting, this paper proposes a polynomial fitting-based prediction algorithm (PFPA) and incorporates it into the model-based multi-objective estimation of distribution algorithm (RM-MEDA) for solving dynamic multi-objective optimization problems. When an environment change is detected, the main mission of PFPA is to predict high-quality search populations for tracking the moving Pareto-optimal set effectively. Firstly, the non-dominated solutions obtained in past environments are utilized to predict high-quality solutions based on a multi-step movement strategy. Secondly, a polynomial fitting-based strategy is designed to fit the distribution of variables according to the obtained search populations, and capture the relationship between variables in the new search environment. Thirdly, some effective search agents are generated for improving population convergence and diversity based on characteristics of variables. To evaluate the performance of the proposed algorithm, experimental results on a set of benchmark functions, with a variety of different dynamic characteristics and difficulties, and two classical dynamic engineering design problems show that PFPA is competitive with some state-of-the-art algorithms.

Fabric wrinkle rating model based on ResNet18 and optimized random vector functional-link network

To improve the accuracy and efficiency of the objective evaluation of the fabric wrinkle model, the R18-COTSA-RVFL model is proposed in this paper. The fabric wrinkle rating model is based on the combination of ResNet18 (R18) and the enhanced random vector functional-link network, in which the enhanced random vector functional-link network replaces the Softmax layer of ResNet18. This model uses ResNet18 (R18) to extract features from wrinkled images. In addition, a random vector functional-link network model improved by the tunicate swarm algorithm was proposed for wrinkle-level classification. First, we used chaotic-logistic maps and opposition-based learning to optimize the initial population of the tunicate swarm algorithm. The proposed search agent fragment mapping method enables the tunicate swarm algorithm simultaneously to optimize the hidden layer node parameters and bias, input layer weight, and activation function of the random vector functional-link network, thus avoiding the inaccurate classification result caused by random parameter selection in the random vector functional-link network. From comparative experiments, the proposed R18-COTSA-RVFL has the highest average classification accuracy and stability.