In the field of dominance robust multi-objective optimization (MOP), how to accurately quantify the robustness of solutions under uncertainty is the challenge in the identification of dominant robust optimal solutions. This paper proposes a historical solution clustering-guided robust Multi-objective optimization algorithm based on decision variable assortment (RMOEA-HSA). It combines decision variable assortment (DVA) with historical solution clustering while introducing a dominance robustness enhancement metric (RobustDR), aiming to accurately quantify the dominance robustness by using systematic analyses of the aggregation patterns of historical solutions. Subsequently, this study develops three strategies: first, employs the aggregated performance of historical solutions to quantify the dominance robustness of individuals; second, an integrated approach combining K-neighbor clustering with an adaptive search mechanism to avoid local optima; last, a DBSCAN and cosine similarity-driven population reduction (DCPR) method for elite individual selection. Superior stability is demonstrated through nine benchmark tests against four existing state-of-the-art algorithms.

Robust Optimization Over Time (ROOT) is used to solve dynamic optimization problems with the aim of finding solutions that can be accepted over a long time. Most of the researches in this field try to seek new robust solutions by predicting the future fitness values of candidate solutions. However, predicting future fitness value is error prone. Therefore, this paper propose a multi-solution robust optimization over time with adaptive population control (MROOT-AC). Firstly, a adaptive population control mechanism was proposed to optimize the population in the unexplored promising region according to the convergence state of the population, so as to improve the diversity of the population. Secondly, in order to prevent the inefficiency and resource waste caused by frequent changes of deployment scheme, a multi-solution archive management mechanism has been proposed. Compared with the existing ROOT algorithms, the experimental results on the Generalized Moving Peak Benchmark (GMPB) show that the proposed algorithm can significantly improve the performance of the robust solution.

This study addresses the challenge of balancing robustness and optimality in robust multi-objective optimization under decision variable uncertainty and limited computational resources. To tackle this challenge, we propose an archive-based evolutionary framework that integrates two complementary strategies. First, a matching-driven two-stage evolutionary association mechanism assigns solutions to weight vectors and decomposes complex multi-objective problems into scalar subproblems, while maintaining a dynamic archive to enhance diversity and computational efficiency. Second, a Penalty Boundary Intersection (PBI)-based three-layer archiving strategy simultaneously evaluates and refines solutions based on convergence and robustness, ensuring an effective trade-off between optimality and stability. To further quantify solution quality, we introduce a novel Robust IGD Harmonic Index (RIHI), which integrates both average and worst-case performance under perturbations. Extensive experiments with benchmark problems show that the proposed algorithm consistently outperforms six state-of-the-art methods in solution quality, robustness, and computational efficiency. Furthermore, we validate the algorithm on a real-world order scheduling problem, demonstrating its ability to generate well-distributed and robust solutions under practical uncertainties.

Evolutionary algorithms (EAs) commonly solve robust multi-objective problems (RMOPs). Current approaches struggle with dynamic disturbances. Their fixed mutation strategies cause this limitation. Moreover, effectively balancing robustness and Pareto optimality remains a fundamental challenge in this field. To address these issues, this study proposes a novel robust multi-objective evolutionary algorithm that integrates reinforcement learning with a tri-mode archive strategy. The proposed method innovatively employs Q-learning to dynamically adjust mutation probabilities, where noise sensitivity serves as the state space and mutation probability as the action space, enabling autonomous policy optimization in dynamic disturbance environments. To achieve an optimal balance between robustness and optimality, a tri-mode elite archive framework is designed to maintain solution sets with high robustness, superior convergence, and broad distribution characteristics. Experimental results on standard benchmark functions demonstrate that the proposed algorithm significantly outperforms six state-of-the-art comparison algorithms in terms of both solution quality and computational efficiency.

To address the slow learning of traditional PSO algorithms and alleviate population collapse in the objective space, this paper proposes a multi-objective particle swarm optimization algorithm using a Logistic-tent chaotic map with GOBL and non-inertial Lévy flight (MOPSOLGN). Firstly, the Logistic-tent chaotic map with generalized opposition-based learning initializes individual positions, avoiding blindness and uncertainty in the initial population and improving its distribution. Secondly, a particle flight method combining an individual competition mechanism with k-means clustering divides particles into losers and winners, using a new non-inertial Lévy flight dynamical equation to balance exploration and exploitation and ensure the algorithm can escape local solutions. Thirdly, a differential mutation strategy enables the population to escape collapse and increases the diversity of the optimal solution set. Comparative experiments with state-of-the-art multi-objective algorithms on benchmark functions verify that MOPSOLGN allows individuals to converge to the real Pareto frontier more quickly and with better distribution.

The multi-objective feature selection problem typically involves two key objectives: minimizing the number of selected features and maximizing classification performance. However, most multi-objective evolutionary algorithms (MOEAs) face challenges in high-dimensional datasets, including low search efficiency and potential loss of search space. To address these challenges, this paper proposes a hybrid algorithm based on fast dimensionality reduction and multi-objective differential evolution with redundant and preference processing (termed DR-RPMODE). In DR-RPMODE, the DR phase uses the freezing and activation operators to remove many irrelevant and redundant features in the high-dimensional datasets, thereby achieving fast dimensionality reduction. Subsequently, the RPMODE algorithm continues the search on the reduced datasets, improving the traditional differential evolutionary framework from two aspects: duplicated and redundant solutions are filtered by redundant handling, and a preference handling method that pays more attention to classification performance is designed for different preference objectives of decision-makers. In the experiment, DR-RPMODE is compared with seven feature selection algorithms on 16 classification datasets. The results indicate that DR-RPMODE outperforms the comparison algorithms on most datasets, demonstrating that it not only achieves outstanding optimization performance but also obtains good classification and scalability results.

This paper regards UAV-assist aerial edge computing as a dynamic multi-objective optimization problem. In order to continuously track the the moving Pareto set, a new Holt-based prediction correction dynamic multi-objective evolutionary algorithm (HDMOEA) is proposed. It includes mainly three main strategies. Firstly, Wilcoxon signed-rank test method is employed to accurately detect environmental change, and the intensity of which is further detected by a new environment perception operator. Secondly, Holt-based prediction correction mechanism is constructed to predict the positions of individuals in the next time Window. The positions are corrected according to a reference point in order to enhance prediction accuracy and accelerate the search speed of the algorithm. Lastly, a new bi-mutation method is proposed used to maintaining the diversity of the population according to the intensity of environmental changes, thereby reduce the likelihood of the population falling into local optima. The proposed algorithm is compared with six state-of-the-art prediction dynamic multi-objective algorithms on the multiple benchmark test sets. The experimental results show that HDMOEA can faster continuous tracking Pareto Frontier, and obtain more accurate Pareto Frontier Set compared with other comparison algorithms.

Since the preferred multi-objective optimization solution set is a local optimal solution with decision maker’s preferences. To improve the performance of its, this paper proposes the angular preference multi-objective optimization algorithms with inverse initialization (AP-MOA). AP-MOA proposes three new strategies. The first is the target initialization strategy. There are two types of cases. For a bi-objective optimization problem, a better initial population is generated in the specified region on the preference information. For the tri-objective optimization problem, a tent mapping is used to generate uniform individuals in the specified region on the preference information. The second is two stage mutation, which is using genetic and differential mutation to produce excellent and stable offspring. The third is the angular preference guiding strategy. Two rays are drawn from the origin of the coordinates based on preference information to delineate a preferred solution region. According to experimental comparison, AP-MOA can converge quickly and obtain a satisfactory set of preference solutions.

Preference multi-objective optimization (PMOP) is hot problem in the field of current optimization. The searching objection of the PMOP which is local target pareto region according to preference information different from general multi-objective optimization. To improve the performance of the PMOP, a new Preference-based multi-objective optimization algorithms under the union mechanisms (UM-NSGAII) is proposed. Two strategies are proposed in UM-NSGAII. Firstly, initial population generated from limiting it to a certain range, which can reap a population dominated by preference information.Mutation individual is also come into being a certain range, which can reap progeny of populations closer to the pareto frontier corresponding to preference information. Secondly, Angle preference strategy is to identify the angle formed by a certain point with an arbitrary point and the origin as an angle preference region, where individuals in this region are selected in preference to individuals outside the region, this allows individuals evolving towards the target pareto region to be retained, facilitating rapid optimization searching.

In order to improve the convergence speed of the multi-objective optimization algorithm while obtaining good distribution and diversity, an adaptive dynamic parameter multi-objective optimization algorithm is proposed (ADPMO). The new algorithm consists of three main strategies. Firstly, a new mutation method based on individual competition mechanism integrated with k-means clustering is proposed, which updates the velocity and position information of the individuals that have failed to compete in each cluster, for improve the diversity of the solution set and avoid premature convergence. Secondly, an adaptively dynamical parameters strategy is proposed. In the process of speed updating, parameters that change dynamically with the number of population iterations, for enhance the convergence speed and convergence of the algorithm. At last, a cross-mutation strategy is introduced, for making the population out of collapse state and increase the diversity of the optimal solution set. Compared with other state- and-art multi-objective algorithms on the two types of benchmark functions, it is verified that the individuals can more converge faster to the real Pareto frontier with good distribution.

To accelerate the convergence of multi-objective optimization algorithm and achieve an optimization solution set with good diversity, this paper proposes the Pacesetter-Lévy Multi-Objective Particle Swarm Optimization using Arnold Chaotic Map with Opposition-Based Learning algorithm (PLMOPSOCO). Firstly, the Arnold Chaotic Map with Opposition-based Learning is proposed to generate some valuable particles in the stage of initialization while maintaining the diversity of population, which can speed up exploration. Secondly, a competition mechanism with k-means clustering is integrated to categorize particles into losers and winners. The flight direction of the loser particles is adjusted according to the Pacesetter-Lévy kinetic equation, an innovative approach to effectively enhance the population’s exploitative ability. Thirdly, a crossover mutation is implemented to assist particles to escape local optima and prevent population stagnation. Comparative studies on benchmark functions demonstrate that this new algorithm is more competitive in terms of both diversity and convergence when compared to several advanced multi-objective algorithms.

Multi-objective optimisation problem in Internet of Things technology has been widely concerned by researchers. The family of multi-objective particle swarm optimisation is among the most representative ones. However, there still exist the shortcomings of overspending and premature convergence. This paper proposes a many-objective particle swarm optimisation algorithm based on opposition-based mutation for elite mechanism. The new algorithm mainly includes three strategies: (1) Opposition-based learning population initialisation strategy, which is designed to avoid the blindness and uncertainty of initial population, and improves the distribution of population and accelerates speed of exploration. (2) Multi-elite opposition mutation mechanism, which is proposed to help particles get away from local optimal positions via a targeted exploration in the search space. (3) Singularity archive technique, which is established to disturb the global evolution trend and further balance the contradiction of convergence and diversity of the population. The effectiveness of the proposed algorithm is verified by comparing 11 algorithms in the simulation experiments.

For multi-objective optimisation problems, a balance between convergence and diversity in multi-objective particle swarm algorithms is the key to approach real Pareto fronts with well-distributed. In order to obtain the Pareto optimal set with good distribution, a multi-objective particle swarm optimisation algorithm based on Levy mutation and information entropy is proposed in this paper. Firstly, an entropy adaptive strategy is proposed to balance the exploration and exploitation ability of the swarm, which guides the flight direction of particles via the adaptively adjusting parameters with information entropy of the swarm. Secondly, a Levy mutation operator is proposed to ensure that the algorithm has the ability to jump out of the local solution. The new mutation operator can control the magnitude of particle mutation by random steps, so that the mutation is more anisotropic and diverse, thus ensuring that the particles still have a large global exploration ability in the late iteration as well as increasing the local exploitation accuracy. Finally, the experimental results in benchmark test functions show that the proposed algorithm has better exploration ability than several compared algorithms and can approach real Pareto front with better distribution.

To accelerate the convergence speed and enhance robustness, a back-diffusion median integrated evolutionary algorithm (BMIEA) is proposed combining the advantages of particle swarm optimization (PSO) and differential evolution algorithm (DE) in this paper. The BMIEA includes three mainly optimization strategies. (1) Firstly, a new meanmedian velocity updating formula is proposed to control optimal path of individuals. It can accelerate the convergence speed via reducing adverse effects of outliers on the population. (2) Secondly, a random differential mutation (RDM) inspired by the DE is devised to avoid the individuals trapping into local optimum via getting one more chance to explore optimal position while exploiting local region in each evolution. (3) Thirdly, a targeted exploration method i.e., back-diffusion operation, is proposed inspired by the duality principle to further accelerate the convergence rate and enhance robustness of algorithm. A series of simulation experiments have verified that BMIEA algorithm has revealed competitiveness compared with 13 state-of-art GOBL-based optimization algorithms. (c) 2022 Elsevier Inc. All rights reserved.

In recent years, the model of improved GAN has been widely applied in the field of machine vision. It not only covers the traditional image processing, but also includes image conversion, image synthesis and so on.. Firstly, this paper describes the basic principles and existing problems of GAN, then introduces several improved GAN models, including Info-GAN, DC-GAN, f-GAN, Cat-GAN and others. Secondly, several improved GAN models for different applications in the field of machine vision are described. Finally, the future trend and development of GAN are prospected.

采用超音速火焰喷涂工艺(HVOF)制得不同配比的Co-W-WC吸收层,之后在上述吸收层表面继续通过Sol-Gel方法制得SiO2复合膜,测试了其微观组织及光学性能.研究结果表明:复合膜中形成了Co、W与WC特征峰,有些Hcp-Co颗粒在喷涂阶段转变成Fcc-Co结构.逐渐提高WC的加入量后,复合膜表面形成了更少的弥散态未熔颗粒,获得了光滑致密组织.W:WC达到1:1时,有助于降低表面粗糙度,使吸收层获得更低发射率.逐渐减小W/WC比值后,反射率发生增大,吸收率下降,表明W具备比WC更强的光谱吸收能力.SiO2薄膜因具备较小的折射率,有助于短波光线透过,因此吸收层可以获得更多入射光.经过100 h热处理后,吸收率增大至0.915,发射率达到0.2516,并未观察到复合膜脱落的情况,表明SiO2复合膜具有稳定的光学性能稳定.

Since the particle swarm optimization (PSO) was proposed to overcome the inherent defects of PSO such as premature convergence and dependent on parameters settings, different PSO variants are devised to optimize different complex optimization problems; NOPSO is an excellent representative among them. This paper ensembles the three types of velocity update formulas proposed in NOPSO and presents a dual-drive opposition-based non-inertial PSO to improve the robustness of algorithm while ensuring the searching efficiency and solving accuracy of optimization procedure. Two main strategies are introduced in the new algorithm: (1) a dual-drive velocity update formula (DDVM) is proposed to control move of particles and (2) an elite differential evolutionary mutation strategy (EDEM) is devised to help particles escape from local optimum. The modified algorithm with two above strategies is proved to be competitive compared with some state-of-the-art OBL-based PSO including NOPSO and can be effectively applied to deep learning in IoTs in the foreseeable future.

Prediction problems are difficult to be carried out in a dynamic environment, for two key questions: one is how to monitor environmental changes, the other is how to make respond timely after the environment has changed. Multi-population strategy is often adopted to address both two key problems. However, there are two stubborn questions that limit and affect the effectiveness of the strategy, i.e., (1) Search overlaps are easy to occur which lead to lose the ability of local exploit, (2) In the course of evolution, the search range of subpopulations tends to assimilate, so the subpopulation will gradually lose the ability to explore whole search space. Therefore, this paper adopts multi-population strategy, and presents a novel intelligence algorithm based on particle swarm optimization to solve the above problem, called adaptively reversed diffusion dual-drive evolutionary algorithm (ARDDEA). (1) Firstly, ARDDEA monitors the environmental changes by setting the global dynamic sentry in each subgroup. (2) Secondly, in order to avoid searching overlapping of the sub-population, a new exclusion strategy is proposed in this paper. A new distance determination method, i.e., between-swarms average Mahalanobis​ distance, is devised in the exclusion strategy to decide the inter-population distance. If the distance is too small between two sub-populations, furtherly, a Hill–valley decision function is used to determine whether they tracked the same peak or not. If so, the inferior subpopulations will be reinitialized by a reverse diffusion operation (RD) proposed in this paper. Besides, (3) a new dual-drive kinetic updating equation is proposed to enhance the search capability of the population. The new algorithm compared with several state-of-art dynamic optimization algorithms on the moving peak problem. The results show that the ARDDEA algorithm can track the optimal solution more effectively in the dynamic environment, and shows strong robustness and adaptability. It is a hope algorithm applied to prediction problems.

针对信息化教学基础上全在线教学转为线下教学衔接策略进行研究,提出将课程划分课题,以理论带动实践逐层深入的衔接模式,进行理实融合,补足全在线教学的缺漏环节,突出理论知识在实际生活中的应用.实践教学中融入思政教育,提升学生的信息素养、职业道德素养,为培养德才兼备的人才提供一定的参考.

针对突发疫情状况,落实教育部"停课不停学"的要求,根据高职学生问卷调查提出"超星平台+学习通+钉钉直播"多维互动混合式授课模式,充分挖掘信息化教学元素,提升特殊时期教学质量.在此教学模式中,教师通过超星平台发布各种教学资源供学生预习,利用钉钉直播和学习通进行多维互动,取得了良好的教学效果;课程中适时融入思政教育,润物于无声,为以后开展线上教学、培养高素质人才提供一定的参考.