Multi-objective Evolution Research Articles

Algorithm for solving multi-objective hybrid flow shop scheduling problem

Aiming at the multi-objective unrelated parallel machine hybrid flow shop scheduling problem, a multi-objective mathematical model is established with the goal of minimizing the maximum completion time, total machine energy consumption and machine processing cost. An improved multi-objective evolution algorithm based on decomposition (IMOEAD) is proposed. The uniform design table is used to generate the initial weight vector to improve the population diversity. Normal distribution crossover and adaptive Gaussian mutation are used to improve the global and local search capabilities of the algorithm. Individuals are selected in the neighborhood of the weight vector to generate new solutions. The non-dominated rank and crowding distance are used to update the external archive. The inverse generation distance, generation distance and number of non-dominated solutions are used as performance indicators. Through a large number of case simulations, the algorithm is compared with the non-dominated sorting genetic algorithm II and the multi-objective evolution algorithm based on decomposition. The results verify the effectiveness of the algorithm.

Multi-Objective Automatic Clustering Algorithm Based on Evolutionary Multi-Tasking Optimization

Data mining technology is the process of extracting hidden knowledge and potentially useful information from a large number of incomplete, noisy, and random practical application data. The clustering algorithm based on multi-objective evolution has obvious advantages compared with the traditional single-objective method. In order to further improve the performance of evolutionary multi-objective clustering algorithms, this paper proposes a multi-objective automatic clustering model based on evolutionary multi-task optimization. Based on the multi-objective clustering algorithm that automatically determines the value of k, evolutionary multi-task optimization is introduced to deal with multiple clustering tasks simultaneously. A set of non-dominated solutions for clustering results is obtained by concurrently optimizing the overall deviation and connectivity index. Multi-task adjacency coding based on a locus adjacency graph was designed to encode the clustered data. Additionally, an evolutionary operator based on relevance learning was designed to facilitate the evolution of individuals within the population. It also facilitates information transfer between individuals with different tasks, effectively avoiding negative transfer. Finally, the proposed algorithm was applied to both artificial datasets and UCI datasets for testing. It was then compared with traditional clustering algorithms and other multi-objective clustering algorithms. The results verify the advantages of the proposed algorithm in clustering accuracy and algorithm convergence.

Evolutionary multi-objective overlapping community detection based on fusion of internal and external connectivity and correction of node intimacy

In the field of community detection, node attribute information plays an important role in community division. Existing methods use topology structure and node attribute information to discover non-overlapping communities. However, so far, attribute information has not been fully utilized in overlapping community detection. To address this, we propose a new overlapping community detection method called “evolutionary multi-objective overlapping community detection based on Fusion of internal and external Connectivity and Correction of Node Intimacy” (FCCNI). Firstly, we propose a fusion strategy based on internal and external connectivity, which integrates some communities with sparse intra-connections and dense inter-connections. This automatically determines, reconfirms, and corrects the number of communities. Secondly, a function is designed to calculate the intimacy between nodes, and the node label with the highest intimacy is selected to correct the current wrong node. The correction strategy is used in two stages of initialization and multi-objective evolution to obtain a more accurate node label. Finally, a method that considers not only the connections of the community, but also the node attribute, is designed to obtain the overlapping community indirectly from the non-overlapping community. The experimental results on five real-life networks and four classical synthetic networks show that FCCNI achieves better overlapping community division, compared with six state-of-the-art comparison algorithms from the literature.

Scaling Multiobjective Evolution to Large Data With Minions: A Bayes-Informed Multitask Approach.

In an era of pervasive digitalization, the growing volume and variety of data streams poses a new challenge to the efficient running of data-driven optimization algorithms. Targeting scalable multiobjective evolution under large-instance data, this article proposes the general idea of using subsampled small-data tasks as helpful minions (i.e., auxiliary source tasks) to quickly optimize for large datasets-via an evolutionary multitasking framework. Within this framework, a novel computational resource allocation strategy is designed to enable the effective utilization of the minions while guarding against harmful negative transfers. To this end, an intertask empirical correlation measure is defined and approximated via Bayes' rule, which is then used to allocate resources online in proportion to the inferred degree of source-target correlation. In the experiments, the performance of the proposed algorithm is verified on: 1) sample average approximations of benchmark multiobjective optimization problems under uncertainty and 2) practical multiobjective hyperparameter tuning of deep neural network models. The results show that the proposed algorithm can obtain up to about 73% speedup relative to existing approaches, demonstrating its ability to efficiently tackle real-world multiobjective optimization involving evaluations on large datasets.

Two-Stage Multiobjective Evolution Strategy for Constrained Multiobjective Optimization

For the past many years, several constrained multiobjective evolutionary algorithms (CMOEAs) have been designed for solving constrained multi-objective optimization problems (CMOPs). In these CMOEAs, some constraint-handling techniques (CHTs) were proposed to balance the convergence and constrained satisfaction, however, they still face some serious challenges, such as premature convergence to the local optimal region and labor-intensive tuning of parameters for a specific CMOP. Furthermore, most of the existing CHTs are derived by solving constrained single-objective optimization. The information hidden from the feasible non-dominated set (FNDS) has not been fully utilized. This study proposed a novel parameter-less constraint handling technique, which divides the entire population into three mutually exclusive subsets dynamically: FNDS, the subset dominated by FNDS, and the subset not dominated by FNDS. According to the proposed division of labor, it is not necessary to balance the convergence and constrained satisfaction in each subset. To avoid being entrapped in local optima, the proposed algorithm adopts a two-stage strategy to solve CMOPs. In the first stage, the proposed algorithm focuses solely on converging toward the unconstrained Pareto front without considering the constrained satisfaction. In the second stage, the FNDS constraint handling technique is adopted to guide the population converging toward constrained Pareto front effectively. The performance of the proposed algorithm was compared to that of nine state-of-the-art CMOEAs, and the comparison results show that the proposed algorithm performs significantly better on the CF, MW, and LIRCMOP test suites.

Hybrid multi-objective evolutionary model compression with convolutional neural networks

Deep learning has been utilized in the fields of image processing, natural language processing and speech recognition. For improving the structure of deep learning, how to compress Convolutional Neural Networks has become a major focus topic. The study proposes a sparsification-based Convolutional Neural Networks hybrid multi-objective evolutionary model compression method based on hybrid multi-objective evolution, in which the process improves the sparsity of the model by sparse pruning and introduces a time-domain residual module to provide more robust generalization performance, and finally the function of the research method is tested. The test outcomes indicate that the Convolutional Neural Networks model generated by the research method drops to the lowest value in the first 10 iterations; in the sparsity test, the total weight density of the model generated by the research method is about 36%; when the latency test is conducted, the latency of the research method is between 3 ms and 14 ms when the weight sparsity is 60%. The above results show that the research method has better model compression effect and can generate Convolutional Neural Networks models with both model accuracy and model size, which can provide new technical support for model compression of convolutional neural networks.

TEAM: A Layered-Cooperation Topology Evolution Algorithm for Multi-Sink Internet of Things

Numerous sensor nodes deployed in the Internet of Things (IoT) can form a large heterogeneous network. The increased energy consumption of sensor nodes and the unbalanced communication load on multiple sink nodes reduce the energy efficiency of the network. Moreover, frequent network attacks also pose severe challenges to topology robustness. Optimizing the network topology to achieve the balance between energy efficiency and robustness is a complex problem. Multi-objective heuristic algorithms based on genetic evolution are commonly used to solve joint optimization problems. However, due to the lack of global search ability caused by the loss of genetic diversity, genetic operations are prone to premature convergence during multi-objective evolution. Therefore, this paper introduces multi-population cooperation into the multi-objective evolution process and proposes a novel layered-cooperation Topology Evolution Algorithm for Multi-sink IoT (TEAM). In TEAM, information entropy is used to measure the effectiveness of load balancing on multiple sink nodes. The crossover and mutation probabilities of different populations are dynamically adjusted to ensure genetic diversity. A layered-cooperation mechanism is designed to avoid premature convergence. Extensive experiments confirm that TEAM can effectively improve the energy efficiency and robustness of network topology while balancing the communication load on multi-sink nodes.

Joint Interdependent Task Scheduling and Energy Balancing for Multi-UAV-Enabled Aerial Edge Computing: A Multiobjective Optimization Approach

To provide a dependency-aware application, multiple UAVs are employed to serve a ground user with a set of interdependent tasks. This leads to a new computing paradigm called as multi-UAV enabled aerial edge computing (MU-AEC). For the large-scale application of MU-AEC, both the task-centric objective and UAV-centric objective should be simultaneously considered. Thus, we focus on the joint interdependent task scheduling and energy balancing for MU-AEC by using a multi-objective optimization approach, which enables a decision maker to identify the optimal solutions corresponding to the best feasible tradeoffs between the two objectives. A constrained multi-objective optimization problem involving two objectives, i.e., the makespan minimization of all tasks and energy balancing among different UAVs, is formulated. In the solution methodology, we propose a constrained decomposition-based multi-objective evolution algorithm. To quickly seek more superior solutions, a local search mechanism by utilizing the objective information, and an improved genetic operator are proposed for remarkable performance improvements. Finally, numerical results demonstrate that compared with the baseline algorithms, our algorithm achieves both advantages in increasing the convergence and diversity of the solutions.

Active broad learning with multi-objective evolution for data stream classification

In a streaming environment, the characteristics and labels of instances may change over time, forming concept drifts. Previous studies on data stream learning generally assume that the true label of each instance is available or easily obtained, which is impractical in many real-world applications due to expensive time and labor costs for labeling. To address the issue, an active broad learning based on multi-objective evolutionary optimization is presented to classify non-stationary data stream. The instance newly arrived at each time step is stored to a chunk in turn. Once the chunk is full, its data distribution is compared with previous ones by fast local drift detection to seek potential concept drift. Taking diversity of instances and their relevance to new concept into account, multi-objective evolutionary algorithm is introduced to find the most valuable candidate instances. Among them, representative ones are randomly selected to query their ground-truth labels, and then update broad learning model for drift adaption. More especially, the number of representative is determined by the stability of adjacent historical chunks. Experimental results for 7 synthetic and 5 real-world datasets show that the proposed method outperforms five state-of-the-art ones on classification accuracy and labeling cost due to drift regions accurately identified and the labeling budget adaptively adjusted.

Multi-objective Evolutionary Algorithm Supported Construction of Architectural Window Design Model Based on Visual Comfort - A Case Study of Office Buildings in Cold Regions

Abstract With the accelerated advancement of big data Internet development, people gradually realize the urgency and importance of energy saving, so the multi-objective optimization research for the performance of office buildings in cold regions is significant, and this paper focuses on the construction of a building window design model based on visual comfort supported by multi-objective evolutionary algorithm- taking office buildings in cold regions as an example, firstly, by reviewing the literature to understand the principle, algorithm, and concept of multi-objective evolution, construct the building window design model based on visual comfort according to the objective function and extract the standard model of inner corridor slab space building in office buildings in cold regions of Harbin and Shenyang. The optimized design of building window parameters was carried out using the established joint simulation and optimization work platform. The results show that for Harbin, as the window heat transfer coefficient increases, the heating energy consumption increases, the cooling energy consumption increases more slowly, and the total energy consumption increases linearly. An increase in window heat transfer coefficient by 0.1W/ (m 2 ∙K) Increases are cooling energy consumption by 0.18%, heating energy consumption by 0.78%, and total energy consumption by 0.46%. For Shenyang, as the window heat transfer coefficient increases, the heating energy consumption increases, the cooling energy consumption increases more slowly, and the total energy consumption increases linearly. If the heat transfer coefficient of the window increases by 0.1W/ (m 2 ∙K), cooling energy consumption increases by 0.18%, heating energy consumption increases by 0.78%, and total energy consumption increases by 0.46%. This study provides a theoretical basis for extracting standard models for other building types while making the results more generalizable and improving the efficiency of sustainable office building design.

An adaptive dual-population based evolutionary algorithm for industrial cut tobacco drying system

Industrial cut tobacco drying is one of the most important processes in cigarette production, which affects the taste, cut tobacco consumption and other indicators of cigarette products. Due to the complicated process and parameters involved, the production of drying system is difficult to improve. In this paper, the model of tobacco drying system is established and optimized. First, an eighth-order nonlinear first-principle model is established, and its corresponding constrained multi-objective optimization problem is constructed based on the multiple requirements in industrial production. Furthermore, an adaptive dual-population based evolutionary algorithm (ADPEA) is proposed in which an assistant population is introduced to balance the feasibility, diversity and convergence. Feasible solutions are preferentially reserved to the next generation in the main population, while diversity and convergence are considered more in the assistant population. The ADPEA is used to optimize the tobacco drying system and is compared with four state-of-the-art multi-objective evolution algorithms. The experimental results reveal that ADPEA has a better performance, and the optimization results could help engineers adjust the process parameters according to the requirements of different batches and brands of cigarette products to ensure that the whole production process can meet the technological requirements while saving energy and reducing emissions.

Disease biomarker identification based on sample network optimization

A large amount of evidence shows that biomarkers are discriminant features related to disease development. Thus, the identification of disease biomarkers has become a basic problem in the analysis of complex diseases in the medical fields, such as disease stage judgment, disease diagnosis and treatment. Research based on networks have become one of the most popular methods. Several algorithms based on networks have been proposed to identify biomarkers, however the networks of genes or molecules ignored the similarities and associations among the samples. It is essential to further understand how to construct and optimize the networks to make the identified biomarkers more accurate. On this basis, more effective strategies can be developed to improve the performance of biomarkers identification. In this study, a multi-objective evolution algorithm based on sample similarity networks has been proposed for disease biomarker identification. Specifically, we design the sample similarity networks to extract the structural characteristic information among samples, which used to calculate the influence of the sample to each class. Besides, based on the networks and the group of biomarkers we choose in every iteration, we can divide samples into different classes by the importance for each class. Then, in the process of evolution algorithm population iteration, we develop the elite guidance strategy and fusion selection strategy to select the biomarkers which make the sample classification more accurate. The experiment results on the five gene expression datasets suggests that the algorithm we proposed is superior over some state-of-the-art disease biomarker identification methods.

A chaotic differential evolution and symmetric direction sampling for large-scale multiobjective optimization

As large-scale multiobjective optimization problems (LSMOPs) contain many decision variables, most existing large-scale multiobjective optimization algorithms (LSMOEAs) require millions or even tens of millions of function evaluations to obtain high-quality solutions, which is unaffordable for many real-world problems. Moreover, few studies are dedicated to solving LSMOPs under limited function evaluations. This article proposes a Chaotic Differential Evolution and Symmetric Direction Sampling for large-scale multiobjective optimization (CDE-SDS). First, based on the decision variable information of nondominated solutions, a chaotic differential evolution strategy is adopted to accelerate the convergence of the population towards the Pareto front. Second, a symmetric direction sampling strategy is employed, which generates sampling solutions by constructing symmetric search directions in a large-scale decision space. Thus, the high-dimensional decision space can be explored more fully. The two strategies optimize the population alternately to balance exploitation and exploration. The performance of CDE-SDS is compared with seven state-of-the-art large-scale multiobjective evolution algorithms on an artificial test suit LSMOP with 500 to 5000 decision variables and a real-world time-varying ratio error estimation problem with 3000 decision variables. Experimental results show that the performance of CDE-SDS significantly outperforms the seven compared algorithms in terms of both diversity and convergence of solutions under limited function evaluations.

EEG-based Emotion Recognition via Transformer Neural Architecture Search

Emotion recognition based on electroencephalogram (EEG) plays an increasingly important role in the field of brain–computer interfaces. Recently, deep learning has been widely applied to EEG decoding owning to its excellent capabilities in automatic feature extraction. Transformer holds great superiority in processing time-series signals due to its long-term dependencies extraction ability. However, most existing transformer architectures are designed manually by human experts, which is a time-consuming and resource-intensive process. In this article, we propose an automatic transformer neural architectures search (TNAS) framework based on multiobjective evolution algorithm (MOEA) for the EEG-based emotion recognition. The proposed TNAS conducts the MOEA strategy that considers both accuracy and model size to discover the optimal model from well-trained supernet for the emotion recognition. We conducted extensive experiments to evaluate the performance of the proposed TNAS on the DEAP and DREAMER datasets. The experimental results showed that the proposed TNAS outperforms the state-of-the-art methods.

A Novel Recommendation Algorithm Based on Multi-objective Evolution

During the past two decades, recommender systems have been playing an increasingly significant role in business decisions for providing personalized recommendations. However, the majority of the present algorithms attach excess importance to accuracy while overlooking users demand for diverse items. We have proposed a novel multi-objective evolutionary recommendation algorithm, which has the capability to look for compromise among accuracy and diversity. Moreover, due to the fact that Pareto optimal solutions of recommendation tend to be sparse, the proposed algorithm implements a new strategy of population initialization with the consideration of the sparse nature of the Pareto solutions. The proposed algorithm will take effect in the recommender systems.

A Pareto Front grid guided multi-objective evolutionary algorithm

For multi-objective optimization problems with irregular Pareto Fronts, most widely used decomposition methods in MOEA/D (multi-objective evolutionary algorithms based on decomposition) have shown to be lack of the ability to balance the diversity and the convergence to track the true Pareto Fronts during the search. This research investigates the recently proposed grid-based decomposition methods, which reflect the inherent characteristics of the neighborhood structure in the solution to address the issues of diversity and convergence. The performance of the grid-based decomposition method, however, depends on the size of its grid segmentation, and its time complexity increases with the number of grids. In order to improve the computational efficiency, we propose a new concept of Pareto Front grid to guide the search in MOEA. In order to reduce the computing time, a new nadir point estimation strategy based on statistical analysis has been proposed to estimate the whole population. In addition, based on the idea of knee point, a novel grid-based knee point selection method is proposed in the environmental selection of the next generation. Finally, a grid-based decomposition multi-objective evolutionary algorithm with Pareto Front Grid (PFG-MOEA) is proposed. Extensive experimental analysis demonstrates the effectiveness of the proposed PFG-MOEA against state-of-the-art multi-objective evolution algorithms. As the extension of the CDG-MOEA (Constrained Decomposition approach with Grids MOEA) algorithm in the literature, PFG-MOEA can obtain better performance by consuming much less computing time.

A clustering-based extended genetic algorithm for the multidepot vehicle routing problem with time windows and three-dimensional loading constraints

Since the multidepot vehicle routing problem with time windows and three-dimensional loading constraints (MDVRPTW-TDLC) is a multiconstraint and combinatorial optimization problem, this study proposes a resource sharing scheme on the basis of customer clustering to achieve a balance of spatial resource spatial distributions, thereby reducing the operating costs in a collaborative multidepot logistics network. Moreover, this study proposes a vehicle compartment partition strategy based on cargo characteristics such as type and sizes to increase vehicle loading rates, after which the proposed MDVRPTW-TDLC is formulated as a bi-objective mixed-integer programming model to minimize the total operating costs while maximizing the average loading rate of vehicles. Subsequently, a two-stage hybrid algorithm combining a three-dimensional (3D) k-harmonic means clustering algorithm and extended nondominated sorting genetic algorithm-II (ENSGA-II) is developed to find the Pareto optimal solutions and solve the MDVRPTW-TDLC optimization model, followed by an introduction of the 3D k-harmonic means clustering algorithm to cluster customers, thereby reducing computational complexity. The proposed ENSGA-II, integrating the improved Clarke–Wright savings algorithm and NSGA-II is then used to obtain the final optimal solutions, followed by the application of a selective exchange mechanism between the clustering algorithm and ENSGA-II to enhance the global and local optimization capability, enabling fast and efficient computation of optimal solutions. Based on our investigations revealed through an algorithm comparison using the multiobjective harmony search algorithm, multiobjective evolution algorithm, multiobjective particle swarm optimization, monarch butterfly optimization, Runge Kutta optimizer, and Harris hawks optimization, ENSGA-II demonstrates better performance in reaching high-quality objective function values. Finally, a real-world case study of Chongqing City, China is employed to verify the efficiency of the proposed model and optimization algorithm, after which a comparison of the optimization results among different vehicle compartment partition strategies demonstrates the efficiency of the proposed method. We also discuss 16 scenarios with and without the vehicle compartment partition strategy with computational results showing an improved vehicle loading rate and reduced logistics operating costs with the adopted vehicle compartment partition strategy. Therefore, our results confirm that the proposed MDVRPTW-TDLC can facilitate the sustainable development of logistics operations, and providing decision support for constructing intelligent logistics systems smart cities.

MOE/RF: A Novel Phishing Detection Model Based on Revised Multiobjective Evolution Optimization Algorithm and Random Forest

MOE/RF: A Novel Phishing Detection Model Based on Revised Multiobjective Evolution Optimization Algorithm and Random Forest

Gray wolf optimization-based self-organizing fuzzy multi-objective evolution algorithm

Gray wolf optimization-based self-organizing fuzzy multi-objective evolution algorithm

Supersonic inlet flow recognition by hybrid-mutation non-dominated sorting genetic algorithm with support vector machines

The recognition of supersonic inlet flow pattern has become a research hotspot in recent years. In this paper, the dual external pressure supersonic inlet is taken as the research object. To explore the flow characteristics of the inlet, time-mean processing on the inlet pressure signal collected by sensors is conducted first, and the features of the inlet pressure data in time domain and frequency domain are extracted, respectively. As feature selection (FS) plays an important role in classification tasks and has been recently studied as a multi-objective optimization problem, two objectives of FS are considered and an improved non-dominated sorting genetic algorithm NSGA2 with hybrid mutation operators using support vector machines (SVM) as classifiers is proposed, aiming to simultaneously select feature subsets and optimize SVMs hyper-parameters. In addition, a way to deal with variation transgression is proposed to make the mutation operator of the single-objective evolution fit well in the multi-objective evolution algorithm. Experimental results on 31 sensor datasets demonstrate that our proposed algorithm can achieve competitive classification accuracy while obtaining a smaller size of feature subset compared with particle swarm optimization algorithm and some multi-objective optimization algorithms using single-objective evolution mutation operators.