Recently, machine learning-embedded large-scale multiobjective evolutionary algorithms (LMOEAs) have shown great promise in solving large-scale multiobjective optimization problems (LMOPs). However, the fast convergence of the population to the true Pareto-optimal front (POF) and even distribution of the obtained Pareto-optimal solutions (POSs) on the POF are not adequately considered when tackling an LMOP. Besides, existing LMOEAs typically pair solutions with a matching rule and employ a network to learn the evolution pattern among the obtained solution pairs. It is difficult to learn various evolution patterns through a simple network, which hinders the collaboration of different patterns for enhancing the search capability. Facing such difficulties, this article proposes an LMOEA with multipattern learning and collaboration (LMOEA-MLC), where a single-hidden-layer multioutput network (SMN) is established to learn inductive and hybrid evolution patterns. Specifically, two inductive ones can be learned with the solution pairs built by two matching rules toward fast convergence and even distribution, respectively. Moreover, the solution pairs considering the fusion of the two inductive ones are collected, enabling SMN to learn a hybrid one and thus making a tradeoff between fast convergence and even distribution. Besides, the learned evolution patterns collaborate to enhance the search capability due to the distinct patterns. To enhance learning speed, SMN’s parameters are updated by an incremental random vector functional link (IRVFL). In our experiments, comprehensive comparisons with eight state-of-the-art LMOEAs demonstrate the significant performance improvement of LMOEA-MLC in handling LMOPs.

Given the high complexity and significant uncertainty of the multiple loads in integrated coal mine energy systems, interval prediction provides more comprehensive information for developing dispatch strategies, thus improving the reliability and robustness of the decision-making process. However, traditional interval prediction evaluation metrics fail to adequately reflect the impact of interval prediction results on the optimal dispatch cost. To address this, this paper proposes a day-ahead interval prediction method for the multivariate load of an integrated coal mine energy system, explicitly considering the dispatch cost. First, multi-task learning is combined with quantile regression to explore the coupling characteristics between multiple loads systematically. Next, an optimization model is constructed that simultaneously accounts for interval prediction performance and dispatch cost to determine the optimal upper and lower quantile combinations, adapting to dynamic load changes. Finally, the upper and lower quantile combinations are optimized using genetic algorithms, and the prediction intervals of multiple loads for the day ahead are obtained based on the optimized quantile combinations and the trained model. The proposed method is applied to an integrated coal mine energy system and compared with existing methods. Experimental results demonstrate that the proposed method ensures the prediction accuracy of the day-ahead multivariate load interval and significantly reduces the system’s dispatch cost.

This article investigates the global predefined-time consensus (PTC) of multiagent system (MAS) via constructing a duplex communication network. Unlike the traditional finite-/fixed-time convergence, our method allows the upper-bound of settling-time to be an explicit constant, which is tunable and can be set beforehand without relating with the network information, controlling parameters, and initial conditions. In particular, our approach uses a smooth, nonchattering consensus scheme that avoids conventional discontinuous functions like signum and absolute value functions. By the Lyapunov stability analysis, the sufficient criterion is deduced for ensuring the PTC of MAS. Finally, simulations confirm the effectiveness of our proposed nonchattering scheme.

The dispatch optimization of coal mine integrated energy system is challenging due to high dimensionality, strong coupling constraints, and multiobjective. Existing constrained multiobjective evolutionary algorithms struggle with locating multiple small and irregular feasible regions when solving the dispatch problem. To address this issue, we here develop a multiform EA framework that incorporates the dispatch-correlated domain knowledge to effectively deal with strong constraints and multiobjective optimization. Possible evolutionary multiform construction strategy based on complex constraint relationship analysis and handling, i.e., constraint-coupled spatial decomposition, constraint strength classification, and constraint handling technique, is first explored. Within the multiform evolutionary optimization framework, two strategies, i.e., an elite-guided knowledge transfer by designing a special crowding distance mechanism to select dominant individuals from each task and a neighborhood-driven dual mutation to effectively balance the diversity and convergence of each optimized task for the differential evolution algorithm, are further developed. The performance of the proposed algorithm in feasibility, convergence, and diversity is demonstrated in a case study of a coal mine integrated energy system (IES) by comparing with CPLEX solver and eight state-of-the-art constrained multiobjective EAs.

Recently, transfer-learning-based dynamic multiobjective optimization algorithms (TL-DMOAs) have been shown to be very promising in solving dynamic multiobjective optimization problems (DMOPs). However, it is difficult for them to model knowledge capable of delineating the Pareto optimal solutions (POSs) found in each historical environment, because the POSs’ distribution cannot be adequately reflected. Besides, existing TL-DMOAs normally focus on acquiring knowledge from historical environments, but neglect correlations behind them for excavating potential knowledge, restricting the performance in generating high-quality initial populations (HIPs). To address these issues, herein a DMOA with multisource and hidden source-based knowledge transfer (DMOA-MHKT) is proposed. First, we design a knowledge extraction strategy by introducing mean shift, a nonparametric clustering method, to cluster the historical POSs. As clusters’ representatives, the cluster centers are considered to represent environmental knowledge, because they can adequately reflect the POSs’ distribution. Second, the most similar historical environment through environmental match and the last one are selected as two explicit sources. In the former source the POSs’ cluster centers are treated as its knowledge. By contrast, based on the POSs’ cluster centers and knee points in the latter source, a scoring method is designed to generate environmental knowledge by depicting the dynamics between two continuous environments. Third, after aligning knowledge of the explicit sources, a hidden source is learned by excavating correlations and potential knowledge behind them, facilitating the generalization enhancement in generating HIPs. The experimental results especially performance comparisons with seven state-of-the-art DMOAs demonstrate that DMOA-MHKT brings significant improvements in solving DMOPs.

Design change is an important issue in complex product development projects. In a complex product with numerous parts (also known as components), the change of one key part may spread to other parts associated with it, generating a chain reaction throughout the entire project. Therefore, it is necessary to select a suitable change plan involving only fewer crucial parts in order to enhance the product’s performance, minimize change cost, and reduce change duration/time. Focusing on the case where the correlation strength between parts cannot be accurately obtained, in this paper we study an interval multi-objective evolutionary algorithm for finding excellent design change plans. Firstly, on the basis of the established multi-layer product network with interval correlation weights, an interval multi-objective optimization model of the product design change planning problem is established, where three new objective functions regarding product performance, carbon trading cost and supply risk are defined. Then, a constraint multi-objective evolutionary algorithm based on interval Pareto dominance is proposed to search for optimal change plans. Several novel operators, including the problem characteristic-guided population update strategy, the probability-based interval Pareto dominance, and the interval constraint handling strategy, are developed to enhance the algorithm’s performance. Finally, the proposed algorithm is compared with eight existing algorithms on the two design change cases, experimental results revealed its effectiveness.

The interactive evolutionary multi-task optimization approach assisted by surrogate models has proven successful in enhancing individualized recommendation performance. However, in light of the dynamically changing user preferences, it becomes imperative to further develop more powerful knowledge sharing strategy to improve the interactive multi-task optimization efficiency. A probability model-assisted search space alignment among multiple tasks method is proposed for knowledge transfer across multi-task environment. Additionally, a diversity maintaining mechanism of the transferred population is proposed to improve the quality and diversity of the initial population in a preference varied multi-task environment. The research demonstrates the effectiveness of the proposed approach in achieving effective knowledge transfer across multi-task environmen. Furthermore, the utilization of a high-quality initial population not only enhances the evolutionary search efficiency of the algorithm but also significantly improves the accuracy, diversity., and novelty of personalized recommendation.

The operation optimization problem of coal mine integrated energy system (CMIES) is characterized by multiobjective, strong constraints, large scale, and mixed variables. It is difficult for existing multiobjective evolutionary algorithms to obtain a set of nondominated solutions with good convergence and uniform distribution, primarily due to the absence of suitable constraint-handling techniques. This research proposes a multitask multiobjective operation optimization framework combining evolutionary algorithm and mathematical programming (MO-EAMP) to address this issue. Within this framework, the main task employs an evolutionary algorithm with global search capability to solve the multiobjective CMIES operation optimization problem. Meanwhile, auxiliary tasks utilize mathematical programming method with robust linear constraint handling capability to solve multiple weighted single-objective CMIES operation optimization problems. During the iteration process of MO-EAMP, the scale and form of auxiliary tasks are adjusted autonomously based on the current state of population, with the aim of guiding the population search toward more promising regions. Finally, the presented algorithm is applied to a coal mine in Shanxi Province, China, and the experimental results demonstrate that the proposed algorithm can obtain a set of optimal operation plans with better convergence and distribution in a shorter time, compared with 7 other existing algorithms.

An intelligent optimization algorithm is only valid for solving some problems, which difficultly solves all operation optimization problems of integrated coal mine energy systems with different characteristics. An optimization paradigm usually consists of multiple operators/strategies, each of which is suitable for solving different problems. It is difficult for an operator/strategy to ensure that the population can evolve forward in the evolution process since the state of the population is changing. To this end, a multi-granularity autonomous intelligent optimization method is proposed to optimize the operation of integrated coal mine energy systems with various scenarios. This method automatically determines appropriate optimization paradigms according to problem characteristics and adaptively adjusts optimization operators/strategies based on population states in the evolution process. For the adaptive adjustment of operators/strategies, this paper proposes an adaptive adjustment strategy based on Q-Learning. Taking a coal mine in Shanxi Province, China as the research object, a series of experiments are conducted, and the experimental results show the effectiveness of the proposed algorithm.

The high prevalence of mental disorders gradually poses a huge pressure on the public healthcare services. Deep learning-based computer-aided diagnosis (CAD) has emerged to relieve the tension in healthcare institutions by detecting abnormal neuroimaging-derived phenotypes. However, training deep learning models relies on sufficient annotated datasets, which can be costly and laborious. Semi-supervised learning (SSL) and transfer learning (TL) can mitigate this challenge by leveraging unlabeled data within the same institution and advantageous information from source domain, respectively. This work is the first attempt to propose an effective semi-supervised transfer learning (SSTL) framework dubbed S3TL for CAD of mental disorders on fMRI data. Within S3TL, a secure cross-domain feature alignment method is developed to generate target-related source model in SSL. Subsequently, we propose an enhanced dual-stage pseudo-labeling approach to assign pseudo-labels for unlabeled samples in target domain. Finally, an advantageous knowledge transfer method is conducted to improve the generalization capability of the target model. Comprehensive experimental results demonstrate that S3TL achieves competitive accuracies of 69.14%, 69.65%, and 72.62% on ABIDE-I, ABIDE-II, and ADHD-200 datasets, respectively. Furthermore, the simulation experiments also demonstrate the application potential of S3TL through model interpretation analysis and federated learning extension.

Many real-world applications can be formulated as expensive multimodal optimization problems (EMMOPs). When surrogate-assisted evolutionary algorithms (SAEAs) are employed to tackle these problems, they not only face the problem of selecting surrogate models but also need to tackle the problem of discovering and updating multiple modalities. Different optimization problems and different stages of evolutionary algorithms (EAs) generally require different types of surrogate models. To address this issue, in this article, we present a multisurrogate-assisted multitasking particle swarm optimization algorithm to seek multiple optimal solutions of EMMOPs at a low computational cost. The proposed algorithm first transforms an EMMOP into a multitasking optimization problem by integrating various surrogate models, and designs a multitasking niche particle swarm algorithm to solve it. Following that, a surrogate model management strategy based on the skill factor and clustering is developed to effectively balance the number of real function evaluations and the prediction accuracy of candidate optimal solutions. In addition, an adaptive local search strategy based on the trust region is proposed to enhance the capability of swarm in exploiting potential optimal modalities. We compare the proposed algorithm with five state-of-the-art SAEAs and seven multimodal EAs on 19 benchmark functions and the building energy conservation problem and experimental results show that the proposed algorithm can obtain multiple highly competitive optimal solutions.

Expensive constraint multimodal optimization problems (ECMMOPs) have such characteristics as expensive objectives and constraints, and multiple optimal modalities simultaneously, which pose severe challenges to evolutionary optimization methods. This article studies an objective-constraint mutual-guided surrogate-assisted particle swarm optimization algorithm for the kind of problem, aiming to discover multiple competing feasible optimal solutions at a lower calculation cost. The algorithm designs first a new two-layer cooperative surrogate model framework based on heterogeneous database to effectively adjust the prediction accuracies of objective surrogates and constraint surrogates on different search regions. An objective-constraint mutual-guided partial evaluation strategy (O-C-PES) is developed to generate high-quality infilling samples for objective and constraint surrogates, respectively, based on which the number of unnecessary real evaluations can be significantly reduced. Moreover, a position feature-guided hybrid update mechanism (PF-HUM) is proposed to find more optimal solutions by searching excellent infeasible and feasible areas at the same time, and a feasible ratio-driven local search (FR-LS) strategy is proposed to improve the algorithm’s exploitation. Compared with four existing surrogate-assisted evolutionary algorithms (EAs) and one constraint multimodal EAs on 21 benchmark problems and three engineering instances, experiment results show that the proposed algorithm can simultaneously obtain multiple highly-competitive feasible optimal solutions with less computational cost.

Personalized search is essentially a complex qualitative optimization problem, and interactive evolutionary algorithms (EAs) have been extended from EAs to adapt to solving it. However, the multisource user-generated contents (UGCs) in the personalized services have not been concerned on in the adaptation. Accordingly, we here present an enhanced restricted Boltzmann machine (RBM)-driven interactive estimation of distribution algorithms (IEDAs) with multisource heterogeneous data from the viewpoint of effectively extracting users’ preferences and requirements from UGCs to strengthen the performance of IEDA for personalized search. The multisource heterogeneous UGCs, including users’ ratings and reviews, items’ category tags, social networks, and other available information, are sufficiently collected and represented to construct an RBM-based model to extract users’ comprehensive preferences. With this RBM, the probability model for conducting the reproduction operator of estimation of distribution algorithms (EDAs) and the surrogate for quantitatively evaluating an individual (item) fitness are further developed to enhance the EDA-based personalized search. The UGCs-driven IEDA is applied to various publicly released Amazon datasets, e.g., recommendation of Digital Music, Apps for Android, Movies, and TV, to experimentally demonstrate its performance in efficiently improving the IEDA in personalized search with less interactions and higher satisfaction.

Band selection (BS) is a widely used dimensionality reduction technique for hyperspectral images. However, most of existing evolutionary algorithms focus on searching a globally optimal band subset under a fixed size, and their obtained band subsets may still contain a large number of redundant bands. In order to simultaneously obtain multiple optimal band subsets with different sizes, this article proposes an unsupervised multitask artificial bee colony (ABC) BS algorithm based on variable-size clustering (MBBS-VC). First, a variable-size band clustering method based on worst class decomposition is developed, based on which the BS problem can be modeled as a multitask optimization problem. Next, a multitask multimicrogroup bee colony algorithm with variable coding length is proposed to simultaneously search multiple optimal band subsets with different sizes. Moreover, several new strategies, including the intergroup collaboration strategy based on bidirectional neighborhood learning and the multimeasure integration judgment (MIJ) mechanism, are designed to improve the performance of MBBS-VC. In this article, the hyperspectral BS problem is transformed into a multitask optimization problem for the first time. Finally, compared with 15 classical BS algorithms on several commonly used datasets, experimental results verify the superiority of the proposed BS algorithm.

Feature selection (FS) in data with class imbalance or missing values has received much attention from researchers due to their universality in real-world applications. However, for data with both the two characteristics above, there is still a lack of the corresponding FS algorithm. Due to the complex coupling relationship between missing data and class imbalance, the need for better FS method becomes essential. To tackle high-dimensional imbalanced data with missing values, this article studies a new evolutionary FS method. First, an improved$F$-measure based on filling risk (RF-measure) is defined to evaluate the influence of missing data on the performance of FS in the case of class imbalance. Following that taking the RF-measure as an objective function, a particle swarm optimization-based FS method with fuzzy clustering (PSOFS-FC) is proposed. Two new problem-specific operators or strategies, i.e., the swarm initialization strategy guided by fuzzy clustering and the local pruning operator based on feature importance, are developed to improve the performance of PSOFS-FC. Compared with state-of-the-art FS algorithms on several public datasets, experimental results show that PSOFS-FC can achieve excellent classification performance with relatively less running time, indicating its superiority on tackling high-dimensional imbalanced data with missing values.

In recent years, the application of multi-objective evolutionary algorithms (MOEAs) to overlapping community detection in complex networks has been a hot research topic. However, the existing MOEAs for detecting overlapping communities show poor scalability to large-scale networks due to the fact that the encoding length of individuals is usually equal to the number of all nodes in the network. To this end, we suggest a reduced mixed representation based multi-objective evolutionary algorithm named RMR-MOEA for large-scale overlapping community detection, where the length of the individual is recursively reduced as the evolution proceeds. Specifically, a mixed representation is adopted for fast encoding and decoding the individual in the population, which consists of two parts: one represents all potential overlapping nodes and the other represents all non-overlapping nodes. Then, in each individual length reduction, two strategies are suggested to respectively shorten the length of each part in the mixed representation, with the aim to greatly reduce the search space. Finally, the experimental results on 10 real-world complex networks demonstrate the effectiveness of the proposed RMR-MOEA in terms of both detection performance and running time, especially on large-scale networks.

Various real-world applications can be classified as expensive multimodal optimization problems. When surrogate-assisted evolutionary algorithms (SAEAs) are employed to tackle these problems, they not only face a contradiction between the precision of surrogate models and the cost of individual evaluations but also have the difficulty that surrogate models and problem modalities are hard to match. To address this issue, this article studies a dual-surrogate-assisted cooperative particle swarm optimization algorithm to seek multiple optimal solutions. A dual-population cooperative particle swarm optimizer is first developed to simultaneously explore/exploit multiple modalities. Following that, a modal-guided dual-layer cooperative surrogate model, which contains one upper global surrogate model and a group of lower local surrogate models, is constructed with the purpose of reducing the individual evaluation cost. Moreover, a hybrid strategy based on clustering and peak-valley is proposed to detect new modalities. Compared with five existing SAEAs and seven multimodal evolutionary algorithms, the proposed algorithm can simultaneously obtain multiple highly competitive optimal solutions at a low computational cost according to the experimental results of testing both 11 benchmark instances and the building energy conservation problem.

Interactive Estimation of Distribution Algorithm (IEDA), by integrating users interactions with Estimation of Distribution Algorithm, is powerful for efficient personalized search when the probability model and fitness function are well designed. We here propose an improved IEDA by using attention mechanism strengthened Restricted Boltzmann Machine (RBM). An attention mechanism assisted RBM model is constructed to approximate the user preferences by inputting item features and user generated contents. Then the attention-enhanced probability model of EDA and the fitness function are developed based on the RBM. In the evolutionary process, the attention-based RBM together with the probability model and fitness function are managed according to new interactions and corresponding information. The proposed algorithm is applied to real-world Amazon data sets usually used in the personalized search or recommendation, and its performance is experimentally demonstrated in better predicting the user preferences to improve the searching efficiency and accuracy.

Influence maximization refers to selecting a group of nodes from a social network, which obtains the largest influence spread under a cascade model. However, most of the existing works only focused on the influence and ignored the diversity of influenced crowd. Thus, scholars have raised the issue of diversified social influence maximization recently, using the category information of nodes to design diversity indicator and introducing a trade-off parameter to balance the two objectives influence and diversity as one single objective for optimization. In fact, the category information of nodes in the network is usually difficult to be collected, thus the definition of diversity based on nodes' categories is not very general and accurate. In addition, it is very difficult to set the trade-off parameter, especially when there is no prior knowledge in real applications. To this end, we employ overlapping community structure information to design the diversity of nodes without any node's additional (e.g. category) information. Due to the two objectives of influence and diversity may be conflicting, a multi-objective evolutionary algorithm named MOEA-DIM is proposed to optimize the two objectives simultaneously, which does not need to set the tradeoff parameter between the two objectives. In MOEA-DIM, a network reduction strategy based on overlapping community structure is suggested to greatly reduce the search space. In addition, a population initialization strategy based on random walk is designed to accelerate the convergence of the algorithm. Experiments on six real-world datasets show that the proposed algorithm MOEA-DIM has promising performance in terms of both effectiveness and efficiency.

Evolutionary feature selection (FS) methods face the challenge of “curse of dimensionality” when dealing with high-dimensional data. Focusing on this challenge, this article studies a variable-size cooperative coevolutionary particle swarm optimization algorithm (VS-CCPSO) for FS. The proposed algorithm employs the idea of “divide and conquer” in cooperative coevolutionary approach, but several new developed problem-guided operators/strategies make it more suitable for FS problems. First, a space division strategy based on the feature importance is presented, which can classify relevant features into the same subspace with a low computational cost. Following that, an adaptive adjustment mechanism of subswarm size is developed to maintain an appropriate size for each subswarm, with the purpose of saving computational cost on evaluating particles. Moreover, a particle deletion strategy based on fitness-guided binary clustering, and a particle generation strategy based on feature importance and crossover both are designed to ensure the quality of particles in the subswarms. We apply VS-CCPSO to 12 typical datasets and compare it with six state-of-the-art methods. The experimental results show that VS-CCPSO has the capability of obtaining good feature subsets, suggesting its competitiveness for tackling FS problems with high dimensionality.

Federated learning obtains a central model on the server by aggregating models trained locally on clients. As a result, federated learning does not require clients to upload their data to the server, thereby preserving the data privacy of the clients. One challenge in federated learning is to reduce the client-server communication since the end devices typically have very limited communication bandwidth. This article presents an enhanced federated learning technique by proposing an asynchronous learning strategy on the clients and a temporally weighted aggregation of the local models on the server. In the asynchronous learning strategy, different layers of the deep neural networks (DNNs) are categorized into shallow and deep layers, and the parameters of the deep layers are updated less frequently than those of the shallow layers. Furthermore, a temporally weighted aggregation strategy is introduced on the server to make use of the previously trained local models, thereby enhancing the accuracy and convergence of the central model. The proposed algorithm is empirically on two data sets with different DNNs. Our results demonstrate that the proposed asynchronous federated deep learning outperforms the baseline algorithm both in terms of communication cost and model accuracy.

Crowdfunding is an emerging Internet fundraising platform in which creators post descriptions of their projects and investors glance over these projects to support or not. With the increasing amount of projects posted in the crowdfunding platform, it is necessary to develop personalized recommendation systems (RSs) by suggesting suitable projects to crowdfunding investors. In this paper, we propose a personalized recommender system for crowdfunding platform to make accurate, high profitable and diverse project recommendations for crowdfunding investors. Specifically, the task of personalized recommendation for crowdfunding platform is modeled as a multi-objective optimization problem. The proposed model maximizes two conflicting performance metrics named as utility-accuracy and topic-diversity. The utility-accuracy is obtained by the probabilistic spreading method, while the topic-diversity is evaluated by recommendation coverage. Then, a multi-objective evolutionary algorithm for personalized recommendation in crowdfunding platform (termed as MOEA-PRCP) is proposed for the two-objective optimization problem. In MOEA-PRCP, a novel initialization strategy is designed for speeding the convergence of the proposed algorithm. Extensive experiments are conducted on a real-world crowdfunding data collected from Indiegogo.com, and the experimental results clearly demonstrate the effectiveness of MOEA-PRCP for personalized recommendation in crowdfunding platform.

Interactive evolutionary algorithms have been applied to personalized search, in which less user fatigue and efficient search are pursued. Motivated by this, we present a fast interactive estimation of distribution algorithm (IEDA) by using the domain knowledge of personalized search. We first induce a Bayesian model to describe the distribution of the new user's preference on the variables from the social knowledge of personalized search. Then we employ the model to enhance the performance of IEDA in two aspects, that is: 1) dramatically reducing the initial huge space to a preferred subspace and 2) generating the individuals of estimation of distribution algorithm(EDA) by using it as a probabilistic model. The Bayesian model is updated along with the implementation of the EDA. To effectively evaluate individuals, we further present a method to quantitatively express the preference of the user based on the human-computer interactions and train a radial basis function neural network as the fitness surrogate. The proposed algorithm is applied to a laptop search, and its superiorities in alleviating user fatigue and speeding up the search procedure are empirically demonstrated.

Multi-objective optimization problems with interval (MOPs-I) uncertainties parameters are common in practice. Evolutionary multi-objective (EMO) algorithms are popularly employed to solve these problems due to their powerful explorations. The comparison strategies among interval objectives of MOPs-I are crucially important for obtaining a superior Pareto front when applying EMOs. By effectively combining two different intervals ranking methods together, i.e., μ and P metrics, we present an improved NSGA-II with a synthesized intervals ranking strategy for optimizing MOPs-I. The characteristics of μ and P in ranking intervals are first analyzed, and then the synthesized ranking method termed as μ ⊕ P is developed to compare and select individuals within the NSGA-II framework. The proposed algorithm is experimentally validated by four MOPs-I functions and a practical problem, and the results empirically demonstrate its merits in obtaining Pareto front with outstanding convergence and spread.

Surrogate-assisted interactive genetic algorithms (IGAs) are found to be very effective in reducing human fatigue. Different from models used in most surrogate-assisted evolutionary algorithms, surrogates in IGA must be able to handle the inherent uncertainties in fitness assignment by human users, where, e.g., interval-based fitness values are assigned to individuals. This poses another challenge to using surrogates for fitness approximation in evolutionary optimization, in addition to the lack of training data. In this paper, a new surrogate-assisted IGA has been proposed, where the uncertainty in subjective fitness evaluations is exploited both in training the surrogates and in managing surrogates. To enhance the approximation accuracy of the surrogates, an improved cotraining algorithm for semisupervised learning has been suggested, where the uncertainty in interval-based fitness values is taken into account in training and weighting the two cotrained models. Moreover, uncertainty in the interval-based fitness values is also considered in model management so that not only the best individuals but also the most uncertain individuals will be chosen to be re-evaluated by the human user. The effectiveness of the proposed algorithm is verified on two test problems as well as in fashion design, a typical application of IGA. Our results indicate that the new surrogate-assisted IGA can effectively alleviate user fatigue and is more likely to find acceptable solutions in solving complex design problems.

Interactive genetic algorithms (IGAs) are effective methods of solving optimization problems with qualitative indices. The problem of user fatigue resulting from his/her evaluations, however, restricts their applications in complex optimization problems. Employing various surrogate models to evaluate (a part of) individuals instead of a user is a feasible approach to solving the above problem. Previous studies, however, have not fully utilized knowledge provided by users with similar preference when constructing these models. The problem of constructing surrogate models by using knowledge of users with similar preference was focused in this study. First, users with similar preference participating the evolution were identified based on the matrix formed by the relationship between users and the “fitness” of allele meaning units and the users' interests in allele meaning units by using the collaborative filtering algorithm based on nearest-neighbor; and then the individuals evaluated by users with similar preference and chosen according to the users' preference similarities and confidence, along with their fitness, were as a part of samples for training the surrogate model of the current user's cognition. The proposed method was applied to an evolutionary fashion design system, and the experimental results show that the proposed method can improve the capability in exploration on the premise of greatly alleviating user fatigue.

Interactive genetic algorithms (IGAs) have been successfully applied to optimize problems with aesthetic criteria by embedding the intelligent evaluations of a user into the evolutionary process. User fatigue caused by frequent interactions, however, often greatly impairs the potentials of IGAs on solving complicated optimization problems. Taking the benefits of collective intelligence into account, we here present an IGA with collective intelligence which is derived from a mechanism of group decision making. An IGA with interval individual fitness is focused here and it can be separately conducted by multiple users at the same time. The collective intelligence of all participated users, represented with social and individual knowledge, is first collected by using a modified group decision making method. Then the strategy of applying the collective intelligence to initialize and guide the single evolution of the IGA is given. With such a multi-user promoted IGA framework, the performance of a single IGA is expected to be evidently improved. In a local network environment, the algorithm is applied to a fashion design system and the results empirically demonstrate that the algorithm can not only alleviate user fatigue but also increase the opportunities of IGAs on finding most satisfactory solutions.

Multi-objective optimization (MOO) problems with interval parameters are popular and important in real-world applications. Previous evolutionary optimization methods aim to find a set of well-converged and evenly-distributed Pareto-optimal solutions. We present a novel evolutionary algorithm (EA) that interacts with a decision maker (DM) during the optimization process to obtain the DM's most preferred solution. First, the theory of a preference polyhedron for an optimization problem with interval parameters is built up. Then, an interactive evolutionary algorithm (IEA) for MOO problems with interval parameters based on the above preference polyhedron is developed. The algorithm periodically provides a part of non-dominated solutions to the DM, and a preference polyhedron, based on which optimal solutions are ranked, is constructed with the worst solution chosen by the DM as the vertex. Finally, our method is tested on two bi-objective optimization problems with interval parameters using two different value function types to emulate the DM's responses. The experimental results show its simplicity and superiority to the posteriori method.

In order to alleviate user fatigue of interactive genetic algorithms with an individual's fuzzy and stochastic fitness, we propose a surrogate model-assisted algorithm by using a directed fuzzy graph to extract user cognition. According to cut-set level and interval dominance probability, we present approaches to construct a directed fuzzy graph of an evolutionary population and calculate an individual's precise fitness based on it. By applying the fuzzy entropy, the chance of data sampling is achieved to obtain reliable samples for training the surrogate model. We adopt a support vector regression machine as the surrogate model, train it using the sampled individuals and their precise fitness, and apply a traditional genetic algorithm to optimize the surrogate model for some generations, providing guided individuals to the user to accelerate the evolution. We quantitatively analyze the performance of the presented algorithm in alleviating user fatigue and increasing more opportunities to look for the satisfactory individuals. Finally, we apply our algorithm to a fashion evolutionary design system to demonstrate its efficiency.

Interactive genetic algorithms with individual's fuzzy fitness well portray the fuzzy uncertainties of a user's cognition. In this paper, we propose an efficient surrogate model-assisted one to alleviate user fatigue by building a classifier and a regressor to approximate the user's cognition. Two reliable training data sets are obtained based on the user's evaluation credibility. Then a support vector classification machine and a support vector regression machine are trained as the surrogate models with these samples. Specifically, the input trained samples are the individuals evaluated by the user, and the output training samples of the classifier and the regressor are widths and centers of these individuals' fuzzy fitness assigned by the user, respectively. These two surrogate models are simultaneously applied to the subsequent evolutions with enlarged population size so as to alleviate user fatigue and enhance the search ability of the algorithm. We constantly update the training data sets and the surrogate models in order to guarantee the approximation precision. Furthermore, we quantitatively analyze the algorithm's performance in alleviating user fatigue and increasing more opportunities to find the optimal solutions. We also apply it to a fashion evolutionary design system to show its efficiency.