Article
Automation & Control Systems
Kai Zhang, Chaonan Shen, Gary G. Yen
Summary: This article proposes a multipopulation-based differential evolution algorithm, called LSMaODE, to efficiently and effectively solve large-scale many-objective optimization problems. The algorithm divides the population into two groups of subpopulations and applies different optimization strategies. The performance is evaluated in both decision and objective dimensions.
IEEE TRANSACTIONS ON CYBERNETICS
(2022)
Article
Automation & Control Systems
Vikas Palakonda, Jae-Mo Kang
Summary: This article proposes a preference-inspired differential evolution algorithm for multi and many-objective optimization, which effectively deals with a wide range of problems. The algorithm generates individuals with good convergence and distribution properties by utilizing a preference-inspired mutation operator and determining local knee points based on a clustering method. Experimental results demonstrate its superior performance compared to eight state-of-the-art algorithms on 35 benchmark problems.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2023)
Article
Operations Research & Management Science
Djaafar Zouache, Fouad Ben Abdelaziz
Summary: In this study, a guided differential evolution algorithm is proposed to solve many-objective optimization problems by using strengthened dominance relation and bi-goal evolution. The guided search strategy, which utilizes adapted differential evolutionary operators for crossover and mutation, allows convergence towards the Pareto front with good solution diversity.
ANNALS OF OPERATIONS RESEARCH
(2022)
Article
Computer Science, Artificial Intelligence
Hongli Bian, Jie Tian, Jialiang Yu, Han Yu
Summary: A novel Entropy Search-based Bayesian Co-Evolutionary Optimization approach (ESB-CEO) is proposed to address the limitations of existing approaches in solving high-dimensional many-objective optimization problems. The ESB-CEO algorithm combines an adaptive acquisition function with the Lp-norm and information entropy to efficiently identify individual solutions that have a significant effect on different search stages, improving the convergence and diversity of the algorithm.
KNOWLEDGE-BASED SYSTEMS
(2023)
Article
Computer Science, Information Systems
Peng Wang, Bing Xue, Jing Liang, Mengjie Zhang
Summary: By identifying relevant features, feature selection methods can maintain or improve classification accuracy and reduce dimensionality. This paper proposes a diversity-based multi-objective differential evolution approach to effectively handle the trade-offs between convergence and diversity. The method detects and removes irrelevant and weakly relevant features to reduce the search space and proposes a new binary mutation operator to produce better feature subsets. Experimental results show that the proposed method outperforms current popular multi-objective feature selection methods on 14 datasets with varying difficulty.
INFORMATION SCIENCES
(2023)
Article
Engineering, Multidisciplinary
Ying Hou, YiLin Wu, Zheng Liu, HongGui Han, Pu Wang
Summary: The DMODE-IEP algorithm improves optimization performance through dynamic adjustment based on evolution progress information. The convergence of the algorithm is proved using probability theory, and testing results demonstrate its superiority in optimization effectiveness compared to other multi-objective optimization algorithms.
SCIENCE CHINA-TECHNOLOGICAL SCIENCES
(2021)
Article
Computer Science, Artificial Intelligence
Xinye Cai, Yushun Xiao, Zhenhua Li, Qi Sun, Hanchuan Xu, Miqing Li, Hisao Ishibuchi
Summary: This article proposes a kernel-based indicator (KBI) for evaluating Pareto front approximations generated by multi/many-objective optimizers. Unlike other distance-based indicators, KBI not only reflects the distance between the solution set and the reference set but also can reflect the distribution of the solution set itself. In addition, a non-dominated set reconstruction approach is proposed to maintain the desirable weak Pareto compliance property of KBI.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2022)
Article
Computer Science, Theory & Methods
Yan-Yang Cheng, Zheng-Yi Chai, Ya-Lun Li
Summary: Multi-task optimization uses knowledge transfer to optimize multiple tasks simultaneously. However, when the number of tasks increases to many-task optimization, the algorithm faces computational burden and degradation of performance due to decreased positive knowledge transfer rate. Existing many-task optimization algorithms have issues in high-dimensional objective space, such as decreased population diversity and slowed optimal solution search speed. To address many-objective and many-task optimization problems, we propose a reference-points-based nondominated sorting approach called MOMaTO-RP. MOMaTO-RP enables knowledge transfer from multiple highly similar tasks and maintains population diversity in high-dimensional objective space, resulting in improved convergence speed. The algorithm is compared to other related algorithms on a classical benchmark set, showing faster convergence speed and better distribution performance.
FUTURE GENERATION COMPUTER SYSTEMS-THE INTERNATIONAL JOURNAL OF ESCIENCE
(2023)
Article
Multidisciplinary Sciences
Mingwei Fan, Jianhong Chen, Zuanjia Xie, Haibin Ouyang, Steven Li, Liqun Gao
Summary: In this paper, an improved multi-objective differential evolution algorithm (MOEA/D/DEM) based on a decomposition strategy is proposed to enhance the search performance for practical multi-objective nutrition decision problems. The algorithm utilizes a neighborhood intimacy factor and a new Gaussian mutation strategy to improve diversity and local search ability. Experimental results show that the proposed algorithm achieves better search capability and obtains competitive results compared to other multi-objective algorithms.
SCIENTIFIC REPORTS
(2022)
Article
Computer Science, Artificial Intelligence
Genggeng Liu, Zhenyu Pei, Nengxian Liu, Ye Tian
Summary: This paper proposes a subspace segmentation based co-evolutionary algorithm for balancing convergence and diversity in many-objective optimization. The decision space is divided into a convergence subspace and a diversity subspace, which are searched in different stages to improve the population convergence and diversity. The algorithm retains elite individuals with better convergence using an adaptively adjusted archive and maintains population diversity by retaining boundary individuals using a penalty factor-based indicator. Experimental results show that the proposed algorithm can balance convergence and diversity well and exhibit competitiveness.
SWARM AND EVOLUTIONARY COMPUTATION
(2023)
Article
Automation & Control Systems
Wenbo Qiu, Jianghan Zhu, Guohua Wu, Huangke Chen, Witold Pedrycz, Ponnuthurai Nagaratnam Suganthan
Summary: The research proposes a general voting-mechanism-based ensemble framework (VMEF) that integrates and cooperates different solution-sorting methods to achieve more robust solution selection. The framework employs a strategy to adaptively allocate total votes based on the contribution of each solution-sorting method, providing good feedback for the optimization process.
IEEE TRANSACTIONS ON SYSTEMS MAN CYBERNETICS-SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Qian Li, Sanyang Liu, Yiguang Bai, Xingshi He, Xin-She Yang
Summary: This paper investigates the robustness of complex networks under the assumption that costs are functions of node degrees. A multi-objective, elitism-based evolutionary algorithm is proposed to address the network disintegration problem. Through information retention and an update mechanism, the algorithm achieves improved convergence rate. Experimental results demonstrate that the proposed method outperforms five other state-of-the-art attack strategies.
KNOWLEDGE-BASED SYSTEMS
(2022)
Article
Computer Science, Artificial Intelligence
Qi Deng, Qi Kang, Liang Zhang, MengChu Zhou, Jing An
Summary: This work proposes an objective space-based population generation method to obtain new individuals in the objective space and then map them to decision variable space and synthesize new solutions. It introduces three new objective vector generation methods and uses a linear mapping method to tightly connect objective space and decision one to jointly determine new-generation solutions. The proposed method exceeds or at least reaches its peers' best level in overall performance while achieving great saving in execution time.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2023)
Article
Computer Science, Artificial Intelligence
Zhenan He, Gary G. Yen, Jinliang Ding
Summary: A novel knee-based decision-making method is proposed to search for solutions of interest from a large number of solutions on the Pareto front, ensuring the performance of these solutions approximates as much as possible the whole Pareto front. Additionally, a new visualization approach is developed to provide information about the shape, location, possible bulge, convergence degree, and distribution of solutions on MaOPs.
IEEE TRANSACTIONS ON EVOLUTIONARY COMPUTATION
(2021)
Article
Automation & Control Systems
Kai Zhang, Zhiwei Xu, Shengli Xie, Gary G. Yen
Summary: In this article, a new evolution strategy called MaOES is proposed to address the challenges faced by existing MaOEAs in solving MaOPs. Inspired by the Vector Equilibrium phenomenon, MaOES efficiently solves diversity preservation and dominance resistance issues using self-adaptive mutation and maximum extension distance strategies, resulting in a well-converged and well-diversified Pareto front.
IEEE TRANSACTIONS ON CYBERNETICS
(2021)
Article
Biochemical Research Methods
Saurav Mallik, Sanghamitra Bandyopadhyay
IEEE-ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS
(2020)
Article
Biology
Angana Chakraborty, Sanghamitra Bandyopadhyay
COMPUTATIONAL BIOLOGY AND CHEMISTRY
(2020)
Article
Materials Science, Coatings & Films
Arpan Mukherjee, Scott Broderick, Krishna Rajan
JOURNAL OF VACUUM SCIENCE & TECHNOLOGY A
(2020)
Article
Chemistry, Physical
Logan Williams, Arpan Mukherjee, Krishna Rajan
JOURNAL OF PHYSICAL CHEMISTRY LETTERS
(2020)
Article
Computer Science, Artificial Intelligence
Snehalika Lall, Debajyoti Sinha, Abhik Ghosh, Debarka Sengupta, Sanghamitra Bandyopadhyay
Summary: The study introduces a feature selection algorithm based on copula that maximizes feature relevance and minimizes redundant information. The proposed CBFS algorithm competes well in maximizing classification accuracy on real and synthetic datasets and demonstrates better noise tolerance compared to other methods.
PATTERN RECOGNITION
(2021)
Article
Computer Science, Artificial Intelligence
Monalisa Pal, Sanghamitra Bandyopadhyay
Summary: This paper introduces an evolutionary framework called LORD for addressing multi-modal multi-objective optimization problems (MMMOPs), which uses decomposition in both objective and decision space. The LORD-II algorithm further extends this framework, demonstrating its dynamics on multi-modal many-objective problems. The efficacy of the frameworks is established through performance comparisons with other algorithms.
SWARM AND EVOLUTIONARY COMPUTATION
(2021)
Article
Biochemical Research Methods
Angana Chakraborty, Burkhard Morgenstern, Sanghamitra Bandyopadhyay
Summary: The newly developed S-conLSH mapping tool uses spaced-context based Locality Sensitive Hashing to achieve faster mapping speed and higher sensitivity on 5 different real and simulated datasets. By utilizing multiple spaced patterns, S-conLSH enables gapped mapping of noisy long reads to the corresponding target locations of a reference genome, making it a promising direction towards alignment-free sequence analysis.
BMC BIOINFORMATICS
(2021)
Article
Computer Science, Artificial Intelligence
Priya Roy, Chandreyee Chowdhury, Mausam Kundu, Dip Ghosh, Sanghamitra Bandyopadhyay
Summary: Indoor localization systems using WiFi signals face challenges due to the significant variation of signal strength with ambient conditions and device configuration. This paper proposes a weighted ensemble classifier based on Dempster-Shafer belief theory to efficiently handle context heterogeneity. Real life experiments show that the technique achieves high localization accuracy at varying granularity levels.
EXPERT SYSTEMS WITH APPLICATIONS
(2021)
Article
Biochemical Research Methods
Sourav Biswas, Sumanta Ray, Sanghamitra Bandyopadhyay
Summary: This article introduces the concepts of network motifs and colored motifs, as well as a method to store colored subgraphs and discover colored motifs using a modified G-trie data structure. The approach utilizes approximate enumeration to reduce runtime and has been applied to find colored motifs in a host pathogen protein-protein interaction network. The study discovered eight motifs, with a majority containing both HIV-1 and human proteins.
IEEE-ACM TRANSACTIONS ON COMPUTATIONAL BIOLOGY AND BIOINFORMATICS
(2021)
Article
Chemistry, Medicinal
Arpan Mukherjee, An Su, Krishna Rajan
Summary: This paper introduces a deep neural network-based toolkit to identify structural motifs within a molecule that contribute to a chemical being an endocrine disruptor. The toolkit combines convolution and long short-term memory (LSTM) architectures and utilizes an active learning-based framework with multiple data sources. Class activation maps (CAMs) generated from feature-extraction layers can pinpoint structural alerts and the chemical environment affecting their specificity.
JOURNAL OF CHEMICAL INFORMATION AND MODELING
(2021)
Article
Chemistry, Physical
Deepesh Giri, Logan Williams, Arpan Mukherjee, Krishna Rajan
Summary: This paper describes a new data-driven framework for computational screening and discovery of metavalent solids, introducing the use of Hirshfeld surface analysis for rapid identification of potential metavalent solids with novel properties.
JOURNAL OF CHEMICAL PHYSICS
(2021)
Article
Biochemical Research Methods
Snehalika Lall, Sumanta Ray, Sanghamitra Bandyopadhyay
Summary: The study introduces a method RgCop based on regularized copula for stable and predictive gene selection in large-scale single cell RNA sequencing data, improving clustering/classification performance and enhancing the robustness of the method.
PLOS COMPUTATIONAL BIOLOGY
(2021)
Article
Biochemical Research Methods
Snehalika Lall, Sumanta Ray, Sanghamitra Bandyopadhyay
Summary: Annotation of cells in single-cell clustering requires a homogeneous grouping of cell populations. sc-CGconv is an unsupervised feature extraction and clustering approach that utilizes copula correlation and graph convolution network to formulate and aggregate cell-cell relationships, which can identify homogeneous clusters with small sample sizes, model the expression co-variability of a large number of genes, preserve cell-to-cell variability, and provide a topology-preserving embedding of cells in low dimensional space.
PLOS COMPUTATIONAL BIOLOGY
(2022)
Article
Computer Science, Artificial Intelligence
Sumanta Ray, Snehalika Lall, Anirban Mukhopadhyay, Sanghamitra Bandyopadhyay, Alexander Schoenhuth
Summary: This article introduces the use of artificial intelligence and deep learning techniques to screen drug repositories and find therapeutic options against COVID-19. By constructing a comprehensive molecular interaction network and predicting connections between drugs and human proteins, novel host-directed therapy options are established, providing a new approach for fighting the virus.
ARTIFICIAL INTELLIGENCE IN MEDICINE
(2022)
Article
Geochemistry & Geophysics
Monidipa Das, Soumya K. Ghosh, Sanghamitra Bandyopadhyay
Summary: This article proposes a MARINE model to address the catastrophic forgetting issue that neural networks encounter when trained in a sequential manner, particularly in the presence of a large degree of subregional variations or heterogeneity in spatial zones. MARINE demonstrates competitive results in spatio-temporal prediction tasks and outperforms other methods in avoiding catastrophic forgetting, especially in highly heterogeneous spatial environments.
IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE SENSING
(2022)