A Multi-Task Decomposition-Based Evolutionary Algorithm for Tackling High-Dimensional Bi-Objective Feature Selection DOI Creative Commons
Hang Xu,

Chaohui Huang,

Jianbing Lin

et al.

Mathematics, Journal Year: 2024, Volume and Issue: 12(8), P. 1178 - 1178

Published: April 14, 2024

Evolutionary algorithms have been widely applied for solving multi-objective optimization problems, while the feature selection in classification can also be treated as a discrete bi-objective problem if attempting to minimize both error and ratio of selected features. However, traditional evolutionary (MOEAs) may drawbacks tackling large-scale selection, due curse dimensionality decision space. Therefore, this paper, we concentrated on designing an multi-task decomposition-based algorithm (abbreviated MTDEA), especially handling high-dimensional classification. To more specific, multiple subpopulations related different tasks are separately initialized then adaptively merged into single integrated population during evolution. Moreover, ideal points these dynamically adjusted every generation, order achieve search preferences directions. In experiments, proposed MTDEA was compared with seven state-of-the-art MOEAs 20 datasets terms three performance indicators, along using comprehensive Wilcoxon Friedman tests. It found that performed best most datasets, significantly better ability promising efficiency.

Language: Английский

Multi-objective exponential distribution optimizer (MOEDO): a novel math-inspired multi-objective algorithm for global optimization and real-world engineering design problems DOI Creative Commons
Kanak Kalita, Janjhyam Venkata Naga Ramesh, Lenka Čepová

et al.

Scientific Reports, Journal Year: 2024, Volume and Issue: 14(1)

Published: Jan. 20, 2024

Language: Английский

Citations

28
Analyzing the impact of feature selection methods on machine learning algorithms for heart disease prediction DOI Creative Commons

Zeinab Noroozi,

Azam Orooji, Leila Erfannia

et al.

Scientific Reports, Journal Year: 2023, Volume and Issue: 13(1)

Published: Dec. 18, 2023

Abstract The present study examines the role of feature selection methods in optimizing machine learning algorithms for predicting heart disease. Cleveland Heart disease dataset with sixteen techniques three categories filter, wrapper, and evolutionary were used. Then seven Bayes net, Naïve (BN), multivariate linear model (MLM), Support Vector Machine (SVM), logit boost, j48, Random Forest applied to identify best models prediction. Precision, F-measure, Specificity, Accuracy, Sensitivity, ROC area, PRC measured compare methods' effect on prediction algorithms. results demonstrate that resulted significant improvements performance some (e.g., j48), whereas it led a decrease other (e.g. MLP, RF). SVM-based filtering have best-fit accuracy 85.5. In fact, best-case scenario, result + 2.3 accuracy. SVM-CFS/information gain/Symmetrical uncertainty highest improvement this index. filter number features selected outperformed terms models' ACC, F-measures. However, wrapper-based improved from sensitivity specificity points view.

Language: Английский

Citations

41
Multi-objective liver cancer algorithm: A novel algorithm for solving engineering design problems DOI Creative Commons
Kanak Kalita, Janjhyam Venkata Naga Ramesh, Róbert Čep

et al.

Heliyon, Journal Year: 2024, Volume and Issue: 10(5), P. e26665 - e26665

Published: March 1, 2024

This research introduces the Multi-Objective Liver Cancer Algorithm (MOLCA), a novel approach inspired by growth and proliferation patterns of liver tumors. MOLCA emulates evolutionary tendencies tumors, leveraging their expansion dynamics as model for solving multi-objective optimization problems in engineering design. The algorithm uniquely combines genetic operators with Random Opposition-Based Learning (ROBL) strategy, optimizing both local global search capabilities. Further enhancement is achieved through integration elitist non-dominated sorting (NDS), information feedback mechanism (IFM) Crowding Distance (CD) selection method, which collectively aim to efficiently identify Pareto optimal front. performance rigorously assessed using comprehensive set standard test benchmarks, including ZDT, DTLZ various Constraint (CONSTR, TNK, SRN, BNH, OSY KITA) real-world design like Brushless DC wheel motor, Safety isolating transformer, Helical spring, Two-bar truss Welded beam. Its efficacy benchmarked against prominent algorithms such grey wolf optimizer (NSGWO), multiobjective multi-verse (MOMVO), (NSGA-II), decomposition-based (MOEA/D) marine predator (MOMPA). Quantitative analysis conducted GD, IGD, SP, SD, HV RT metrics represent convergence distribution, while qualitative aspects are presented graphical representations fronts. source code available at: https://github.com/kanak02/MOLCA.

Language: Английский

Citations

15
Multi-objective Geometric Mean Optimizer (MOGMO): A Novel Metaphor-Free Population-Based Math-Inspired Multi-objective Algorithm DOI Creative Commons
Sundaram B. Pandya, Kanak Kalita, Pradeep Jangir

et al.

International Journal of Computational Intelligence Systems, Journal Year: 2024, Volume and Issue: 17(1)

Published: April 11, 2024

Abstract This research introduces a novel multi-objective adaptation of the Geometric Mean Optimizer (GMO), termed Multi-Objective (MOGMO). MOGMO melds traditional GMO with an elite non-dominated sorting approach, allowing it to pinpoint Pareto optimal solutions through offspring creation and selection. A Crowding Distance (CD) coupled Information Feedback Mechanism (IFM) selection strategy is employed maintain amplify convergence diversity potential solutions. efficacy capabilities are assessed using thirty notable case studies. encompasses nineteen benchmark problems without constraints, six constraints five engineering design challenges. Based on optimization results, proposed better 54.83% in terms GD, 64.51% IGD, 67.74% SP, 70.96% SD, HV 77.41% RT. Therefore, has for solving un-constraint, constraint real-world application. Statistical outcomes from compared those Equilibrium (MOEO), Decomposition-Based Symbiotic Organism Search (MOSOS/D), Non-dominated Sorting Genetic Algorithm (NSGA-II), Multi-Verse Optimization (MOMVO) Plasma Generation (MOPGO) algorithms, utilizing identical performance measures. comparison reveals that consistently exhibits robustness excels addressing array The source code available at https://github.com/kanak02/MOGMO .

Language: Английский

Citations

12
Integrated Multiobjective Energy Management for a Smart Microgrid Incorporating Electric Vehicle Charging Stations and Demand Response Programs Under Uncertainty DOI Creative Commons
Rahman Hasani, Mohammad Mohammadi, Amin Samanfar

et al.

International Journal of Energy Research, Journal Year: 2025, Volume and Issue: 2025(1)

Published: Jan. 1, 2025

This paper presents an innovative 24‐h scenario–based microgrid energy management system (MG‐EMS) designed to achieve cost reduction and emission under conditions of uncertainty. Furthermore, a multiobjective hybrid heuristic algorithm, named particle swarm optimization lightning search algorithm (hMOPSO‐LSA), is introduced tackle the MG‐EMS problem. combines LSA MOPSO algorithm. The MG investigation comprises photovoltaic (PV) wind turbine (WT) units, combined heat power (CHP) system, employs multicarrier storage technology, specifically, power‐to‐gas (P2G) technology electric vehicle (EV) parking lot (PL). Flexible loads are incorporated into enhance through participation in demand response program (DRP). proposed model utilizes probability density functions (PDFs) for modeling uncertainties Roulette wheel (RW) method scenario selection. simulations, carried out MATLAB, encompass two different sections. In first part, accuracy efficiency were validated by solving standard DTLZ benchmark comparing results with those several other algorithms. second was using model, solved hMOPSO‐LSA both without flexible their inclusion. To provide comprehensive evaluation, problem compared against three algorithms: flower pollination (MOFPA), MOPSO, dragonfly (MODA). demonstrate that achieves higher findings indicate DRP 6.43% 8.21% emissions. Additionally, P2G proves effective reduction, contributing 6.87% required gas supply within MG.

Language: Английский

Citations

1
A Comprehensive Review of Metaheuristics for Hyperparameter Optimization in Machine Learning DOI

Ramachandran Narayanan,

N. Ganesh

Published: March 29, 2024

Hyperparameter optimization is a critical step in the development and fine-tuning of machine learning (ML) models. Metaheuristic techniques have gained significant popularity for addressing this challenge due to their ability search hyperparameter space efficiently. In review, we present detailed analysis various metaheuristic ML, encompassing population-based, single solution-based, hybrid approaches. We explore application metaheuristics Bayesian neural architecture search, two prominent areas within field. Moreover, provide comparative these based on established criteria evaluate performance diverse ML applications. Finally, discuss future directions open challenges with special emphasis opportunities improvement metaheuristics. Other crucial issues like adaptability new paradigms, computational complexity, scalability are also discussed critically. This review aims researchers practitioners comprehensive understanding state-of-the-art tuning, thereby facilitating informed decisions advancements

Language: Английский

Citations

5
Metaheuristic Algorithms and Their Applications in Different Fields DOI
Abrar Yaqoob, Navneet Kumar Verma, Rabia Musheer Aziz

et al.

Published: March 29, 2024

A potent method for resolving challenging optimization issues is provided by metaheuristic algorithms, which are heuristic approaches. They provide an effective technique to explore huge solution spaces and identify close ideal or optimal solutions. iterative often inspired natural social processes. This study provides comprehensive information on algorithms the many areas in they used. Heuristic well-known their success handling issues. a tool problem-solving. Twenty such as tabu search, particle swarm optimization, ant colony genetic simulated annealing, harmony included article. The article extensively explores applications of these diverse domains engineering, finance, logistics, computer science. It underscores particular instances where have found utility, optimizing structural design, controlling dynamic systems, enhancing manufacturing processes, managing supply chains, addressing problems artificial intelligence, data mining, software engineering. paper thorough insight into versatile deployment across different sectors, highlighting capacity tackle complex wide range real-world scenarios.

Language: Английский

Citations

4
MORKO: A Multi-objective Runge–Kutta Optimizer for Multi-domain Optimization Problems DOI Creative Commons
Kanak Kalita, Pradeep Jangir, Sundaram B. Pandya

et al.

International Journal of Computational Intelligence Systems, Journal Year: 2025, Volume and Issue: 18(1)

Published: Jan. 8, 2025

Abstract In the current landscape, there is a rapid increase in creation of new algorithms designed for specialized problem scenarios. The performance these unfamiliar or practical settings often remains untested. This paper presents development, multi-objective Runge–Kutta optimizer (MORKO), which built upon principles elitist non-dominated sorting and crowding distance. goal to achieve superior efficiency, diversity, robustness solutions. MORKO effectiveness further enhanced by incorporating various strategies that maintain balance between diversity execution efficiency. approach not only directs search toward optimal regions but also ensures process does become stagnant. efficiency compared against renowned like marine predicator algorithm (MOMPA), gradient-based (MOGBO), evolutionary based on decomposition (MOEA/D), genetic (NSGA-II) several test benchmarks such as ZDT, DTLZ, constraint (CONSTR, TNK, SRN, BNH, OSY KITA) real-world engineering design (brushless DC wheel motor, safety isolating transformer, helical spring, two-bar truss, welded beam, disk brake, tool spindle cantilever beam) problems. We used unique, non-overlapping metrics this comparison suggested fresh correlation analysis technique exploration. outcomes were rigorously tested confirmed using non-parametric statistical evaluations. proves excel deriving comprehensive varied solutions many tests challenges, owing its multifaceted features. Looking ahead, has potential applications complex management tasks.

Language: Английский

Citations

0
An efficient multi-objective parrot optimizer for global and engineering optimization problems DOI Creative Commons

Mohammed R. Saad,

Marwa M. Emam,

Essam H. Houssein

et al.

Scientific Reports, Journal Year: 2025, Volume and Issue: 15(1)

Published: Feb. 11, 2025

Abstract The Parrot Optimizer (PO) has recently emerged as a powerful algorithm for single-objective optimization, known its strong global search capabilities. This study extends PO into the Multi-Objective (MOPO), tailored multi-objective optimization (MOO) problems. MOPO integrates an outward archive to preserve Pareto optimal solutions, inspired by behavior of Pyrrhura Molinae parrots. Its performance is validated on Congress Evolutionary Computation 2020 (CEC’2020) benchmark suite. Additionally, extensive testing four constrained engineering design challenges and eight popular confined unconstrained test cases proves MOPO’s superiority. Moreover, real-world helical coil springs automotive applications conducted depict reliability proposed in solving practical Comparative analysis was performed with seven published, state-of-the-art algorithms chosen their proven effectiveness representation current research landscape-Improved Manta-Ray Foraging Optimization (IMOMRFO), Gorilla Troops (MOGTO), Grey Wolf (MOGWO), Whale Algorithm (MOWOA), Slime Mold (MOSMA), Particle Swarm (MOPSO), Non-Dominated Sorting Genetic II (NSGA-II). results indicate that consistently outperforms these across several key metrics, including Set Proximity (PSP), Inverted Generational Distance Decision Space (IGDX), Hypervolume (HV), (GD), spacing, maximum spread, confirming potential robust method addressing complex MOO

Language: Английский

Citations

0
An Adaptive Initialization and Reproduction-Based Evolutionary Algorithm for Tackling Bi-Objective Feature Selection in Classification DOI Open Access
Hang Xu

Symmetry, Journal Year: 2025, Volume and Issue: 17(5), P. 671 - 671

Published: April 28, 2025

As a commonly used method in classification, feature selection can be treated as bi-objective optimization problem, whose objectives are to minimize both the classification error and number of selected features, suitable for multi-objective evolutionary algorithms (MOEAs) tackle. However, due discrete environment increasing traditional MOEAs could face shortcomings searching abilities, especially large-scale or high-dimensional datasets. Thereby, this work, an adaptive initialization reproduction-based algorithm (abbreviated AIR) is proposed, specifically designed addressing classification. In AIR, mechanism AI) reproduction AR) have been by analyzing characteristics currently solutions order improve their search abilities balance convergence diversity performances. Moreover, designing also utilizes implicit symmetry generated around some interpolation axes objective space. experiments, AIR comprehensively compared with five state-of-the-art list 20 real-life datasets, its statistical performance being overall best terms several indicators.

Language: Английский

Citations

0