Showing papers on "Multi-swarm optimization published in 2021"

Surrogate-Assisted Multipopulation Particle Swarm Optimizer for High-Dimensional Expensive Optimization

Abstract: Surrogate-assisted evolutionary algorithms (SAEAs) are well suited for computationally expensive optimization. However, most existing SAEAs only focus on low- or medium-dimensional expensive optimization. Thus, a novel SAEA for high-dimensional expensive optimization, denoted as surrogate-assisted multipopulation particle swarm optimizer (SA-MPSO), is proposed and fully investigated in this work. The proposed algorithm employs a parameter-free clustering technique, denoted as affinity propagation clustering, to generate several subswarms. A surrogate-assisted learning strategy-based particle swarm optimizer is proposed for guiding the search of each subswarm. Furthermore, a model management strategy is adapted to choose the promising particles for real fitness evaluations. Finally, a subswarm diversity maintenance scheme and a surrogate-based trust region local search technique are introduced to enhance both exploration and exploitation. The experimental results on commonly used benchmark test problems with dimensions varying from 30 to 100 and airfoil design problem have shown that SA-MPSO outperforms some state-of-the-art methods.

Enhanced multi-swarm cooperative particle swarm optimizer

Abstract: In this paper, a novel multi-swarm particle swarm optimizer driven by delayed-activation (DA) strategy and repulsive mechanism, named as enhanced multi-swarm cooperative particle swarm optimizer (EMCPSO) is proposed. EMCPSO is designed to make use of the advantage of multi-swarm technique and overcome the problem of premature convergence of original PSO. In this algorithm, the whole population is partitioned into four identical sub-swarms. The best particle of each sub-swarm, sbest, is used to estimate the evolutionary state of the group. If the sbest can continuously improve its solution's quality, that sub-swarm evolves independently without communicating with other counterparts. Otherwise, based on a non-ascending sequence, a delayed-activation (DA) strategy will be triggered. With information sharing among multi-swarm, activating exemplar is constructed to promote the stagnant sub-swarm to search for better solutions again. On the other hand, a repulsive mechanism is introduced to prevent the whole population from gathering together prematurely. In this way, more potential regions of the search space can be explored by EMCPSO. The experiment results on CEC 2017 problem set demonstrate the superior performance of the proposed EMCPSO in terms of solution accuracy and convergence speed.

Impact of population topology on particle swarm optimization and its variants: An information propagation perspective

Abstract: Particle swarm optimization is one of the most effective optimization algorithms motivated by bird flocking behaviours. Population topology is a key aspect of particle swarm optimization research. However, after more than twenty years of research, the effects of the population topology are still poorly understood. Previous research has established that the information propagation speed determined by the population topology has an important impact on the algorithm performance; however, the impact of information propagation speed on particle swarm optimization and its variants has not yet been investigated. In this paper, information propagation in particle swarms is described and, hence, a method of simulating information propagation in particle swarms is introduced, which is used to obtain the information propagation speed. The correlation between the information propagation speed and algorithm performance is clarified through numerical simulation. The results show that the information propagation speed has a strong negative correlation with the population diversity of particle swarm optimization and its variants in the early iterations, regardless of the adopted test function and population diversity measure. The results also show that when optimizing problems with the same property, the impact of population topology on the optimization results of particle swarm optimization and variant algorithms is similar. Further more, this study provides some guidance on the population topology selection for particle swarm optimization and its variants. These findings contribute to our understanding of the impact of population topology on particle swarm optimization and its variants, and provide a basis for population topology selection for particle swarm optimization and its variants.

Multi-lead ECG signal analysis using RBFNN-MSO algorithm

Abstract: In this paper, we present a method for electrocardiogram beat based on multi swarm optimization and radial basis function neural network. ECG is a non-surgical method for measuring and recording the electrical activity of the heart and on many occasions, an experienced cardiologist may not be available on the patient’s site. Therefore, a type of automated ECG analysis is required for the patient to take the electrocardiogram by a general practitioner or paramedical team attending the patient’s location. There is a need for automated ECG analysis. Finally, this paper gives the best analysis methodology for the automated analysis of multi-channel ECG signals. Diagnosis may be affected by the presence of artifacts and noise in multi-channel ECG signals. Some researchers calculated dynamic cutoff frequency parameter from noisy ECG signals to remove noise using the neural network method of Radial Basis function with particle swarm improvement method (RBFNN-PSO). But PSO only has Swarm and it takes a lot of time to give a response. To overcome these limitations, an improved version of the RBFNN-PSO algorithm called Radial Basis Function Neural Network with Multi Swarm Optimization (RBFNN-MSO) has been proposed. Finally, the cutoff frequency parameter is determined by the RBFNN-MSO methodology that is applied to digital low-frequency filters for impulse response (FIR). The next step after removing noise from multi-channel ECG signals is the feature extraction and reduction process. 24 features of the patient’s multi-channel ECG signals are extracted. The next part of the research is divided into two steps. The first step is whether or not the patient’s ECG signals are affected. The vector machine is supported with particle swarm improvement (SVM-PSO) and another way is to support the vector machine with multi swarm improvement (SVM-MSO) to detect ECG signals of the affected patient or not. Finally, SVM-MSO offers greater accuracy compared to SVM-PSO. When compared to all other existing architecture results, they used the rating with 86% of all the test accuracy. But in this paper, our proposed work has 90% overall in different situations. In another point of view also,our proposed work has proven that average accuracy is over 85% even then train data set is small.

Multi-swarm optimization model for multi-cloud scheduling for enhanced quality of services

Abstract: Cloud services gain more attention due to its accessibility, performance, and cost factors. Cloud offers a wide range of services and completes the task without any delay due to its scheduling policies. Task scheduling is an important factor in cloud computing applications. The performance of applications increases due to an effective scheduling strategy. The cloud resources are allocated to the tasks through task scheduling. Factors like customer satisfaction, resource utilization, better performance make task scheduling crucial for service providers. Depending on the scheduling schemes support in clouds, scheduling is categorized into single cloud or multi-cloud scheduling. Multi-cloud environment provides diverse resources and significantly reduces the cost and commercial limitations. However, reducing the cost functions and makespan are the major factors considered to avoid customer dissatisfaction. But it is essential to concentrate on other factors, such as throughput, delay, Makespan, waiting time, response time, utilization, and efficiency to improve the quality of services. This research work presents a Multi-Swarm Optimization model for Multi-Cloud Scheduling for Enhanced Quality of Services for a multi-cloud environment. Experimental results demonstrate that the proposed approach performs better in all aspects compared to existing techniques, such as Adaptive energy-efficient scheduling, single objective particle swarm optimization scheduling, and improves the quality of services.

MSLDOCK: Multi-Swarm Optimization for Flexible Ligand Docking and Virtual Screening.

Abstract: Autodock and its various variants are widely utilized docking approaches, which adopt optimization methods as search algorithms for flexible ligand docking and virtual screening. However, many of them have their limitations, such as poor accuracy for dockings with highly flexible ligands and low docking efficiency. In this paper, a multi-swarm optimization algorithm integrated with Autodock environment is proposed to design a high-performance and high-efficiency docking program, namely, MSLDOCK. The search algorithm is a combination of the random drift particle swarm optimization with a novel multi-swarm strategy and the Solis and Wets local search method with a modified implementation. Due to the algorithm's structure, MSLDOCK also has a multithread mode. The experimental results reveal that MSLDOCK outperforms other two Autodock-based approaches in many aspects, such as self-docking, cross-docking, and virtual screening accuracies as well as docking efficiency. Moreover, compared with three non-Autodock-based docking programs, MSLDOCK can be a reliable choice for self-docking and virtual screening, especially for dealing with highly flexible ligand docking problems. The source code of MSLDOCK can be downloaded for free from https://github.com/lcmeteor/MSLDOCK.

Multi-Swarm Optimization for Extracting Multiple-Choice Tests From Question Banks

Abstract: In this study, a novel method for generating multiple-choice tests is presented, which extracts the required number of tests of the same levels of difficulty in a single attempt and approximates the difficulty level requirement given by users. We propose an approach using parallelism and Pareto optimization for multi-swarm migration in a particle swarm optimization (PSO) algorithm. Multi-PSO is proposed for shortening the computing time. The proposed migration of PSOs increases the diversity of tests and controls the overlap of extracted tests. The experimental results show that the proposed method can generate many tests from question banks satisfying predefined levels of difficulty. Additionally, the developed method is shown to be effective in terms of many criteria when compared with other methods such as manually extracted tests, a simulated annealing algorithm (SA), random methods and PSO-based approaches in terms of the number of successful solutions, accuracy, standard deviation, search speed, and the number of questions overlapping between the exam questions, as well as for changing the search space, changing the number of individuals, changing the number of swarms, and changing the difficulty requirements.