China
Africa
Explore chapters and articles related to this topic
Advanced Algorithms for the Layout Problem
Published in Sunderesh S. Heragu, Facilities Design, 2022
In Chapter 9, we discussed hybrid algorithms. An algorithm that is a combination of two or more types of algorithms, for example, construction and improvement, is considered a hybrid algorithm. An algorithm that has the characteristics of optimal and heuristic algorithms can also be considered a hybrid algorithm. This type of hybrid algorithm is essentially an optimal layout algorithm that is terminated before verifying that the best solution obtained is optimal. Some optimal algorithms have been modified to stop the search after a preset computer run time has been reached, and then the best solution obtained is improved further using an improvement algorithm. Such algorithms can be found in Burkard and Stratman (1978), Bazaraa and Sherali (1980), and Bazaraa and Kirca (1983).
Hybrid Intelligence Techniques for Handwritten Digit Recognition
Published in Siddhartha Bhattacharyya, Anirban Mukherjee, Indrajit Pan, Paramartha Dutta, Arup Kumar Bhaumik, Hybrid Intelligent Techniques for Pattern Analysis and Understanding, 2017
Hemant Jain, Ryan Serrao, B.K. Tripathy
This hybrid algorithm tends to combine the strengths of the two individual algorithms. The back-propagation with momentum (BPM) algorithm has a strong tendency to find the locally optimistic result, while on the other hand, particle swarm optimization (PSO) has a stronger tendency to find the globally optimistic result [27]. This hybridization helps find the most optimized results. The BPM algorithm mainly involves two stages, the first being the propagation of the activation values from the input to the output layer. Let zinj be the input received to the hidden layer and b the bias. Xi,i=1,2,3,...,n forms the input layer and zj is the output value obtained when passed through the activation function.
Big Data Optimization in Electric Power Systems: A Review
Published in Ahmed F. Zobaa, Trevor J. Bihl, Big Data Analytics in Future Power Systems, 2018
Iman Rahimi, Abdollah Ahmadi, Ahmed F. Zobaa, Ali Emrouznejad, Shady H.E. Abdel Aleem
The authors have compared results with several optimization algorithms such as culture PSO (CPSO), modified PSO (MPSO), orthogonal teaching learning-based optimization (OTLBO), and teaching learning-based optimization (TLBO), GSA. Regarding robustness, the suggested method has better performance than other solution optimization methods. Moreover, the results show that hybrid algorithm has saved computational time significantly. Quality solution and the convergence speed of the hybrid algorithm possess superior performance than other optimization algorithms. Using of renewable energy has attracted the attention of power system planners across the world. Rajesh et al. (2016) applied differential evolution algorithm in a model of a solar plant to minimize both emission and cost. In the model, the data were gathered from demand and plants, and then the model is generated based on assumption. After several studies, the model is developed, and a solution methodology has been selected for the proposed model. A sensitivity analysis was applied to the proposed model, and finally, the future power system model is generated with characteristics such as total cost, capacity additions, emission level. Naderi et al. (2017) proposed a fuzzy adaptive, comprehensive-learning PSO known as FAHCLPSO for the large-scale power dispatch optimization problem. Objective functions for the proposed algorithm include minimizing the active power transmission losses and improving the voltage profile of the system. The authors have validated the performance of their suggested algorithm with three different tests, including IEEE 30-bus, IEEE 118-bus, and IEEE 354-bus test systems. The authors have claimed that the proposed algorithm (FAHCLPSO) was the first applied for optimal reactive power dispatch. They have used fuzzy logic to enhance the searchability of the algorithm.
Breast cancer detection in digital mammography using a novel hybrid approach of Salp Swarm and Cuckoo Search algorithm with deep belief network classifier
Published in The Imaging Science Journal, 2021
R. Reenadevi, B. Sathiyabhama, S. Sankar, Digvijay Pandey
Hybrid optimization algorithms are important in improving optimization algorithm search capability. The main objective of hybridization is to combine the advantages of each algorithm to form a hybrid algorithm while minimizing any significant disadvantages. In recent decades, compared to several algorithms, hybrid optimization algorithms are showing greater performance in solving medical image applications [7]. Some of the hybrid optimization algorithms include SSA-Particle Swarm Optimization (PSO), PSO-Firefly, Ant colony optimization-PSO, Crow search based on chaos theory with Fuzzy C-Means (FCM) clustering, and Monkey algorithm with Krill Herd [8]. Hence, the accuracy of classification depends on feature selection from large medical image data. In addition, deep learning classifiers are extensively used in medical image classification [9].This motivation energies to develop an optimization-based feature selection technique with DBN for breast cancer classification, by integrating pre-processing, feature extraction, feature selection, and classification [10].
Truss topology, shape and sizing optimization by fully stressed design based on hybrid grey wolf optimization and adaptive differential evolution
Published in Engineering Optimization, 2018
Natee Panagant, Sujin Bureerat
The strong and weak points of FSD-SHADE and FSD-GWO, to some extent, complement each other. Therefore, the two algorithms are hybridized, leading to FSD-GWADE. The hybrid algorithm shows improvements in overall performance as it gains all the advantages from its base algorithms. The algorithm also shows great improvement compared to GWADE without using FSD. The algorithm can provide low mean weight with high CCR. The CCR results of FSD-GWADE are higher than 70% in all tests. Mean weights found by FSD-GWADE and FSD-SHADE are close for the 45-bar, 15-bar, 25-bar, 39-bar and 47-bar problems, while FSD-GWADE finds significantly improved mean weight for the 68-bar problem. The mean rank results from performing multi-comparison using the Friedman test also confirm that FSD-GWADE can outperform all algorithms in Table 4. FSD-GWADE provides the lowest mean rank for all test problems.
A hybrid metaheuristic algorithm to achieve sustainable production: involving employee characteristics in the job-shop matching problem
Published in Journal of Industrial and Production Engineering, 2023
Bingtao Quan, Sujian Li, Kuo-Jui Wu
The genetic algorithm is an optimization method developed by imitating biological evolution mechanisms in nature [29]. Mayer et al. [30] utilized single and multiobjective optimizations by applying a genetic algorithm, which resulted in the lowest cost and lowest carbon emissions in grid-connected and off-grid scenarios, respectively. To effectively solve image classification tasks, Sun et al. [31] proposed an automatic CNN architecture design method by genetic algorithms. To overcome traditional controller shortcomings, an improved genetic algorithm optimization fuzzy controller was proposed by [32]. In the article, (Rostami et al. [33] proposes a genetic algorithm based on community detection for the purpose of feature selection, which is realized in three steps. Considering the problems of manual scheduling of medical treatment, Squires et al. [34] presented a novel genetic algorithm for the intelligent scheduling of transcranial magnetic stimulation appointments. However, to resolve more complicated problems, a hybrid algorithm that combines a genetic algorithm with another algorithm is usually employed. Li et al. [35] proposed the combination of a genetic algorithm and an artificial neural network to optimize the fuel efficiency and waste emissions from a new internal combustion engine. The example verifies that the prediction performance of the hybrid prediction system is better than that of the benchmark model [36]. Prior studies are increasing application trend for applying genetic algorithms or hybrid algorithms that combine genetic algorithms with other methods, because the genetic algorithm has the characteristics of global search and simple implementation in the hybrid algorithm.