China
Africa
Explore chapters and articles related to this topic
State of the Art in Optimal Design and Control of Urban Wastewater Systems
Published in Carlos Alberto Vélez Quintero, Optimization of Urban Wastewater Systems using Model Based Design and Control, 2020
A random search is the simplest stochastic search strategy, as it simply evaluates a given number of randomly selected solutions. Like enumeration, though, these strategies are not efficient for many MOPs because of their failure to incorporate problem domain knowledge. Random searches can generally expect to do no better than enumerative ones. In general, Monte Carlo methods employs a pure random search where any selected trial solution is fully independent of any previous choice and its outcome (Osyczka 1985). The current “best” solution and associated decision variables are stored as a comparator. Tabu search is a metastrategy developed to avoid getting “stuck” on local optima. It keeps a record of both visited solutions and the “paths” which reached them in different “memories.” This information restricts the choice of solutions to evaluate next. Tabu search is often integrated with other optimization methods (Glover and Laguna 1997).
Study of the optimization evaluation algorithm of improved genetic-BP neural network in the monitoring and diagnosis of shortwave transmitting system
Published in Jimmy C.M. Kao, Wen-Pei Sung, Civil, Architecture and Environmental Engineering, 2017
Mutation is also a way to generate new individuals. In genetic algorithm, the method of replacing some gene values in the individual chromosome coding by other values is used to achieve mutation simulation. In the process of inheritance, the mutation is usually randomly generated. Compared with selection and crossover operation, the individual’s local mutation, as the supplement of the new individual method, can avoid the loss of some information and ensure the effectiveness of the genetic process. Genetic algorithm achieves a global search by the crossover operation and improves local random search ability by mutation operation. In this paper, the method of uniform mutation is used, namely setting the gene locus in the chromosome coding as the mutation point, and randomly selecting a numerical value to replace the original gene value in the value range of the gene corresponding to the mutation probability pm, thus generating new individuals at a lower probability.
Disassembly Sequencing Problem: Resolving the Complexity by Random Search Techniques
Published in Surendra M. Gupta, A. J. D. (Fred) Lambert, Environment Conscious Manufacturing, 2007
Mukul Tripathi, Shubham Agrawal, M. K. Tiwari
Recently, random search techniques have gained increasing importance because of their ability to provide highly efficient and competitive solutions in a very short span of time. The complexities in manufacturing are growing, and the need to provide competitive alternatives at the earliest has motivated the development and the implementation of these search techniques on complex optimization problems. In this chapter, we will briefly illustrate different AI techniques and their implementation on the disassembly sequencing problem.
Genetic algorithm-based integrated optimization of active control systems for civil structures subjected to random seismic excitations
Published in Engineering Optimization, 2020
Zhen Mei, Zixiong Guo, Lincong Chen, Haifeng Wang, Yichao Gao
The placement optimization of control devices is one of the hottest research topics in the field of structural vibration control. For example, Lopez and Soong (2002) proposed sequential search methods for determining the optimal placement of dampers. Abdullah (1999) used a gradient-based algorithm to optimize the placement of control devices. In this approach, each additional device was optimized one at a time. Peng, Ghanem, and Li (2013) developed a sequential procedure with a minimum storey controllability index gradient to search for the optimal topology of control devices. Although these methods have been found to be effective, they may not necessarily yield the optimal placement of control devices; that is, they may become trapped in local optima (Abdullah, Richardson, and Hanif 2001; Abbasi and Markazi 2014). The reason for this is that the addition of each control device to the structure and control system changes the system dynamics. In essence, the position optimization of control devices belongs to the discrete optimization problem. Random search methods, such as simulated annealing and genetic algorithms, are global optimization algorithms, and can be used to solve discrete optimization problems. In particular, genetic algorithms have been applied in research on the position optimization of control devices (Furuya, Hamazaki, and Fujita 1998; Liu, Yang, and Li 2003; Wongprasert and Symans 2004).
A survey on simulation optimization for the manufacturing system operation
Published in International Journal of Modelling and Simulation, 2018
Ran Liu, Xiaolei Xie, Kaiye Yu, Qiaoyu Hu
Random search can work on an infinite parameter space. The technique selects points at random from the overall search region [23]. Considering a production-distribution system with no analytical constraints, Li et al. [24] adopt random search for problem solving where both the performance measure and the constraints are estimated via a stochastic, discrete-event simulation. Earl et al. [25] develop two scheduling methods to minimize expected costs for several products across multiple finite capacity resources: the first sub-optimizes the operations sequence by perturbation, using mean operation durations; the second generates a schedule of start times directly by random search with an embedded simulation of candidate schedules for evaluation. Matinnejad et al. [26] combine random search, adaptive random search, hill climbing, and simulated annealing algorithms for Model-in-the-Loop testing of continuous controllers. Generally speaking, random search has not been adopted alone much in manufacturing operations. Sometimes it assists other methods to simplify the search process. The major shortcoming of random search is the slow convergence to optimum accompanied by lots of computational efforts.
1st MICCAI workshop on deep learning in medical image analysis
Published in Computer Methods in Biomechanics and Biomedical Engineering: Imaging & Visualization, 2018
Gustavo Carneiro, João Manuel R. S. Tavares, Andrew P. Bradley, João Paulo Papa, Jacinto C. Nascimento, Jaime S. Cardoso, Zhi Lu, Vasileios Belagiannis
Additionally, there is a little effort on model selection of deep learning techniques, which poses an interesting problem, since we may face hundreds of parameters, being a near-exhaustive search on this high-dimensional search space impractical. The problem gets worse in large image-based data-sets, which have been commonly used in several recent papers. Given the large amount of parameters, some authors have argued that a random search may perform well for some applications. However, a hand tuning of the parameters may limit our understanding about how well the techniques can generalise and describe data.