检索结果-内蒙古大学图书馆

Integrating estimation of distribution algorithms versus Q-learning into Meta-RaPS for solving the 0-1 multidimensional knapsack problem

引用

COMPUTERS & INDUSTRIAL ENGINEERING 2017年 112卷 706-720页

作者： Arin, Arif Rabadi, Ghaith Middle East Tech Univ Ankara Turkey Old Dominion Univ Engn Management & Syst Engn Engn Syst BldgRoom 2102 Norfolk VA 23529 USA

Finding near-optimal solutions in an acceptable amount of time is a challenge when developing sophisticated approximate approaches. A powerful answer to this challenge might be reached by incorporating intelligence into metaheuristics. We propose integrating two methods into Meta-RaPS (Metaheuristic for Randomized Priority Search), which is currently classified as a memoryless metaheuristic. The first method is the estimation of distribution algorithms (EDA), and the second is utilizing a machine learning algorithm known as Q-Learning. To evaluate their performance, the proposed algorithms are tested on the 0-1 Multidimensional Knapsack Problem (MKP). Meta-RaPS EDA appears to perform better than Meta-RaPS Q-Learning. However, both showed promising results compared to other approaches presented in the literature for the 0-1 MKP. (C) 2016 Elsevier Ltd. All rights reserved.

关键词： Machine learning estimation of distribution algorithms Q-learning Meta-RaPS 0-1 multidimensional knapsack problem

来源：评论

学校读者我要写书评

暂无评论

Evolutionary and estimation of distribution algorithms for Unconstrained, Constrained, and Multiobjective Noisy Combinatorial Optimisation Problems

引用

EVOLUTIONARY COMPUTATION 2023年第3期31卷 259-285页

作者： Aishwaryaprajna Rowe, Jonathan E. Univ Birmingham Sch Comp Sci Birmingham W Midlands England Alan Turing Inst London England

We present an empirical study of a range of evolutionary algorithms applied to various noisy combinatorial optimisation problems. There are three sets of experiments. The first looks at several toy problems, such as OneMax and other linear problems. We find that UMDA and the Paired-Crossover Evolutionary Algorithm (PCEA) are the only ones able to cope robustly with noise, within a reasonable fixed time budget. In the second stage, UMDA and PCEA are then tested on more complex noisy problems: SubsetSum, Knapsack, and SetCover. Both perform well under increasing levels of noise, with UMDA being the better of the two. In the third stage, we consider two noisy multiobjective problems (CountingOnesCountingZeros and a multiobjective formulation of SetCover). We compare several adaptations of UMDA for multiobjective problems with the Simple Evolutionary Multiobjective Optimiser (SEMO) and NSGA-II. We conclude that UMDA, and its variants, can be highly effective on a variety of noisy combinatorial optimisation, outperforming many other evolutionary algorithms.

关键词： Noisy combinatorial optimisation noisy multiobjective optimisation expected runtime crossover estimation of distribution algorithms NSGA-II

来源：评论

学校读者我要写书评

暂无评论

Space complexity of estimation of distribution algorithms

引用

EVOLUTIONARY COMPUTATION 2005年第1期13卷 125-143页

作者： Gao, Y Univ Alberta Dept Comp Sci Edmonton AB T6G 2E8 Canada

In this paper, we investigate the space complexity of the estimation of distribution algorithms (EDAs), a class of sampling-based variants of the genetic algorithm. By analyzing the nature of EDAs, we identify criteria that characterize the space complexity of two typical implementation schemes of EDAs, the factorized distribution algorithm and Bayesian network-based algorithms. Using random additive functions as the prototype, we prove that the space complexity of the factorized distribution algorithm and Bayesian network-based algorithms is exponential in the problem size even if the optimization problem has a very sparse interaction structure.

关键词： estimation of distribution algorithms space complexity additive fitness functions graphical models and bayesian networks treewidth

来源：评论

学校读者我要写书评

暂无评论

Feature selection for intrusion detection using new multi-objective estimation of distribution algorithms

引用

APPLIED INTELLIGENCE 2019年第12期49卷 4237-4257页

作者： Maza, Sofiane Touahria, Mohamed Univ Ferhat Abbas Setif 1 Dept Comp Sci Setif 19000 Algeria

The manipulation of a large number of features has become a critical problem in Intrusion Detection Systems(IDS). Therefore, Feature Selection (FS) is integrated to select the significant features, in order to avoid the computational complexity, and improve the classification performance. In this paper, we present a new multi-objective feature selection algorithm MOEDAFS (Multi-Objective estimation of distribution algorithms (EDA) for Feature Selection). The MOEDAFS is based on EDA and Mutual Information (MI). EDA is used to explore the search space and MI is integrated as a probabilistic model to guide the search by modeling the redundancy and relevance relations between features. Therefore, we propose four probabilistic models for MOEDAFS. MOEDAFS selects the better feature subsets (non-dominated solutions) that have a better detection accuracy and smaller number of features. MOEDAFS uses two objective functions (minimizing classification Error Rate (ER) and minimizing the Number of Features(NF)). In order to demonstrate the performance of MOEDAFS, a comparative study is designed by internal and external comparison on NSL-KDD dataset. Internal comparison is performed between the four versions of MOEDAFS. External comparison is organized against some well-known deterministic, metaheuristic, and multi-objective feature selection algorithms that have a single and Multi-solution. Experimental results demonstrate that MOEDAFS outperforms recent algorithms.

关键词： Intrusion detection system Feature selection Multi-Objective optimization estimation of distribution algorithms Mutual information

来源：评论

学校读者我要写书评

暂无评论

Delineation of site-specific management zones using estimation of distribution algorithms

引用

INTERNATIONAL TRANSACTIONS IN OPERATIONAL RESEARCH 2023年第4期30卷 1703-1729页

作者： Velasco, Jonas Vicencio, Salvador Lozano, Jose A. Cid-Garcia, Nestor M. CONACYT Ctr Invest Matemat CIMAT AC Fray Bartolome Casas 314 Aguascalientes 20259 Aguascalientes Mexico Ctr Invest Matemat CIMAT AC Fray Bartolome Casas 314 Aguascalientes 20259 Aguascalientes Mexico Univ Basque Country UPV EHU Basque Ctr Appl Math BCAM Paseo Manuel Lardizabal 1 Donostia San Sebastian 20018 Spain CONACYT Ctr Invest Ciencias Informac Geoespacial Lab Nacl Geointeligencia Circuito Tecnopolo Norte 107Tecnopolo Pocitos 2 Aguascalientes 20313 Aguascalientes Mexico

In this paper, we present a novel methodology to solve the problem of delineating homogeneous site-specific management zones (SSMZ) in agricultural fields. This problem consists of dividing the field into small regions for which a specific rate of inputs is required. The objective is to minimize the number of management zones, which must be homogeneous according to a specific soil property: physical or chemical. Furthermore, as opposed to oval zones, SSMZ with rectangular shapes are preferable since they are more practical for agricultural technologies. The methodology we propose is based on evolutionary computation, specifically on a class of the estimation of distribution algorithms (EDAs). One of the strongest contributions of this study is the representation used to model the management zones, which generates zones with orthogonal shapes, for example, L or T shapes, and minimizes the number of zones required to delineate the field. The experimental results show that our method is efficient to solve real field and randomly generated instances. The average improvement of our method consists in reducing the number of management zones in the agricultural fields concerning other operations research methods presented in the literature. The improvement depends on the size of the field and the level of homogeneity established for the resulting management zones.

关键词： site‐ specific management zones estimation of distribution algorithms orthogonal shapes evolutionary computation combinatorial optimization

来源：评论

学校读者我要写书评

暂无评论

Addressing the advantages of using ensemble probabilistic models in estimation of distribution algorithms for scheduling problems

引用

INTERNATIONAL JOURNAL OF PRODUCTION ECONOMICS 2013年第1期141卷 24-33页

作者： Chen, Shih-Hsin Chen, Min-Chih Nanhua Univ Dept Elect Commerce Management Dalin 62248 Taiwan WuFeng Univ Dept Informat Management Wufeng 62153 Taiwan

estimation of distribution algorithms (EDAs) have recently been recognized as a prominent alternative to traditional evolutionary algorithms due to their increasing popularity. The core of EDAs is a probabilistic model which directly impacts performance of the algorithm. Previous EDAs have used a univariate, bi-variate, or multi-variable probabilistic model each time. However, application of only one probabilistic model may not represent the parental distribution well. This paper advocates the importance of using ensemble probabilistic models in EDAs. We combine the univariate probabilistic model with the bi-variate probabilistic model which learns different population characteristics. To explain how to employ the two probabilistic models, we proposed the Ensemble Self-Guided Genetic Algorithm (eSGGA). The extensive computation results on two NP-hard scheduling problems indicate the advantages of adopting two probabilistic models. Most important of all, eSGGA can avoid the computation effort overhead when compared with other EDAs employing two models. As a result, this paper might point out a next generation approach for EDAs. (c) 2012 Elsevier B.V. All rights reserved.

关键词： estimation of distribution algorithms Single machine scheduling problem Permutation flowshop scheduling problem Self-Guided Genetic Algorithm

来源：评论

学校读者我要写书评

暂无评论

Structure Learning of Bayesian Networks by estimation of distribution algorithms with Transpose Mutation

引用

JOURNAL OF APPLIED RESEARCH AND TECHNOLOGY 2013年第4期11卷 586-596页

作者： Kim, D. W. Ko, S. Kang, B. Y. Chung Ang Univ Sch Comp Sci & Engn Seoul 156756 South Korea Kyungpook Natl Univ Sch Mech Engn Taegu 702701 South Korea

estimation of distribution algorithms (EDAs) constitute a new branch of evolutionary optimization algorithms that were developed as a natural alternative to genetic algorithms (GAs). Several studies have demonstrated that the heuristic scheme of EDAs is effective and efficient for many optimization problems. Recently, it has been reported that the incorporation of mutation into EDAs increases the diversity of genetic information in the population, thereby avoiding premature convergence into a suboptimal solution. In this study, we propose a new mutation operator, a transpose mutation, designed for Bayesian structure learning. It enhances the diversity of the offspring and it increases the possibility of inferring the correct arc direction by considering the arc directions in candidate solutions as bi-directional, using the matrix transpose operator. As compared to the conventional EDAs, the transpose mutation-adopted EDAs are superior and effective algorithms for learning Bayesian networks.

关键词： estimation of distribution algorithms Mutation Bayesian network Structure learning Optimization

来源：评论

学校读者我要写书评

暂无评论

A novel classification learning framework based on estimation of distribution algorithms

引用

INTERNATIONAL JOURNAL OF COMPUTING SCIENCE AND MATHEMATICS 2012年第4期3卷 353-366页

作者： Fan, Jiancong Xu, Qiang Liang, Yongquan Shandong Univ Sci & Technol Coll Informat Sci & Engn Qingdao Shandong Peoples R China Shandong Univ Sci & Technol Coll Informat Sci & Engn Comp Sci Qingdao Shandong Peoples R China

estimation of distribution algorithms (abbr. EDAs) is a relatively new branch of evolutionary algorithms. EDAs replace search operators with the estimation of the distribution of selected individuals + sampling from the population. In an EDAs, this explicit representation of the population is replaced with a probability distribution over the choices available at each position in the vector that represents a population member. In this paper, an estimation of distribution learning framework and the corresponding learning algorithm are proposed and the relevant properties of the framework are analysed on the basis of probability. The framework provides a basis and a principle criterion for designing and analysing evolutionary learning algorithms based on EDAs. The probability is the core tool of EDAs. EDA-based learning algorithms are required to estimate the population distribution by the sample distributions. The learning framework proposed can guide and regulate the design processes of learning algorithms and strategies based on EDAs. The framework involved in relevant learning problems is analysed from the perspectives of probability by properties analysis, proof and verification. The experiment results show that the framework proposed is feasible for realising learning from datasets and has better learning performances than some other relevant evolutionary learning methods.

关键词： evolutionary computation classification estimation of distribution algorithms EDAs machine learning

来源：评论

学校读者我要写书评

暂无评论

Optimizing logistic regression coefficients for discrimination and calibration using estimation of distribution algorithms

引用

TOP 2008年第2期16卷 345-366页

作者： Robles, V. Bielza, C. Larranaga, P. Gonzalez, S. Ohno-Machado, L. Univ Politecn Madrid Dept Arquitectura & Tecnol Sistemas Informat Madrid Spain Univ Politecn Madrid Dept Inteligencia Artificial Madrid Spain Harvard Univ Div Hlth Sci & Technol Boston MA 02115 USA MIT Boston MA USA

Logistic regression is a simple and efficient supervised learning algorithm for estimating the probability of an outcome or class variable. In spite of its simplicity, logistic regression has shown very good performance in a range of fields. It is widely accepted in a range of fields because its results are easy to interpret. Fitting the logistic regression model usually involves using the principle of maximum likelihood. The Newton-Raphson algorithm is the most common numerical approach for obtaining the coefficients maximizing the likelihood of the data. This work presents a novel approach for fitting the logistic regression model based on estimation of distribution algorithms (EDAs), a tool for evolutionary computation. EDAs are suitable not only for maximizing the likelihood, but also for maximizing the area under the receiver operating characteristic curve (AUC). Thus, we tackle the logistic regression problem from a double perspective: likelihood-based to calibrate the model and AUC-based to discriminate between the different classes. Under these two objectives of calibration and discrimination, the Pareto front can be obtained in our EDA framework. These fronts are compared with those yielded by a multiobjective EDA recently introduced in the literature.

关键词： Logistic regression Evolutionary algorithms estimation of distribution algorithms Calibration and discrimination 62J12 90C59 90C29

来源：评论

学校读者我要写书评

暂无评论

estimation of distribution algorithms as Logistic Regression Regularizers of Microarray Classifiers

引用

METHODS OF INFORMATION IN MEDICINE 2009年第3期48卷 236-241页

作者： Bielza, C. Robles, V. Larranaga, P. Univ Politecn Madrid Dept Inteligencia Artificial E-28040 Madrid Spain Univ Politecn Madrid Dept Arquitectura & Tecnol Sistemas Informat E-28040 Madrid Spain

Objectives: The "large k (genes), small N (samples)" phenomenon complicates the problem of microarray classification with logistic regression. The indeterminacy of the maximum likelihood solutions, multicollinearity of predictor variables and data over-fitting cause unstable parameter estimates. Moreover, computational problems arise due to the large number of predictor (genes) variables. Regularized logistic regression excels as a solution. However, the difficulties found here involve an objective function hard to be optimized from a mathematical viewpoint and a careful required tuning of the regularization parameters. Methods: Those difficulties are tackled by introducing a new way of regularizing the logistic regression. estimation of distribution algorithms (EDAs), a kind of evolutionary algorithms, emerge as natural regularizers. Obtaining the regularized estimates of the logistic classifier amounts to maximizing the likelihood function via our EDA, without having to be penalized. Likelihood penalties add a number of difficulties to the resulting optimization problems, which vanish in our case. Simulation of new estimates during the evolutionary process of EDAs is performed in such a way that guarantees their shrinkage while maintaining their probabilistic dependence relationships learnt. The EDA process is embedded in an adapted recursive feature elimination procedure, thereby providing the genes that are best markers for the classification. Results: The consistency with the literature and excellent classification performance achieved with our algorithm are illustrated on four microarray data sets: Breast, Colon, Leukemia and Prostate. Details on the last two data sets are available as supplementary material. Conclusions: We have introduced a novel EDA-based logistic regression regularizer. It implicitly shrinks the coefficients during EDA evolution process while optimizing the usual likelihood function. The approach is combined with a gene subset selection pr

关键词： Logistic regression regularization estimation of distribution algorithms DNA microarrays

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：