Genetic Algorithm Learning Research Articles

Research on concentration retrieval of gas FTIR spectra by interval extreme learning machine and genetic algorithm

This paper proposed a novel effective quantitative analysis method for FTIR spectroscopy of polluted gases, which select the best wavenumbers based on the idea of interval dividing. Meanwhile, genetic algorithm was adopted to optimize the connect weights and thresholds of the input layer and the hidden layer of extreme learning machine (ELM) because of its global search ability. Firstly, the whole spectrum region was divided into several subintervals; Secondly, the quantitative analysis model was established in each subinterval by using optimized GA-ELM; Thirdly, the best combination of subintervals was selected according to the generalized performance of each subinterval model by computing the parameters root mean square error (RMSE) and determined coefficients r. In this paper, the mixture of CO, CO2 and N2 O gases were selected as the research object and the whole spectrum range was from 2 140 to 2 220 cm-1. The experiment results showed that the RMSE of model established with the selected wavenumbers was 154. 996 3, the corresponding r can reach 0. 987 4, and the running time was just 0. 8 seconds, which indicated that the concentration retrieval model established with the proposed Interval-GA-ELM (iGELM) method can not only reduce the modeling time, but also can improve the stability and predict accuracy, especially under the condition of the exist of interferents, which providing an effective approach to the remote analysis of polluted gases.

Classification of hyperspectral remote sensing images based on simulated annealing genetic algorithm and multiple instance learning

A hybrid feature selection and classification strategy was proposed based on the simulated annealing genetic algorithm and multiple instance learning (MIL). The band selection method was proposed from subspace decomposition, which combines the simulated annealing algorithm with the genetic algorithm in choosing different cross-over and mutation probabilities, as well as mutation individuals. Then MIL was combined with image segmentation, clustering and support vector machine algorithms to classify hyperspectral image. The experimental results show that this proposed method can get high classification accuracy of 93.13% at small training samples and the weaknesses of the conventional methods are overcome.

Fuzzy-Weighted Similarity Measures for Memory-Based Collaborative Recommender Systems

Memory-based collaborative recommender system (CRS) computes the similarity between users based on their declared ratings. However, not all ratings are of the same importance to the user. The set of ratings each user weights highly differs from user to user according to his mood and taste. This is usually reflected in the user’s rating scale. Accordingly, many efforts have been done to introduce weights to the similarity measures of CRSs. This paper proposes fuzzy weightings for the most common similarity measures for memory-based CRSs. Fuzzy weighting can be considered as a learning mechanism for capturing the preferences of users for ratings. Comparing with genetic algorithm learning, fuzzy weighting is fast, effective and does not require any more space. Moreover, fuzzy weightings based on the rating deviations from the user’s mean of ratings take into account the different rating scales of different users. The experimental results show that fuzzy weightings obviously improve the CRSs performance to a good extent.

PERFORMANCE ANALYSIS OF LINKAGE LEARNING TECHNIQUES IN GENETIC ALGORITHMS

One variance of Genetic Algorithms is a Linkage Learning Genetic Algorithm (LLGA) enhances the efficiencies of Simple Genetic Algorithm (SGA) while solving NP hard Problems. Discovery of Linkage Learning Technique is an important task in GA. Almost all existing Linkage Learning Techniques follow either random approach or probabilistic approac hes. This makes repeated passes over the population to determine the relatio nship between individuals. SGA with random linkage technique is simple but may take long time to converge to the optimal solut ions. This paper uses a linkage learning operator c alled Gene Silencing which is an inspired mechanism from biological systems. T he Gene Silencing mechanism is used to improve the linkages by preserving the building blocks in an individual from the disru ption of recombination processes such as Crossover and Mutation. It converges quickly to the optimal solution without compromisin g the diversification on search spaces. To prove th is phenomenon, the Travelling Sales Person problem (TSP) has been chosen to retain the order of cities in a tour. Experim ents carried out on different TSP benchmark instances taken from TSPLIB which is a standard library for TSP problems. These benchmark instances have also been applied on various linkage learning techniques and analyses the performance of these techniques w ith Gene Silencing (GS) mechanism. The performance analysis has been made on experimental results with respect to optimal solu tion and convergence speed.

Motion recognition for 3D human motion capture data using support vector machines with rejection determination

This article presents a motion recognition strategy with rejection ability to extract the meaningful actions according to a given set of motion classes, or categories or types and reject such input patterns whose categories are not known. During the online recognition phrase, the multiple one-versus-one support vector machines are aggregated with the majority voting strategy over the most recent frames in a sliding window to predict the most probable type at each instance. And then, the corresponding index motion map is utilized to determine whether the predicted type should be accepted or not. The motion will be considered to be unknown when consecutive multiple frames are rejected. As a contribution, an adjusted self-organizing map algorithm is proposed to automatically learn the index motion map for each motion class, where the map size and topology are dynamically tuned by the intrinsic characteristics of the trained motions dataset. At the postprocessing step, the procedure is enhanced by an efficient key patterns-based verification strategy, which significantly improves the recognition precision. As a further contribution, we introduce a genetic algorithm learning algorithm to automatically learn the necessary key patterns for each class base on the previous learned index motion map. We evaluate our motion recognition model on various experiments conducted on synthetic data and real data from the freely available sets of motion capture database (HDM05). Experiment results show that the proposed strategy can not only classify motions correctly, but also identify the existence of unknown motion types.

A feature construction approach for genetic iterative rule learning algorithm

This paper presents a proposal that introduces the use of feature construction in a fuzzy rule learning algorithm. This is done by means of the combination of two different approaches together with a new learning strategy. The first of these two approaches consists of using relations in the antecedent of fuzzy rules while the second one employs functions in the antecedent of that rules. Thus, the method we propose tries to integrate these two models so that, using a learning strategy that allows us to start learning more general rules and finish the process learning more specific ones, we are able to increase the amount of information extracted from the initial variables. The experimental results show that the proposed method obtains a good trade-off among accuracy, interpretability and time needed to get the model in relation to the rest of algorithms using feature construction involved in the comparison.

Developing a hybrid intelligent model for forecasting problems: Case study of tourism demand time series

Forecasting tourism demand is a crucial issue in the tourism industry and is generally seen to be one of the most complex functions of tourism management. With the accurate forecasted trends and patterns that indicate the sizes, directions and characteristics of future international tourist flows, the government and private sectors can have a well-organized tourism strategy and provide a better infrastructure to serve the visitors and develop a suitable marketing strategy to gain benefit from the growing tourism. With the aim of developing accurate forecasting tools in the tourism industry, this study presents a new hybrid intelligent model that is called Modular Genetic-Fuzzy Forecasting System (MGFFS) by a combination of genetic fuzzy expert systems and data preprocessing. MGFFS is developed in three stage architecture. The first stage is data preprocessing. Some statistical tests are used to choose the key lags that are to be considered in the time series model. Then data transformation and K-means clustering have been applied to develop a modular model for reducing the complexity of the whole data space to become something more homogeneous. In the second stage, extraction of the TSK type fuzzy rule-based system for each cluster will be carried out by means of an efficient genetic learning algorithm that uses symbiotic evolution for fitness assignment. In the last stage, the testing data are first clustered and tourism demand forecasting is done by means of each cluster’s fuzzy system. Results show that forecasting accuracy of MGFFS is relatively better than other approaches in literature such as Classical Time Series models, Neuro-Fuzzy systems, and neural network, according to MAPE and RMSE evaluations. Powerful non-parametric statistical tests such as Friedman, Bonferroni, Holm and Hochberg are also used for comparing the performance of MGFFS with others. Based on the statistical tests, MGFFS is better than other models in accuracy and can be used as a suitable forecasting tool in tourism demand forecasting problems.

A new method for expert target recognition system: Genetic wavelet extreme learning machine (GAWELM)

In last year’s, the expert target recognition has been become very important topic in radar literature. In this study, a target recognition system is introduced for expert target recognition (ATR) using radar target echo signals of High Range Resolution (HRR) radars. This study includes a combination of an adaptive feature extraction and classification using optimum wavelet entropy parameter values. The features used in this study are extracted from radar target echo signals. Herein, a genetic wavelet extreme learning machine classifier model (GAWELM) is developed for expert target recognition. The GAWELM composes of three stages. These stages of GAWELM are genetic algorithm, wavelet analysis and extreme learning machine (ELM) classifier. In previous studies of radar target recognition have shown that the learning speed of feedforward networks is in general much slower than required and it has been a major disadvantage. There are two important causes. These are: (1) the slow gradient-based learning algorithms are commonly used to train neural networks, and (2) all the parameters of the networks are fixed iteratively by using such learning algorithms. In this paper, a new learning algorithm named extreme learning machine (ELM) for single-hidden layer feedforward networks (SLFNs) Ahern, Delisle, et al., 1989; Al-Otum & Al-Sowayan, 2011; Avci, Turkoglu, & Poyraz, 2005a, 2005b; Biswal, Dash, & Panigrahi, 2009; Frigui et al., in press; Cao, Lin, & Huang, 2010; Guo, Rivero, Dorado, Munteanu, & Pazos, 2011; Famili, Shen, Weber, & Simoudis, 1997; Han & Huang, 2006; Huang, Cai, Chen, & Liu, 2011; Huang, Chen, & Siew, 2006; Huang & Siew, 2005; Huang, Liu, Gao, & Guo, 2009; Jiang, Liu, Li, & Tang, 2011; Kubrusly & Levan, 2009; Le, Tamura, & Matsumoto, 2011; Lhermitte et al., 2011; Martínez-Martínez et al., 2011; Matlab, 2011; Nelson, Starzyk, & Ensley, 2002; Nejad & Zakeri, 2011; Tabib, Sathe, Deshpande, & Joshi, 2009; Tang, Sun, Tang, Zhou, & Wei, 2011, which randomly choose hidden nodes and analytically determines the output weights of SLFNs, to eliminate the these disadvantages of feedforward networks for expert target recognition area. Then, the genetic algorithm (GA) stage is used for obtaining the feature extraction method and finding the optimum wavelet entropy parameter values. Herein, the optimal one of four variant feature extraction methods is obtained by using a genetic algorithm (GA). The four feature extraction methods proposed GAWELM model are discrete wavelet transform (DWT), discrete wavelet transform–short-time Fourier transform (DWT–STFT), discrete wavelet transform–Born–Jordan time–frequency transform (DWT–BJTFT), and discrete wavelet transform–Choi–Williams time–frequency transform (DWT–CWTFT). The discrete wavelet transform stage is performed for optimum feature extraction in the time–frequency domain. The discrete wavelet transform stage includes discrete wavelet transform and calculating of discrete wavelet entropies. The extreme learning machine (ELM) classifier is performed for evaluating the fitness function of the genetic algorithm and classification of radar targets. The performance of the developed GAWELM expert radar target recognition system is examined by using noisy real radar target echo signals. The applications results of the developed GAWELM expert radar target recognition system show that this GAWELM system is effective in rating real radar target echo signals. The correct classification rate of this GAWELM system is about 90% for radar target types used in this study.

A FILTER PROPOSAL FOR INCLUDING FEATURE CONSTRUCTION IN A GENETIC LEARNING ALGORITHM

In system identification process often a predetermined set of features is used. However, in many cases it is difficult to know a priori whether the selected features were really the more appropriate ones. This is the reason why the feature construction techniques have been very interesting in many applications. Thus, the current proposal introduces the use of these techniques in order to improve the description of fuzzy rule-based systems. In particular, the idea is to include feature construction in a genetic learning algorithm. The construction of attributes in this study will be restricted to the inclusion of functions defined on the initial attributes of the system. Since the number of functions and the number of attributes can be very large, a filter model, based on the use of information measures, is introduced. In this way, the genetic algorithm only needs to explore the particular new features that may be of greater interest to the final identification of the system. In order to manage the knowledge provided by the new attributes based on the use of functions we propose a new model of rule by extending a basic learning fuzzy rule-based model. Finally, we show the experimental study associated with this work.

Good Prediction of Gas Price between MLFF and GMDH Neural Network

Studies show neural networks have better results in predicting of financial time series in comparison to any linear or non-linear functional form to model the price movement. Neural networks have the advantage of simulating the non-linear models when little a priori knowledge of the structure of problem domains exist or the number of immeasurable input variables are great and system has a chaotic characteristics. Among different methods, MLFF neural network with back-propagation learning algorithm and GMDH neural network with genetic learning algorithms are used to predict gas price of the Henry Hob database covering 01/01/2004-13/7/2009 period. This paper uses moving average crossover inputs and the results confirms (1) the fact there is short-term dependence in gas price movements, (2) the EMA moving average has better result and also (3) by means of the GMDH neural networks, prediction accuracy in comparison to MLFF neural networks, can be improved.

Extreme learning machine based genetic algorithm and its application in power system economic dispatch

In this paper a novel optimization algorithm, which utilizes the key ideas of both genetic algorithm (GA) and extreme learning machine (ELM), is proposed. Traditional genetic algorithm employs genetic operations, such as selection, mutation and crossover to generate the optimal solution. In practice, the child solutions generated by crossover and mutation are largely random and therefore cannot ensure the fast convergence of the algorithm. To tackle the weakness of traditional GA, the ELM is introduced to estimate the nonlinear functional relationships between the parent population and child population generated by genetic operations. The trained downward-climbing and upward-climbing ELMs are then employed to generate candidate solutions, which forms the new population together with the solutions given by genetic operations. The proposed algorithm is applied to the power system economic dispatch problem. As demonstrated in case studies, the modified genetic algorithm is able to locate local minima faster and escape from local minima with a greater probability. The proposed algorithm can therefore ensure the faster convergence and provide more economical dispatch plans.

Social Learning and Monetary Policy Rules

We analyse the effects of social learning in a monetary policy context. Social learning might be viewed as more descriptive of actual learning behaviour in complex market economies. In our model, Taylor Principle governs uniqueness and expectational stability of rational expectations equilibrium (REE) under homogeneous recursive algorithms. We find that the Taylor Principle is not necessary for convergence to REE minimum state variable (MSV) equilibrium under social learning. Sunspot equilibria exist in the indeterminate region. Our agents cannot co-ordinate on a sunspot equilibrium in general form specification, however, they can co-ordinate on common factor specification. We contribute to the use of genetic algorithm learning in stochastic environments.

A genetic optimization algorithm and perceptron learning rules for a bi-criteria parallel machine scheduling

This work considers scheduling problems minding the setup and removal times of jobs rather than processing times. For some production systems, setup times and removal times are so important to be considered independent of processing times. In general, jobs are performed according to the automatic machine processing in production systems, and the processing times are considered to be constant regardless of the process sequence. As the human factor can influence the setup and removal times, when the setup process is repetitive the setup times decreases. This fact is considered as learning effect in scheduling literature. In this study, a bi-criteria m-identical parallel machines scheduling problem with learning effects of setup and removal times is considered. The learning effect is proposed using a perceptron neural network algorithm. The objective function of the problem is minimization of the weighted sum of total earliness and tardiness. A mathematical programming model is developed for the problem, which is NP-hard. Results of computational tests show that the LINGO 9 software is effective in solving problems with up to 25 jobs and five machines. Therefore, for larger sized problems, a genetic algorithm for optimization is developed.

Adaptive Neuro-Fuzzy Controler With Genetic Training For Mobile Robot Control

In this paper, we investigate the use of adaptive techniques in the optimization of navigation of Khepera mobile robot in an unstructured and dynamic environment. We optimize the performance of our simplified fuzzy controller using neural network that utilizes genetic algorithm learning. The adaptation of the system involves the tuning of the control rules thereby trimming the control actions, and adjusting the fuzzy controller output gain. We realised an improved performance in our adaptive neuro-fuzzy controller with genetic training for various implemented behaviours on the robot.

Recursive Learning of Genetic Algorithm Featuring Incremental Attribute Learning for Higher Dimensional Classification Problems

Rule-based classifiers trained by Genetic Algorithms (GAs) have been one of the most prevailing solutions for pattern classification problem. This paper introduces an algorithm named Recursive Learning of Genetic Algorithms featuring Incremental Attribute Learning (RLGA-IAL) developed from the Recursive Learning of Genetic Algorithms with Task Decomposition and Varied Rule Set (RLGA). Instead of training all the attributes in a batch, RLGA-IAL integrates the attributes sequentially. By projecting a large multidimensional search space to single-dimensional space spaces with integration, it reduces the difficulty in deriving the classification rules.Though a series of experiments, RLGA-IAL shows a successful and promising performance in classification problems with the dimension of the datasets ranging from 5 to 60.

A combining genetic learning algorithm and risk matrix model using in optimal production program

One of the important issues for any enterprises is the compromise optimal solution between inverse of multi objective functions. The prediction of the production cost and/or profit per unit of a product and deal with two obverse functions at same time can be extremely difficult, especially if there is a lot of conflict information about production parameters. But the most important is how much risk of this compromise solution. For this reason, the research introduce and developed a strong and cabable model of genatic algorithim combining with risk management matrix to increase the quality of decisions as it is based on quantitive indicators, not on qualititive evaluation. Research results show that integration of genetic algorithm and risk management matrix model has strong significant in the decision making where it power and time to make the right decision and improve the quality of the decision making as well.

Portfolio analysis and asset management: simple genetic algorithm approach

The aim of this paper is to derive a modified computationally efficient genetic learning algorithm to derive the optimal rebalancing weights to engineer a structured portfolio out of different assets. The stochastic target function is formulated as an expected squared cost of hedging error, which is assumed to be partly and harmonically dependent on the governing Markovian process underlying the individual asset returns and partly on randomness. A simple haploid genetic algorithm is developed as an alternative numerical scheme, which is deemed to be computationally more efficient than numerically deriving an explicit solution to the formulated optimisation model.

An Efficient Inductive Genetic Learning Algorithm for Fuzzy Relational Rules

Abstract Fuzzy modelling research has traditionally focused on certain types of fuzzy rules. However, the use of alternative rule models could improve the ability of fuzzy systems to represent a specific problem. In this proposal, an extended fuzzy rule model, that can include relations between variables in the antecedent of rules is presented. Furthermore, a learning algorithm based on the iterative genetic approach which is able to represent the knowledge using this model is proposed as well. On the other hand, potential relations among initial variables imply an exponential growth in the feasible rule search space. Consequently, two filters for detecting relevant potential relations are added to the learning algorithm. These filters allows to decrease the search space complexity and increase the algorithm efficiency. Finally, we also present an experimental study to demonstrate the benefits of using fuzzy relational rules.

An active learning genetic algorithm for integrated process planning and scheduling

In traditional approaches, process planning and scheduling are carried out sequentially, where scheduling is done separately after the process plan has been generated. However, the functions of these two systems are usually complementary. The traditional approach has become an obstacle to improve the productivity and responsiveness of the manufacturing system. If the two systems can be integrated more tightly, greater performance and higher productivity of a manufacturing system can be achieved. Therefore, the research on the integrated process planning and scheduling (IPPS) problem is necessary. In this paper, a new active learning genetic algorithm based method has been developed to facilitate the integration and optimization of these two systems. Experimental studies have been used to test the approach, and the comparisons have been made between this approach and some previous approaches to indicate the adaptability and superiority of the proposed approach. The experimental results show that the proposed approach is a promising and very effective method on the research of the IPPS problem.

A Hybrid Automatic System for the Diagnosis of Lung Cancer Based on Genetic Algorithm and Fuzzy Extreme Learning Machines

An automatic system for the diagnosis of lung cancer has been proposed in this manuscript. The proposed method is based on combination of genetic algorithm (GA) for the feature selection and newly proposed approach, namely the extreme learning machines (ELM) for the classification of lung cancer data. The dimension of the feature space is reduced by the GA in this scheme and the effective features are selected in this way. The data are then fed to a fuzzy inference system (FIS) which is trained by the fuzzy extreme learning machines approach. The results on real data indicate that the proposed system is very effective in the diagnosis of lung cancer and can be used for clinical applications.