Acquisition of optimal attribute subset through genetic algorithm using GNP‐based class association rule mining

Attribute selection is a technique to prune less relevant information and discover high‐quality knowledge. It is especially useful for the classification of a large database, because the preprocessing of data increases the possibility that predictor attributes given to the mining algorithm become more relevant to the class attribute. In this paper, a method to acquire the optimal attribute subset for the genetic network programming (GNP) based class association rule mining has been proposed, and this attribute selection process using genetic algorithm (GA) leads to a higher accuracy for classification. Class association rule mining through GNP is conducted with a small subset of data rather than the original large number of attributes; thus simple but important rules are obtained for classification while the local optimal problem is avoided. Simulation results with educational data show that the classification accuracy is largely improved from 52.73 to 74.54%, when classification is made using the optimal attribute subset. © 2014 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.     

A nonlinear model to rank association rules based on semantic similarity and genetic network programing

Many methods have been studied for mining association rules efficiently. However, because these methods usually generate a large number of rules, it is still a heavy burden for the users to find the most interesting ones. In this paper, we propose a novel method for finding what the user is interested in by assigning several keywords, like searching documents on the Web using search engines. By considering both the semantic similarity between the rules and keywords, and the statistical information like support, confidence, chi-squared value, etc. we could rank the rules by a new method named RuleRank, where evolutionary methods are applied to find the optimal ranking model. Experiments show that our approach is effective for the users to find what they want. Copyright © 2009 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.     

Analysis of energy consumption of elevator group supervisory control system based on genetic network programming

Elevator Group Supervisory Control System (EGSCS) is a traffic system, which provides the transportation services for passengers in modern buildings. As the elevator systems include uncertainty due to the future arrival of the passengers, it difficult to model, analyze, and optimize the elevator group supervisory control system. Recently, artificial intelligence technology has been used in such complex systems. Genetic Network Programming(GNP), a graph‐based evolutionary method extended from genetic algorithm and genetic programming, has been already applied to EGSCS. On the other hand, since energy consumption is becoming one of the greatest challenges in the society, it should be taken as one of the criteria of the elevator operations. The elevators with maximum energy efficiency are therefore required. In this paper, the GNP is used to solve EGSCS with energy consumption (EC). Moreover, the idle car assignment has been embedded in the proposed method. Finally, the simulations show that some factors should be introduced into GNP in order to deal with the higher EC in the light traffic of the elevator systems. © 2011 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.     

Image denoising using pulse coupled neural network with an adaptive Pareto genetic algorithm

Recently, pulse coupled neural network (PCNN) attracts much attention in image denoising as a nonlinear filtering technique. The PCNN‐based anisotropic diffusion (PCNN‐AD) method has been proposed previously for flicker noise reduction and its effectiveness has been demonstrated. Using the visual characteristics of PCNN, PCNN‐AD has also solved the problem of AD that AD is not able to suppress the isolate noise. However, there are still two drawbacks in PCNN‐AD, that is, time consuming and PCNN parameters' estimation. In order to improve the efficiency and the denoising performance of PCNN‐AD, a PCNN‐based method with an adaptive Pareto genetic algorithm (GA‐PCNN) has been proposed to restrain from additive white Gaussian noise (AWGN) in this paper. GA‐PCNN firstly integrates the PCNN and AD as a parallel system, then, optimizes the parameters of a simplified PCNN by the adaptive Pareto GA. Experimental results indicate that GA‐PCNN has better performances than the previous denoising techniques, i.e. median filter, Wiener filter, AD filter, and PCNN‐AD. The effectiveness of GA‐PCNN on AWGN reduction and edge preservation are shown finally. The results will also contribute to denoising in CMOS image sensors in the future. © 2011 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.     

Evolving graph‐based chromosome by means of variable size genetic network programming with binomial distribution

Genetic network programming (GNP) is a graph‐based evolutionary algorithm with fixed size, which has been proven to solve complicated problems efficiently and effectively. In this paper, variable size genetic network programming (GNPvs) with binomial distribution has been proposed, which will change the size of the individuals and obtain their optimal size during evolution. The proposed method will select the number of nodes to move from one parent GNP to another parent GNP during crossover to implement the new feature of GNP. The probability of selecting the number of nodes to move satisfies a binomial distribution. The proposed method can keep the effectiveness of crossover, improve the performance of GNP, and find the optimal size of the individuals. The well‐known testbed Tileworld is used to show the numerical results in the simulations. © 2013 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.     

An extended probabilistic model building genetic network programming using both of good and bad individuals

Classical estimation of distribution algorithms (EDAs) generally use truncation selection to estimate the distribution of the good individuals while ignoring the bad ones. However, various researches in evolutionary algorithms (EAs) have reported that the bad individuals may affect and help solving the problem. This paper proposes a new method to use the bad individuals by studying the substructures rather than the entire individual structures to solve reinforcement learning (RL) problems, which generally factorize their entire solutions to the sequences of state–action pairs. This work was studied in a recent graph‐based EDA named probabilistic model building genetic network programming (PMBGNP), which could solve RL problems successfully, to propose an extended PMBGNP. The effectiveness of this work is verified in an RL problem, namely robot control. Compared to other related work, results show that the proposed method can significantly speed up the evolution efficiency. © 2013 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.     

Acoustical parameters of polyethylene glycol/water mixtures

Ultrasonic velocity (v) and density (d) have been measured for polyethylene glycol/water mixtures at 30°C. The adiabatic compressibility (β _ad), molar compressibility (β), specific acoustic impendance (Z), Rao number (R) and van der Waals constant (b) have been computed. The variations ofv, d, β _ad,β, Z, R andb with mole ratio of water/ether group oxygen have been studied. The association between the components and the formation of tetrahydrate have been reported.     

A genetic network programming with learning approach for enhanced stock trading model

In this paper, an enhancement of stock trading model using Genetic Network Programming (GNP) with Sarsa Learning is described. There are three important points in this paper: First, we use GNP with Sarsa Learning as the basic algorithm while both Technical Indices and Candlestick Charts are introduced for efficient stock trading decision-making. In order to create more efficient judgment functions to judge the current stock price appropriately, Importance Index (IMX) has been proposed to tell GNP the timing of buying and selling stocks. Second, to improve the performance of the proposed GNP-Sarsa algorithm, we proposed a new method that can learn the appropriate function describing the relation between the value of each technical index and the value of the IMX. This is an important point that devotes to the enhancement of the GNP-Sarsa algorithm. The third point is that in order to create more efficient judgment functions, sub-nodes are introduced in each node to select appropriate stock price information depending on the situations and to determine appropriate actions (buying/selling). To confirm the effectiveness of the proposed method, we carried out the simulation and compared the results of GNP-Sarsa with other methods like GNP with Actor Critic, GNP with Candlestick Chart, GA and Buy&Hold method. The results shows that the stock trading model using GNP-Sarsa outperforms all the other methods.     

Adaptability analysis of genetic network programming with reinforcement learning in dynamically changing environments

Genetic network programming (GNP) has been proposed as one of the evolutionary algorithms and extended with reinforcement learning (GNP-RL). The combination of evolution and learning can efficiently evolve programs and the fitness improvement has been confirmed in the simulations of tileworld problems, elevator group supervisory control systems, stock trading models and wall following behavior of Khepera robot. However, its adaptability in testing environments, where the situations dynamically change, has not been analyzed in detail yet. In this paper, the adaptation mechanism in the testing environment is introduced and it is confirmed that GNP-RL can adapt to the environmental changes using a robot simulator WEBOTS, especially when unexperienced sensor troubles suddenly occur. The simulation results show that GNP-RL works well in the testing even if wrong sensor information is given because GNP-RL has a function to automatically change programs using alternative actions. In addition, the analysis on the effects of the parameters of GNP-RL is carried out in both training and testing simulations.     

Genetic relation algorithm with guided mutation for the large-scale portfolio optimization

The survey of the relevant literatures shows that there have been many studies for portfolio optimization problems and that the number of studies which have investigated the optimum portfolio using evolutionary computation is quite large. But, almost none of these studies deals with genetic relation algorithm (GRA), where GRA is one of the evolutionary methods with graph structure. This study presents an approach to large-scale portfolio optimization problems using GRA with a new operator, called guided mutation. In order to pick up the most efficient portfolio, GRA considers the correlation coefficient between stock brands as strength, which indicates the relation between nodes in each individual of GRA. Guided mutation generates offspring according to the average value of correlation coefficients in each individual, which means to enhance the exploitation ability of evolution of GRA. A genetic relation algorithm with guided mutation (GRA/G) for the portfolio optimization is proposed in this paper. Genetic network programming (GNP), which was proposed in our previous research, is used to validate the performance of the portfolio generated with GRA/G. The results show that GRA/G approach is successful in portfolio optimization.