Abstract: Although the Differential Evolution (DE) algorithm has been shown to be a simple yet powerful evolutionary algorithm for optimizing continuous functions, users are still faced with the problem of preliminary testing and hand-tuning of the evolutionary parameters prior to commencing the actual optimization process. As a solution, self-adaptation has been found to be highly beneficial in automatically and dynamically adjusting evolutionary parameters such as crossover rates and mutation rates. In this paper, we present a first attempt at self-adapting the population size parameter in addition to self-adapting crossover and mutation rates. Firstly, our main objective is to demonstrate the feasibility of self-adapting the population size parameter in DE. Using De Jong's F1–F5 benchmark test problems, we showed that DE with self-adaptive populations produced highly competitive results compared to a conventional DE algorithm with static populations. In addition to reducing the number of parameters used in DE, the proposed algorithm actually outperformed the conventional DE algorithm for one of the test problems. It was also found that that an absolute encoding methodology for self-adapting population size in DE produced results with greater optimization reliability compared to a relative encoding methodology.

Abstract: This paper compares various training methods available for training multi-layer perceptron (MLP) type of artificial neural networks (ANNs) for modelling the rainfall-runoff process. The training methods investigated include the popular back-propagation algorithm (BPA), real-coded genetic algorithm (RGA), and a self-organizing map (SOM). A SOM was used to first classify the input-output space into different categories and then develop feed-forward MLP models for each category using BPA. The daily average rainfall and streamflow data derived from an existing catchment were employed to develop all ANN models investigated in this study. A wide variety of standard statistical performance evaluation measures were employed to evaluate the performances of various ANN models developed. The results obtained in this study indicate that the approach of first classifying the input-output space into different categories using SOM and then developing separate ANN models for different classes trained using BPA performs better than the approach of developing a single ANN rainfall-runoff model trained using BPA. The ANN rainfall-runoff model trained using RGA was able to provide a better generalization of the complex, dynamic, non-linear, and fragmented rainfall-runoff process in comparison with the other approaches investigated in this study. It has been found that the RGA trained ANN model significantly outperformed the ANN model trained using BPA, and was also able to overcome certain limitations of the ANN rainfall-runoff model trained using BPA reported by many researchers in the past. It is noted that the performances of various ANN models should to be evaluated using a wide variety of statistical performance indices rather than relying on a few global error statistics normally employed that are similar in nature to the global error minimized at the output layer of an ANN.

Abstract: A multilayer neural network based on multi-valued neurons (MLMVN) is considered in the paper. A multi-valued neuron (MVN) is based on the principles of multiple-valued threshold logic over the field of the complex numbers. The most important properties of MVN are: the complex-valued weights, inputs and output coded by the kth roots of unity and the activation function, which maps the complex plane into the unit circle. MVN learning is reduced to the movement along the unit circle, it is based on a simple linear error correction rule and it does not require a derivative. It is shown that using a traditional architecture of multilayer feedforward neural network (MLF) and the high functionality of the MVN, it is possible to obtain a new powerful neural network. Its training does not require a derivative of the activation function and its functionality is higher than the functionality of MLF containing the same number of layers and neurons. These advantages of MLMVN are confirmed by testing using parity n, two spirals and "sonar" benchmarks and the Mackey---Glass time series prediction.

Abstract: In this paper a new algorithm for edge detection using ant colony search is proposed. The problem is represented by a directed graph in which nodes are the pixels of an image. To adapt the problem, some modifications on original ant colony search algorithm (ACSA) are applied. A large number of experiments are employed to determine suitable algorithm parameters. We drive an experimental relationship between the size of the image to be analyzed and algorithm parameters. Several experiments are made and the results suggest the effectiveness of the proposed algorithm.

Abstract: A study is presented to compare the performance of three types of artificial neural network (ANN), namely, multi layer perceptron (MLP), radial basis function (RBF) network and probabilistic neural network (PNN), for bearing fault detection. Features are extracted from time domain vibration signals, without and with preprocessing, of a rotating machine with normal and defective bearings. The extracted features are used as inputs to all three ANN classifiers: MLP, RBF and PNN for two- class (normal or fault) recognition. Genetic algorithms (GAs) have been used to select the characteristic parameters of the classifiers and the input features. For each trial, the ANNs are trained with a subset of the experimental data for known machine conditions. The ANNs are tested using the remaining set of data. The procedure is illustrated using the experimental vibration data of a rotating machine. The roles of different vibration signals and preprocessing techniques are investigated. The results show the effectiveness of the features and the classifiers in detection of machine condition.

Abstract: This paper presents a novel approach by combining SOM and fuzzy rule base for flow time prediction in semiconductor manufacturing factory. Flow time of a new order is highly related to the shop floor status; however, the semiconductor manufacturing processes are highly complicated and involve more than hundred of production steps. There is no governing function identified so far among the flow time of a new order and these shop flow status. Therefore, a simulation model which mimics the production process of a real wafer fab located in Hsin-Chu Science-based Park of Taiwan is built and flow time and related shop floor status are collected and fed into the SOM for classification. Then, corresponding fuzzy rule base is selected and applied for flow time prediction. Genetic process is further applied to fine-tune the composition of the rule base. Finally, using the simulated data, the effectiveness of the proposed method is shown by comparing with other approaches.

Abstract: An investigation is conducted on two well-known similarity-based learning approaches to text categorization: the k-nearest neighbors (kNN) classifier and the Rocchio classifier. After identifying the weakness and strength of each technique, a new classifier called the kNN model-based classifier (kNN Model) is proposed. It combines the strength of both kNN and Rocchio. A text categorization prototype, which implements kNN Model along with kNN and Rocchio, is described. An experimental evaluation of different methods is carried out on two common document corpora: the 20-newsgroup collection and the ModApte version of the Reuters-21578 collection of news stories. The experimental results show that the proposed kNN model-based method outperforms the kNN and Rocchio classifiers, and is therefore a good alternative for kNN and Rocchio in some application areas.

Abstract: In this paper, a genetic algorithm (GA) is proposed as a search strategy for not only positive but also negative quantitative association rule (AR) mining within databases. Contrary to the methods used as usual, ARs are directly mined without generating frequent itemsets. The proposed GA performs a database-independent approach that does not rely upon the minimum support and the minimum confidence thresholds that are hard to determine for each database. Instead of randomly generated initial population, uniform population that forces the initial population to be not far away from the solutions and distributes it in the feasible region uniformly is used. An adaptive mutation probability, a new operator called uniform operator that ensures the genetic diversity, and an efficient adjusted fitness function are used for mining all interesting ARs from the last population in only single run of GA. The efficiency of the proposed GA is validated upon synthetic and real databases.

Abstract: An intense research around classifier fusion in recent years revealed that combining performance strongly depends on careful selection of classifiers to be combined Classifier performance depends, in turn, on careful selection of features, which could be further restricted by the subspaces of the data domain On the other hand, there is already a number of classifier fusion techniques available and the choice of the most suitable method depends back on the selections made within classifier, features and data spaces In all these multidimensional selection tasks genetic algorithms (GA) appear to be one of the most suitable techniques providing reasonable balance between searching complexity and the performance of the solutions found In this work, an attempt is made to revise the capability of genetic algorithms to be applied to selection across many dimensions of the classifier fusion process including data, features, classifiers and even classifier combiners In the first of the discussed models the potential for combined classification improvement by GA-selected weights for the soft combining of classifier outputs has been investigated The second of the proposed models describes a more general system where the specifically designed GA is applied to selection carried out simultaneously along many dimensions of the classifier fusion process Both, the weighted soft combiners and the prototype of the three-dimensional fusion-classifier-feature selection model have been developed and tested using typical benchmark datasets and some comparative experimental results are also presented

Abstract: In this paper, we present a new approach for training set selection in large size data sets. The algorithm consists on the combination of stratification and evolutionary algorithms. The stratification reduces the size of domain where the selection is applied while the evolutionary method selects the most representative instances. The performance of the proposal is compared with seven non-evolutionary algorithms, in stratified execution. The analysis follows two evaluating approaches: balance between reduction and accuracy of the subsets selected, and balance between interpretability and accuracy of the representation models associated to these subsets. The algorithms have been assessed on large and huge size data sets. The study shows that the stratified evolutionary instance selection consistently outperforms the non-evolutionary ones. The main advantages are: high instance reduction rates, high classification accuracy and models with high interpretability.

Abstract: This paper presents new findings in the design and application of biologically plausible neural networks based on spiking neuron models, which represent a more plausible model of real biological neurons where time is considered as an important feature for information encoding and processing in the brain. The design approach consists of an evolutionary strategy based supervised training algorithm, newly developed by the authors, and the use of different biologically plausible neuronal models. A dynamic synapse (DS) based neuron model, a biologically more detailed model, and the spike response model (SRM) are investigated in order to demonstrate the efficacy of the proposed approach and to further our understanding of the computing capabilities of the nervous system. Unlike the conventional synapse, represented as a static entity with a fixed weight, employed in conventional and SRM-based neural networks, a DS is weightless and its strength changes upon the arrival of incoming input spikes. Therefore its efficacy depends on the temporal structure of the impinging spike trains. In the proposed approach, the training of the network free parameters is achieved using an evolutionary strategy where, instead of binary encoding, real values are used to encode the static and DS parameters which underlie the learning process. The results show that spiking neural networks based on both types of synapse are capable of learning non-linearly separable data by means of spatio-temporal encoding. Furthermore, a comparison of the obtained performance with classical neural networks (multi-layer perceptrons) is presented.

Abstract: Recently, Adaboost has been compared to greedy backfitting of extended additive models in logistic regression problems, or “Logitboost". The Adaboost algorithm has been applied to learn fuzzy rules in classification problems, and other backfitting algorithms to learn fuzzy rules in modeling problems but, up to our knowledge, there are not previous works that extend the Logitboost algorithm to learn fuzzy rules in classification problems.In this work, Logitboost is applied to learn fuzzy rules in classification problems, and its results are compared with that of Adaboost and other fuzzy rule learning algorithms. Contradicting the expected results, it is shown that the basic extension of the backfitting algorithm to learn classification rules may produce worse results than Adaboost does. We suggest that this is caused by the stricter requirements that Logitboost demands to the weak learners, which are not fulfilled by fuzzy rules. Finally, it is proposed a prefitting based modification of the Logitboost algorithm that avoids this problem

Abstract: An algorithm of vague fault-tree analysis is proposed in this paper to calculate fault interval of system components from integrating expert's knowledge and experience in terms of providing the possibility of failure of bottom events. We also modify Tanaka et al's definition and extend the new usage on vague fault-tree analysis in terms of finding most important basic system component for managerial decision-making. In numerical verification, the fault of automatic gun is presented as a numerical example. For advanced experiment, a fault tree for the reactor protective system is adopted as simulation example and we compare the results with other methods. This paper also develops vague fault tree decision support systems (VFTDSS) to generate fault-tree, fault-tree nodes, then directly compute the vague fault-tree interval, traditional reliability, and vague reliability interval.

Abstract: We characterize the class of copulas that can be constructed from the diagonal section by means of the functional equation C(x,y)+|x−y|=C(x∨y,x∨y), for all (x,y) in the unit square such that C(x,y)>0. Some statistical properties of this class are given.

Abstract: Accurate software estimation such as cost estimation, quality estimation and risk analysis is a major issue in software project management. In this paper, we present a soft computing framework to tackle this challenging problem. We first use a preprocessing neuro-fuzzy inference system to handle the dependencies among contributing factors and decouple the effects of the contributing factors into individuals. Then we use a neuro-fuzzy bank to calibrate the parameters of contributing factors. In order to extend our framework into fields that lack of an appropriate algorithmic model of their own, we propose a default algorithmic model that can be replaced when a better model is available. One feature of this framework is that the architecture is inherently independent of the choice of algorithmic models or the nature of the estimation problems. By integrating neural networks, fuzzy logic and algorithmic models into one scheme, this framework has learning ability, integration capability of both expert knowledge and project data, good interpretability, and robustness to imprecise and uncertain inputs. Validation using industry project data shows that the framework produces good results when used to predict software cost.

Abstract: In the frame of designing a knowledge discovery system, we have developed stochastic models based on high-order hidden Markov models. These models are capable to map sequences of data into a Markov chain in which the transitions between the states depend on the n previous states according to the order of the model. We study the process of achieving information extraction from spatial and temporal data by means of an unsupervised classification. We use therefore a French national database related to the land use of a region, named Ter Uti, which describes the land use both in the spatial and temporal domain. Land-use categories (wheat, corn, forest, ...) are logged every year on each site regularly spaced in the region. They constitute a temporal sequence of images in which we look for spatial and temporal dependencies. The temporal segmentation of the data is done by means of a second-order Hidden Markov Model (HMM2) that appears to have very good capabilities to locate stationary segments, as shown in our previous work in speech recognition. The spatial classification is performed by defining a fractal scanning of the images with the help of a Hilbert---Peano curve that introduces a total order on the sites, preserving the relation of neighborhood between the sites. We show that the HMM2 performs a classification that is meaningful for the agronomists. Spatial and temporal classification may be achieved simultaneously by means of a two levels HMM2 that measures the a posteriori probability to map a temporal sequence of images onto a set of hidden classes.

Abstract: This paper focuses on developing a new multiple criteria decision-making (MCDM) methodology for subdivision arrangement of Ro-Ro vessels, which integrates the multi-objective optimisation with a fuzzy multi-attributive group decision-making (FMAGDM) technique. The study concentrates on the task of finding and then evaluating (or ranking) the finite number of pareto-optimal design alternatives (PODAs). A genetic algorithm based multi-objective optimisation technique, namely MOGA, is employed for optimisation purpose in terms of chosen design parameters such as damage stability, survivability, and cargo capacity. MOGA is the methodology where the solution space is searched for a set of PODAs, from which experts can express their opinions and choose the best PODA. The subjectiveness and imprecision of the ranking process is modelled as linear trapezoidal fuzzy numbers by means of linguistic terms. An attribute based aggregation technique for homogeneous and heterogeneous groups of experts is employed and used for dealing with the fuzzy opinion aggregation. Finally, a real case study on the subdivision arrangement of a Ro-Ro vessel is conducted to illustrate the effectiveness of the approach.

Abstract: The paper addresses the problem of extracting acronyms and their expansions from text. We propose a support vector machines (SVM) based approach to deal with the problem. First, all likely acronyms are identified using heuristic rules. Second, expansion candidates are generated from surrounding text of acronyms. Last, SVM model is employed to select the genuine expansions. Analysis shows that the proposed approach has the advantages of saving over the conventional rule based approaches. Experimental results show that our approach outperforms the baseline method of using rules. We also show that the trained SVM model is generic and can adapt to other domains easily.

Abstract: The technique of direct and inverse fuzzy (F-)transforms of three different types is introduced and approximating properties of the inverse F-transforms are described. A method of lossy image compression and reconstruction on the basis of the F-transform is presented.

Abstract: This paper presents a novel evolutionary design approach in a multi-agent design environment. Multi-agent system architecture offers a promising framework for dynamically managing cooperative agents in a distributed environment while the tree structure based generic algorithm provides a foundation for supporting evolutionary and innovative design abilities. Design is a complex knowledge discovery process. Creative design is a human trait that is not easily converted into a computational tool. Rather than to implement the innovative design by computers, this environment is used to stimulate the imagination of designers and extend their thinking space. It wants to explore a feasible and useful evolutionary approach in a distributed environment that will give the designers concrete help for the creative designs. This approach is illustrated by a mobile phone design example, which used binary algebraic expression tree to form sketch shapes and a feature based product tree to produce component combination choices. Because evolution is guided by human selectors, the evolutionary algorithm is not complex. It shows that approach is able to generate some creative solutions, demonstrating the power of explorative evolution.

Abstract: This paper presents two novel and useful defuzzification methods for fuzzy set outputs. Two algorithms based on root mean square (RMS) to obtain a new defuzzification procedure are proposed. In order to validate the efficacy of the proposed algorithms the results are compared with the existing defuzzification methods such as weighted average, centroid (COG) and mean of maxima. The satisfaction of a set of essential constraints is also dealt with which motivates a step towards rational defuzzification algorithm. These new methods RMS1and RMS2 stand on par with the most commonly used COG method in every respect. In addition, the value obtained by RMS2 is always higher and hence when a higher value is needed or desirable this can be employed advantageously.

Abstract: In the study of weighted fuzzy production rules (WFPRs) reasoning, we often need to consider those rules whose consequences are represented by two or more propositions connected by “AND” or “OR”. To enhance the representation capability of those rules, this paper proposes two types of knowledge representation parameters, namely, the input weight and the output weight, for a rule. A Generalized Fuzzy Petri Net (GFPN) is also presented for WFPR reasoning. Furthermore, this paper gives a similarity measure to improve the evaluation method of WFPRs and the multilevel fuzzy reasoning in which the consequences and their certainty factors are deduced synchronously by using a GFPN.

Abstract: Evolutionary relationships among species are usually (1) illustrated by means of a phylogenetic tree and (2) inferred by optimising some measure of fitness, such as the total evolutionary distance between species or the likelihood of the tree (given a model of the evolutionary process and a data set). The combinatorial complexity of inferring the topology of the best tree makes phylogenetic inference an ideal candidate for evolutionary algorithms. However, difficulties arise when different data sets provide conflicting information about the inferred `best' tree(s). We apply the techniques of multi-objective optimisation to phylogenetic inference for the first time. We use the simplest model of evolution and a four species problem to illustrate the method.

Abstract: Information systems are one of the most rapidly changing and vulnerable systems, where security is a major issue. The number of security-breaking attempts originated inside the organizations are increasing steadily. Attacks made in this way, usually done by ``authorized'' users of the system, cannot be immediately traced. As the idea of filtering the traffic at the entrance door, by using firewalls and the like, is not completely successful, the use of intrusion detection systems should be considered to increase the defense capacity of an information system. This paper presents a framework for a statistical anomaly prediction system using a neuro-genetic forecasting model, which predicts unauthorized invasions of user, based on previous observations and takes further action before intrusion occurs. In this paper, we propose an evolutionary time-series model for short-term database intrusion forecasting using genetic algorithm owing to its global search capability. The experimental results show that the combination strategy(neuro-genetic) can quicken the learning speed of the network and improve the predicting precision compared to the traditional artificial neural network. This paper also focuses on detecting significant changes of transaction intensity for intrusion prediction. The experimental study is performed using real time data provided by a major Corporate Bank. Furthermore, a comparative evaluation of the proposed neuro-genetic model with the traditional feed-forward network trained by the back-propagation with momentum and adaptive learning rate using sum square error on a prediction data set has been presented and a better prediction accuracy has been observed.

Abstract: Guessing the degree of difficulty of a problem before seeing its solution is notoriously hard not only for beginners, but also for the most experienced professionals Can we develop a method to evaluate, in some objective way, the difficulty of an open problem? This note proposes such a measure which can be used for a fairly large class of finitely refutable conjectures which includes, for example, Riemann Hypothesis and the Goldbach’s Conjecture According to our measure, Riemann Hypothesis is more complex than Goldbach’s Conjecture We also show, in a nonconstructive way, that the Collatz 3x+1 Problem is finitely refutable; consequently, our method cannot be applied, hence stronger versions of this problem are studied

Abstract: We present a novel neural model for image compression called the direct classification (DC) model. The DC is a hybrid between a subset of the self-organizing Kohonen (SOK) model and the adaptive resonance theory (ART) model. The DC is a fast and efficient neural classification engine. The DC training utilizes the accuracy of the winner-takes-all feature of the SOK model and the elasticity/speed of the ART1 model. The DC engine has experimentally achieved much better results than the state-of-the-art peer image compression techniques (e.g., JPEG2000 and DjVu wavelet technology) especially in the domains of colored documents and still satellite images. We include a comprehensive analysis of the most important parameters of our DC system and their effects on system performance.