An Effective Hybrid Semi-Parametric Regression Strategy for Rainfall Forecasting Combining Linear and Nonlinear Regression

Rainfall forecasting is an important research topic in disaster prevention and reduction. The characteristic of rainfall involves a rather complex systematic dynamics under the influence of different meteorological factors, including linear and nonlinear pattern. Recently, many approaches to improve forecasting accuracy have been introduced. Artificial neural network (ANN), which performs a nonlinear mapping between inputs and outputs, has played a crucial role in forecasting rainfall data. In this paper, an effective hybrid semi-parametric regression ensemble (SRE) model is presented for rainfall forecasting. In this model, three linear regression models are used to capture rainfall linear characteristics and three nonlinear regression models based on ANN are able to capture rainfall nonlinear characteristics. The semi-parametric regression is used for ensemble model based on the principal component analysis technique. Empirical results reveal that the prediction using the SRE model is generally better than those obtained using other models in terms of the same evaluation measurements. The SRE model proposed in this paper can be used as a promising alternative forecasting tool for rainfall to achieve greater forecasting accuracy and improve prediction quality.

Recursive Learning of Genetic Algorithms with Task Decomposition and Varied Rule Set

Rule-based Genetic Algorithms (GAs) have been used in the application of pattern classification (Corcoran & Sen, 1994), but conventional GAs have weaknesses. First, the time spent on learning is long. Moreover, the classification accuracy achieved by a GA is not satisfactory. These drawbacks are due to existing undesirable features embedded in conventional GAs. The number of rules within the chromosome of a GA classifier is usually set and fixed before training and is not problem-dependent. Secondly, conventional approaches train the data in batch without considering whether decomposition solves the problem. Thirdly, when facing large-scale real-world problems, GAs cannot utilise resources efficiently, leading to premature convergence. Based on these observations, this paper develops a novel algorithmic framework that features automatic domain and task decomposition and problem-dependent chromosome length (rule number) selection to resolve these undesirable features. The proposed Recursive Learning of Genetic Algorithm with Task Decomposition and Varied Rule Set (RLGA) method is recursive and trains and evolves a team of learners using the concept of local fitness to decompose the original problem into sub-problems. RLGA performs better than GAs and other related solutions regarding training duration and generalization accuracy according to the experimental results.

Parallel Single and Multiple Objectives Genetic Algorithms

This paper critically reviews the reported research on parallel single and multi-objective genetic algorithms. Many early efforts on single and multi-objective genetic algorithms were introduced to reduce the processing time needed to reach an acceptable solution. However, some parallel single and multi-objective genetic algorithms converged to better solutions as compared to comparable sequential single and multiple objective genetic algorithms. The authors review several representative models for parallelizing single and multi-objective genetic algorithms. Further, some of the issues that have not yet been studied systematically are identified in the context of parallel single and parallel multi-objective genetic algorithms. Finally, some of the potential applications of parallel multi-objective GAs are discussed.

Differential Operators Embedded Artificial Bee Colony Algorithm

Artificial Bee Colony (ABC) is one of the most recent nature inspired (NIA) algorithms based on swarming metaphor. Proposed by Karaboga in 2005, ABC has proven to be a robust and efficient algorithm for solving global optimization problems over continuous space. However, it has been observed that the structure of ABC is such that it supports exploration more in comparison to exploitation. In order to maintain a balance between these two antagonist factors, this paper suggests incorporation of differential evolution (DE) operators in the structure of basic ABC algorithm. The proposed algorithm called DE-ABC is validated on a set of 10 benchmark problems and the numerical results are compared with basic DE and basic ABC algorithm. The numerical results indicate that the presence of DE operators help in a significant improvement in the performance of ABC algorithm.

Feature Selection Based on Minimizing the Area Under the Detection Error Tradeoff Curve

Feature selection is crucial to select an “optimized” subset of features from the original feature set based on a certain objective function. In general, feature selection removes redundant or irrelevant data while retaining classification accuracy. This paper proposes a feature selection algorithm that aims to minimize the area under the curve of detection error trade-off (DET) curve. Particle swarm optimization (PSO) is employed to search for the optimal feature subset. The proposed method is implemented in face recognition and iris recognition systems. The result shows that the proposed method is able to find an optimal subset of features that sufficiently describes iris and face images by removing unwanted and redundant features and at the same time improving the classification accuracy in terms of total error rate (TER).

LZW Encoding in Genetic Algorithm

To solve a problem using Genetic Algorithms (GAs), a solution must be encoded into a binary string. The length of the binary string represents the size of the problem. As the length of the binary string increases, the size of the search space also increases at an exponential rate. To reduce the search space, one approach is to use a compressed encoding chromosome. This paper presents a genetic algorithm, called LZWGA, that uses compressed chromosomes. An LZWGA chromosome must be decompressed using an LZW decompression algorithm before its fitness can be evaluated. By using compressed encoding, the search space is reduced dramatically. For one-million-bit problem, the search space of the original problem is 21000000 or about 9.90x10301029 points. When using a compressed encoding, the search space was reduced to 8.37x10166717 points. LZWGA can solve one-million-bit OneMax, RoyalRoad, and Trap functions.

Experimental Study on Recent Advances in Differential Evolution Algorithm

The Differential Evolution (DE) is a well known Evolutionary Algorithm (EA), and is popular for its simplicity. Several novelties have been proposed in research to enhance the performance of DE. This paper focuses on demonstrating the performance enhancement of DE by implementing some of the recent ideas in DE’s research viz. Dynamic Differential Evolution (dDE), Multiple Trial Vector Differential Evolution (mtvDE), Mixed Variant Differential Evolution (mvDE), Best Trial Vector Differential Evolution (btvDE), Distributed Differential Evolution (diDE) and their combinations. The authors have chosen fourteen variants of DE and six benchmark functions with different modality viz. Unimodal Separable, Unimodal Nonseparable, Multimodal Separable, and Multimodal Nonseparable. On analyzing distributed DE and mixed variant DE, a novel mixed-variant distributed DE is proposed whereby the subpopulations (islands) employ different DE variants to cooperatively solve the given problem. The competitive performance of mixed-variant distributed DE on the chosen problem is also demonstrated. The variants are well compared by their mean objective function values and probability of convergence.

Simultaneous Tolerance Synthesis for Manufacturing and Quality using Evolutionary Algorithms

Tolerance plays a major role in the manufacturing industry, as it affects product design, manufacturing, and quality of the product. This paper considers product design, manufacturing, and quality simultaneously, and introduces a concurrent approach for tolerance allocation using evolutionary algorithms. A non-linear model that minimizes the combination of manufacturing cost and quality loss simultaneously, in a single objective function has been considered. In the proposed work, evolutionary algorithms (that is, Genetic Algorithms (GA), Differential Evolution (DE), and Particle Swarm Optimization (PSO)) have been used to determine the optimal tolerances at the minimum manufacturing and quality loss cost. The application of the proposed methodology has been demonstrated on a simple mechanical assembly.

Application of Genetic Algorithm to Minimize the Number of Objects Processed and Setup in a One-Dimensional Cutting Stock Problem

This paper presents the application of the one new approach using Genetic Algorithm in solving One-Dimensional Cutting Stock Problems in order to minimize two objectives, usually conflicting, i.e., the number of processed objects and setup while simultaneously treating them as a single goal. The model problem, the objective function, the method denominated SingleGA10 and the steps used to solve the problem are also presented. The obtained results of the SingleGA10 are compared to the following methods: SHP, Kombi234, ANLCP300 and Symbio10, found in literature, verifying its capacity to find feasible and competitive solutions. The computational results show that the proposed method, which only uses a genetic algorithm to solve these two objectives inversely related, provides good results.

Usage Profile Generation from Web Usage Data Using Hybrid Biclustering Algorithm

Biclustering has the potential to make significant contributions in the fields of information retrieval, web mining, and so forth. In this paper, the authors analyze the complex association between users and pages of a web site by using a biclustering algorithm. This method automatically identifies the groups of users that show similar browsing patterns under a specific subset of the pages. In this paper, mutation operator from Genetic Algorithms is incorporated into the Binary Particle Swarm Optimization (BPSO) for biclustering of web usage data. This hybridization can increase the diversity of the population and help the particles effectively escape from the local optimum. It detects optimized user profile group according to coherent browsing behavior. Experiments are performed on a benchmark clickstream dataset to test the effectiveness of the proposed algorithm. The results show that the proposed algorithm has higher performance than existing PSO methods. The interpretation of this biclustering results are useful for marketing and sales strategies.