THE DESIGN AND ANALYSIS OF EXPERIMENTS. By Oscar Kempthorne. New York, John Wiley and Sons, Inc., 1952. 631 pp. $8.50. This book by a teacher of statistics (as well as a consultant for \"experimenters\") is a comprehensive study of the philosophical background for the statistical design of experiment. It is necessary to have some facility with algebraic notation and manipulation to be able to use the volume intelligently. The problems are presented from the theoretical point of view, without such practical examples as would be helpful for those not acquainted with mathematics. The mathematical justification for the techniques is given. As a somewhat advanced treatment of the design and analysis of experiments, this volume will be interesting and helpful for many who approach statistics theoretically as well as practically. With emphasis on the \"why,\" and with description given broadly, the author relates the subject matter to the general theory of statistics and to the general problem of experimental inference. MARGARET J. ROBERTSON

The Design and Analysis of Experiments

Grey Wolf Optimizer

Data Mining Practical Machine Learning Tools and Techniques

The Whale Optimization Algorithm

Differential evolution (DE) is arguably one of the most powerful stochastic real-parameter optimization algorithms in current use. DE operates through similar computational steps as employed by a standard evolutionary algorithm (EA). However, unlike traditional EAs, the DE-variants perturb the current-generation population members with the scaled differences of randomly selected and distinct population members. Therefore, no separate probability distribution has to be used for generating the offspring. Since its inception in 1995, DE has drawn the attention of many researchers all over the world resulting in a lot of variants of the basic algorithm with improved performance. This paper presents a detailed review of the basic concepts of DE and a survey of its major variants, its application to multiobjective, constrained, large scale, and uncertain optimization problems, and the theoretical studies conducted on DE so far. Also, it provides an overview of the significant engineering applications that have benefited from the powerful nature of DE.

/pdf/differential-evolution-a-survey-of-the-state-of-the-art-3lz840pz2t.pdf

Differential Evolution: A Survey of the State-of-the-Art

Metaheuristics and Parallelism

http://www.lcc.uma.es/~eat/pdf/cor08a.pdf

DNA fragment assembly using a grid-based genetic algorithm

Software Product Lines (SPLs) are families of related software products, each with its own set of feature combinations. Their commonly large number of products poses a unique set of challenges for software testing as it might not be technologically or economically feasible to test of all them individually. SPL pairwise testing aims at selecting a set of products to test such that all possible combinations of two features are covered by at least one selected product. Most approaches for SPL pairwise testing have focused on achieving full coverage of all pairwise feature combinations with the minimum number of products to test. Though useful in many contexts, this single-objective perspective does not reflect the prevailing scenario where software engineers do face trade-offs between the objectives of maximizing the coverage or minimizing the number of products to test. In contrast and to address this need, our work is the first to propose a classical multi-objective formalisation where both objectives are equally important. In this paper, we study the application to SPL pairwise testing of four classical multi-objective evolutionary algorithms. We developed three seeding strategies - techniques that leverage problem domain knowledge - and measured their performance impact on a large and diverse corpus of case studies using two well-known multi-objective quality measures. Our study identifies the performance differences among the algorithms and corroborates that the more domain knowledge leveraged the better the search results. Our findings enable software engineers to select not just one solution (as in the case of single-objective techniques) but instead to select from an array of test suite possibilities the one that best matches the economical and technological constraints of their testing context.

/pdf/comparative-analysis-of-classical-multi-objective-on55ekursj.pdf

Comparative analysis of classical multi-objective evolutionary algorithms and seeding strategies for pairwise testing of Software Product Lines

In this paper, we present a novel implementation of a Cellular Genetic Algorithm (cGA) model for a multi-GPU platform using NVIDIA's CUDA technology. This multi-GPU cGA model is compared first against a serial version in CPU and then versus an implementation on a single GPU. We divide the different operations of the cGA into distinct sets of instructions called kernels. Using the multi-GPU platform we observe that the speedup with respect to the CPU version ranges from 8 to 771, while it is similar to that of the GPU, with a little overhead in the multi-GPU case. Our results demonstrate that multi-GPU desktops can serve as cost-effective parallel computing platforms to obtain accurate results in very short time, although they need special considerations in order to improve on regular single GPUs.

A multi-GPU implementation of a Cellular Genetic Algorithm

An open issue in multi-objective optimization is designing metaheuristics that reach the Pareto front using a low number of function evaluations. In this paper, we adopt a benchmark composed of three well-known problem families (ZDT, DTLZ, and WFG) and analyze the behavior of six state-of-the-art multi-objective metaheuristics, namely, NSGA-II, SPEA2, PAES, OMOPSO, AbYSS, and MOCell, according to their convergence speed, i.e., the number of evaluations required to obtain an accurate Pareto front. By using the hypervolume as a quality indicator, we measure the algorithms converging faster, as well as their hit rate over 100 independent runs. Our study reveals that modern multi-objective metaheuristics such as MOCell, OMOPSO, and AbYSS provide the best overall performance, while NSGA-II and MOCell achieve the best hit rates.

Enrique Alba

Papers

Metaheuristics and Parallelism

DNA fragment assembly using a grid-based genetic algorithm

Comparative analysis of classical multi-objective evolutionary algorithms and seeding strategies for pairwise testing of Software Product Lines

A multi-GPU implementation of a Cellular Genetic Algorithm

A Study of Convergence Speed in Multi-objective Metaheuristics