No abstract

No abstract

In today's world of engineering evolution, the need for optimized design has led to development of a plethora of optimization algorithms. Right from hardware engineering design problems that need optimization of design parameters to software applications that require reduction of data sets, optimization algorithms play a vital role. These algorithms are either based on statistical measures or on heuristics. Traditional optimization algorithms use statistical methods in which the optimal solution may not be the global minimal point. These standard optimization techniques are more application specific and demand different parameter sets for different applications. Rather, the bio‐inspired meta‐heuristic algorithms act like black boxes enabling multiple applications with definite global optimal solutions. This review work gives an insight of various bio‐inspired optimization algorithms including dragonfly algorithm, the whale optimization algorithm, gray wolf optimizer, moth‐flame optimization algorithm, cuckoo optimization algorithm, artificial bee colony algorithm, ant colony optimization, grasshopper optimization algorithm, binary bat algorithm, salp algorithm, and the ant lion optimizer. The biological behaviors of the living things that lead to modeling of these algorithms have been discussed in detail. The parametric setting of each algorithm has been studied and their evaluation with benchmark test functions has been reviewed. Also their application to real‐world engineering design problems has been discussed. Based on these characteristics, the possibility to extend these algorithms to data set optimization, feature set reduction, or optimization has been discussed. This article is categorized under: Algorithms and Computational Methods > Algorithms Algorithms and Computational Methods > Computational Complexity Algorithms and Computational Methods > Genetic Algorithms and Evolutionary Computing

Spatial autore g ressive (AR) models have been extensively used to represent texture ima g es in machine learnin g applications. This work emphasizes the contribution of 2D autore g ressive models for analysis and synthesis of textural ima g es. Autore g ressive model parameters as a feature set of texture ima g e represent texture and used for synthesis. Yule walker Least Square (LS) method has used for parameter estimation.The test statistics for choice of proper nei g hbourhood (N) has also been su gg ested. The Brodatz texture ima g e album has chosen for the experimentation.Parameters have estimated from the textures. The test statistics decides the best nei g hbourhood or proper order of the model.The synthesized texture ima g e and the ori g inal texture ima g e have compared for perceptual similarities. It is been inferred that the proper nei g hbourhood for a g iven texture is unique and solely depends on the properties of the texture.

No abstract