Editorial for the special issue of Information Sciences Journal (ISJ) on "Nature-inspired algorithms for large scale global optimization"
Editorial Editorial for the special issue of Information Sciences Journal (ISJ) on "Nature-inspired algorithms for large scale global optimization" In the past few decades, many nature-inspired optimization algorithms have been developed successfully for solving a wide range of optimization problems. Evolutionary Algorithms (EAs), Simulated Annealing, Differential Evolution (DE), Particle Swarm Optimization (PSO), Ant Colony Optimization (ACO), Estimation of Distribution Algorithms (EDA) are
... t some representative examples among many others. These meta-heuristic algorithms do not rely on gradient information, and are less likely to be stuck on local optima because of their use of a population of candidate solutions, thereby offering significant advantages over traditional single-point and derivative-dependent methods. Although these techniques have demonstrated excellent search capabilities for solving small or medium-sized optimization problems, they still encounter serious challenges when applied to solve large scale optimization problems, i.e., problems with several hundreds to thousands, or millions of variables. Only very few attempts have been made by nature-inspired optimization algorithms on problems with millions of variables [1,2]. Many real-world optimization problems involve a large number of decision variables. For example, in shape optimization, a large number of shape design variables are often used to represent complex shapes, such as turbine blades, aircraft wings, and heat exchangers, etc. With the arrival of big data, there is an unprecedented demand to solve optimization problems with a large number of feature variables, training instances, and classes  . The recent advance in the area of machine learning has witnessed very large scale optimization problems encountered in training deep neural network architectures (so-called deep learning), some of which are involved in optimizing over a billion of connection weights in a very large neural network  . This is something unheard of even 10 years ago. How well nature-inspired optimization algorithms handle this kind of real-world large scale global optimization (LSGO) problems still remains an open question. In recent years, researches on scaling up nature-inspired optimization algorithms such as EAs to tackle large scale optimization problems have gathered momentum on both theoretical and empirical studies. LSGO has attracted an increasing attention in the optimization research community. Most noticeable are the LSGO special sessions, competitions, and tutorials organized at the IEEE flagship conference Congress on Evolutionary Computation (CEC) since 2008. Several LSGO benchmark test function suites specifically designed to cater for the LSGO competitions have been highly cited [5, 6] . This special issue aims to highlight the latest development in the area of nature-inspired algorithms for handling LSGO problems. We have received over 50 submissions to this special issue. All of the submissions were rigorously peer-reviewed by LSGO experts. Only 9 papers were finally recommended for publication. The first paper "Non-rigid Multi-modal Medical Image Registration by Combining L-BFGS-B with Cat Swarm Optimization" by Yang et al., proposes a cooperative coevolutionary model hybridizing the classic L-BFGS-B method with the cat swarm optimization method, for tackling the non-rigid multi-modal image registration problem. The L-BFGS-B method is used here as an efficient local search method. An important contribution of this work is the adoption of a block grouping method which is much more effective in capturing the interdependency among variables. The second paper "Decomposition-based Evolutionary Algorithm for Large Scale Constrained Problems" by Sayed et al., proposes VIIC (Variable Interaction Identification Technique for Constrained problems), which aims specifically for solving constrained LSGO problems. The paper also proposes a set of constrained LSGO benchmark functions, and develops DEVIIC (decomposition-based EA optimization model and VIIC). A challenging industrial LSGO problem with constraints is used for evaluating the proposed DEVIIC method. The third paper "Solving the large-scale hybrid flow shop scheduling problem with limited buffers by a hybrid artificial bee colony algorithm" by Li and Pan, combines an artificial bee colony (ABC) with tabu search (TS) for solving the large scale hybrid flow shop (HFS) scheduling problem with limited buffers. A self-adaptive neighbourhood strategy is adopted to balance the exploitation and exploration capabilities of the algorithm. The fourth paper "Greedy discrete particle swarm optimization for large-scale social network clustering" by Cai et al., proposes a new discrete PSO method for discovering community structures in social networks. More specifically in http://dx.