Creative thinking is arguably the pinnacle of cerebral functionality. Like no other mental faculty, it has been omnipotent in transforming human civilizations. Probing the neural basis of this most extraordinary capac...
详细信息
Creative thinking is arguably the pinnacle of cerebral functionality. Like no other mental faculty, it has been omnipotent in transforming human civilizations. Probing the neural basis of this most extraordinary capacity, however, has been doggedly frustrated. Despite a flurry of activity in cognitive neuroscience, recent reviews have shown that there is no coherent picture emerging from the neuroimaging work. Based on this, we take a different route and apply two well established paradigms to the problem. First is the evolutionary framework that, despite being part and parcel of creativity research, has no informed experimental work in cognitive neuroscience. Second is the emerging prediction framework that recognizes predictive representations as an integrating principle of all cognition. We show here how the prediction imperative revealingly synthesizes a host of new insights into the way brains process variation-selection thought trials and present a new neural mechanism for the partial sightedness in human creativity. Our ability to run offline simulations of expected future environments and action outcomes can account for some of the characteristic properties of cultural evolutionary algorithms running in brains, such as degrees of sightedness, the formation of scaffolds to jump over unviable intermediate forms, or how fitness criteria are set for a selection process that is necessarily hypothetical. Prospective processing in the brain also sheds light on how human creating and designing - as opposed to biological creativity - can be accompanied by intentions and foresight. This paper raises questions about the nature of creative thought that, as far as we know, have never been asked before.
This work presents theoretical results on the development of a statistical convergence criterion for evolutionary algorithms. An analytical formula is derived for the probability of success in isotropic Gaussian mutat...
详细信息
ISBN:
(纸本)9781467384186
This work presents theoretical results on the development of a statistical convergence criterion for evolutionary algorithms. An analytical formula is derived for the probability of success in isotropic Gaussian mutation operators over spherical functions, and statistical criteria are proposed for evaluating, with predefined confidence levels, the convergence of (1 + 1) and (1 + lambda) Evolution Strategies. The results presented are intended as a first approach to the development of statistically based stop criteria for evolutionary optimizers, and as a contribution for the broader application of statistical modeling to the development and study of population-based algorithms.
The cooperation of evolutionary algorithms in a simple hierarchical model was proposed. Three adaptive variants of differential evolution and covariance-matrix-adaptation evolutionary strategy were chosen for cooperat...
详细信息
ISBN:
(纸本)9783319198248;9783319198231
The cooperation of evolutionary algorithms in a simple hierarchical model was proposed. Three adaptive variants of differential evolution and covariance-matrix-adaptation evolutionary strategy were chosen for cooperation. The performance of the model was tested on the learning-based CEC 2015 suite of 15 functions at levels of the dimension 10 and 30. The results showed that the proposed cooperation is beneficial. The proposed model was superior significantly compared to any of individual algorithms included in the cooperation. The cooperation of the selected four algorithms appeared beneficial especially in the problems of dimension D = 30, where the cooperative model outperformed all the algorithms including other cooperative models in comparison.
This paper studies the idea of separating the explored and unexplored regions in the search space to improve change detection and optima tracking. When an optimum is found, a simple sampling technique is used to estim...
详细信息
ISBN:
(纸本)9783319165486;9783319165493
This paper studies the idea of separating the explored and unexplored regions in the search space to improve change detection and optima tracking. When an optimum is found, a simple sampling technique is used to estimate the basin of attraction of that optimum. This estimated basin is marked as an area already explored. Using a special tree-based data structure named KD-Tree to divide the search space, all explored areas can be separated from unexplored areas. Given such a division, the algorithm can focus more on searching for unexplored areas, spending only minimal resource on monitoring explored areas to detect changes in explored regions. The experiments show that the proposed algorithm has competitive performance, especially when change detection is taken into account in the optimisation process. The new algorithm was proved to have less computational complexity in term of identifying the appropriate sub-population/region for each individual. We also carry out investigations to find out why the algorithm performs well. These investigations reveal a positive impact of using the KD-Tree.
Ensemble classifiers have become popular in recent years owing to their ability to produce robust predictive models that generalise well to previously unseen data. In principle, evolutionary algorithms (EAs) are well ...
详细信息
ISBN:
(纸本)9781479974924
Ensemble classifiers have become popular in recent years owing to their ability to produce robust predictive models that generalise well to previously unseen data. In principle, evolutionary algorithms (EAs) are well suited to ensemble generation since they result in a pool of trained classifiers. However, in practice they are infrequently used for this purpose. Current research trends in the EA community focus on relatively complex mechanisms for building ensembles, such as co-evolution and multi-objective optimisation. In this paper, we take a back-to-basics approach, studying whether conventional EAs, augmented with simple niching strategies, can be used to form accurate ensembles. We focus on crowding for this, considering both deterministic and probabilistic variants. We also consider the effect of different similarity measures. Our results suggest that simple niching methods can lead to accurate ensemble classifiers and that the choice of similarity measure is not a significant factor. A further study using heterogeneous classifier models within the population showed no added benefit.
A main focus of current research on evolutionary multiobjective optimization (EMO) is the study of the effectiveness of EMO algorithms for problems with many objectives. Among the several techniques that have led to t...
详细信息
ISBN:
(纸本)9783319159348;9783319159331
A main focus of current research on evolutionary multiobjective optimization (EMO) is the study of the effectiveness of EMO algorithms for problems with many objectives. Among the several techniques that have led to the development of more effective algorithms, decomposition and component-wise design have presented particularly good results. But how do they compare? In this work, we conduct a systematic analysis that compares algorithms produced using the MOEA/D decomposition-based framework and the AutoMOEA component-wise design framework. In particular, we identify a version of MOEA/D that outperforms the best known MOEA/D algorithm for several scenarios and confirms the effectiveness of decomposition on problems with three objectives. However, when we consider problems with five objectives, we show that MOEA/D is unable to outperform SMS-EMOA, being often outperformed by it. Conversely, automatically designed AutoMOEAs display competitive performance on three-objective problems, and the best and most robust performance among all algorithms considered for problems with five objectives.
Multi-Objective evolutionary algorithms(MOEAs) have been gaining increased popularity and usage in different fields of engineering. For real world large scale optimization problems with large variable/search space, us...
详细信息
ISBN:
(纸本)9781479974924
Multi-Objective evolutionary algorithms(MOEAs) have been gaining increased popularity and usage in different fields of engineering. For real world large scale optimization problems with large variable/search space, using a large population of individuals in proportion to the size of search space is ubiquitous. Solving such problems with current state of the art algorithms like NSGA-II [1] is pervasive. The strength of NSGA-II lies in its non-dominance selection procedure and non-dominance based sorting of a population of individuals. Although, the non-dominated sort is computationally efficient for a small population (10(2) - 10(3)) of solutions but becomes computationally expensive and slow for a large population (10(4) - 10(5)) of solutions. Also, various archive based algorithms [2], [3] have been proposed in past which make use of a large population apart from the principal population. Therefore, there is a huge need for a scalable and parallel implementation of NSGA-II. With advent of consumer level Graphics processing units(GPUs) and advancement of CUDA framework we try to fill this research gap using GPGPU architecture. In this paper we propose a parallel GPU based implementation of NSGA-II with major focus on non-dominated sorting. The proposed approach can be easily coupled with the original form of NSGA-II to solve real world problems using large populations.
Echo State Network (ESN) is a special type of neural network with a randomly generated structure called the reservoir. The performance of ESN is sensitive to the reservoir parameters, which have to be tuned for best p...
详细信息
ISBN:
(纸本)9781467378253
Echo State Network (ESN) is a special type of neural network with a randomly generated structure called the reservoir. The performance of ESN is sensitive to the reservoir parameters, which have to be tuned for best performance. Tuning of the reservoir parameters using evolutionary algorithms can be slow and produce inconsistent results. In this paper, we present a simple method for generating reservoirs based on templates that makes the reservoir matrices deterministic with respect to the parameters. Compared with the traditional method where the reservoir matrices are random, tuning of the reservoir parameters with an evolutionary algorithm needs less time, less number of cost function evaluations, and produces more reliable results using the proposed method.
In this paper we argue that to produce good optimization performances, the exploration of the solution space does not need to be carried out in the unorderly fashion most evolutionary algorithms use. Other strategies ...
详细信息
ISBN:
(纸本)9781479974924
In this paper we argue that to produce good optimization performances, the exploration of the solution space does not need to be carried out in the unorderly fashion most evolutionary algorithms use. Other strategies that seek to minimize the cost involved in successive evaluation processes should be explored. This does not imply a fundamental change on how evolutionary algorithms work, but rather, it brings some structure onto how solution spaces are explored by contemplating decoding cost as one of the elements to be minimized when sampling. The traditional implementations of most evolutionary algorithms assume that any point in the solution space can be evaluated any time and at no cost. However, this is not always the case and often each step of the process only part of the solution space is available for evaluation giving rise to a class of problems we have called Constrained Sampling optimization problems over which evolutionary algorithms are quite inefficient. To address these problems we have proposed a modification of the general strategy of evolutionary algorithms to address these constraints efficiently. Here, we study the effects of this approach when applied to problems that are not constrained, thus modifying the way the solution space is explored. This study is carried out to determine how these modification impact the performance of a set of popular evolutionary algorithms over a representative set of benchmark functions corresponding to fitness landscapes with a variety of characteristics. We show that by restricting the sampling capabilities of most algorithms, the cost of the optimization procedure is reduced for most types of fitness landscapes without affecting their results.
In the paper several types of evolutionary algorithms have been tested regarding the dynamic nonlinear multivariable system model. We have defined three problems regarding the observed system: the first is the so-call...
详细信息
In the paper several types of evolutionary algorithms have been tested regarding the dynamic nonlinear multivariable system model. We have defined three problems regarding the observed system: the first is the so-called grey box identification where we search for the characteristic of the system's valve. the second problem is black box identification where we search the model of the system with the usage of system's measurements and the third one is a system's controller design. We solved these problems with the usage of genetic algorithms differential evolution, evolutionary strategies, genetic programming and a de eloped approach called AMEBA algorithm. All methods have been proven to be very useful for solving problems of the grey box identification and design of the controller for the mentioned system but AMEBA algorithm have also been successfully used in black box identification problem where it generated a suitable model. (C) 2015, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.
暂无评论