Search the digital library catalog

Abstract: Reporting the empirical results of swarm and evolutionary computation algorithms is a challenging task with many possible difficulties. These difficulties stem from the stochastic nature of such algorithms, as well as their inability to guarantee an optimal solution in polynomial time. This research deals with measuring the performance of stochastic optimization algorithms, as well as the confidence intervals of the empirically obtained statistics. Traditionally, the arithmetic mean is used for measuring average performance, but we propose quantiles for measuring average, peak and bad-case performance, and give their interpretations in a relevant context for measuring the performance of the metaheuristics. In order to investigate the differences between arithmetic mean and quantiles, and to confirm possible benefits, we conducted experiments with 7 stochastic algorithms and 20 unconstrained continuous variable optimization problems. The experiments showed that median was a better measure of average performance than arithmetic mean, based on the observed solution quality. Out of 20 problem instances, a discrepancy between the arithmetic mean and median happened in 6 instances, out of which 5 were resolved in favor of median and 1 instance remained unresolved as a near tie. The arithmetic mean was completely inadequate for measuring average performance based on the observed number of function evaluations, while the 0.5 quantile (median) was suitable for that task. The quantiles also showed to be adequate for assessing peak performance and bad-case performance. In this paper, we also proposed a bootstrap method to calculate the confidence intervals of the probability of the empirically obtained quantiles. Considering the many advantages of using quantiles, including the ability to calculate probabilities of success in the case of multiple executions of the algorithm and the practically useful method of calculating confidence intervals, we recommend quantiles as the standard measure of peak, average and bad-case performance of stochastic optimization algorithms.
Keywords: algorithmic performance, experimental evaluation, metaheuristics, quantile, confidence interval, stochastic algorithms, evolutionary computation, swarm intelligence, experimental methodology
Published in DKUM: 28.03.2025; Views: 0; Downloads: 8
Full text (490,48 KB)
This document has many files! More...

Abstract: The design of neural network architecture is becoming more difficult as the complexity of the problems we tackle using machine learning increases. Many variables influence the performance of a neural model, and those variables are often limited by the researcher's prior knowledge and experience. In our master's thesis, we will focus on becoming familiar with evolutionary neural network design, anomaly detection techniques, and a deeper knowledge of autoencoders and their potential for application in unsupervised learning. Our practical objective will be to build a neural architecture search based on swarm intelligence, and construct an autoencoder architecture for anomaly detection in the MNIST dataset.
Keywords: neural architecture search, machine learning, swarm intelligence
Published in DKUM: 18.10.2021; Views: 1269; Downloads: 113
Full text (3,18 MB)
This document has many files! More...

Abstract: Swarm intelligence is a modern and efficient mechanism for solving hard problems in computer science, engineering, mathematics, economics, medicine and optimization. Swarm intelligence is the collective behavior of decentralized and self-organized systems. This research area is a branch of artificial intelligence and could be viewed as some kind of family relationship with evolutionary computation because both communities share a lot of common characteristics. To date, a lot of swarm intelligence algorithms have been developed and applied to several real-world problems. The main focus of this thesis is devoted to the bat algorithm which is a member of the swarm intelligence community, as developed recently. In line with this, a comprehensive analysis of papers was performed tackling this algorithm. Some hybridizations of the original algorithm were proposed because the preliminary results of this algorithm regarding the optimization of benchmark functions with higher dimensions had not too promising. Extensive experiments showed that the hybridizing the original bat algorithm has beneficial effects on the results of the original bat algorithm. Finally, an experimental study was performed during which we researched for the dependence of an applied randomized method on the results of the original bat algorithm. The results of this study showed that selecting the randomized method had a crucial impact on the results of the original bat algorithm and that the bat algorithm using Levy flights is also suitable for solving the harder optimization problems.
Keywords: swarm intelligence, evolutionary computation, bat algorithm, hybridization, review
Published in DKUM: 06.09.2013; Views: 3471; Downloads: 335
Full text (8,69 MB)