Figure - uploaded by Adolph J. Vogel
Content may be subject to copyright.
Source publication
Parameter tuning aims to find suitable parameter values for heuristic optimisation algorithms that allows for the practical application of such algorithms. Conventional tuning approaches view the tuning problem as two distinct problems, namely, a stochastic problem to quantify the performance of a parameter vector and a deterministic problem for fi...
Context in source publication
Context 1
... are therefore able to propose appropriate values for each of the three parameters. This is done by performing a coarse 3 3 full factorial experiment, i.e. three values per parameter for the three ESPO parameters as tabulated in Table 3. It is evident that the parameters values in the full factorial experiment differ significantly allowing for a proper sensitivity versus robustness study to be performed. ...Citations
... To save computation resources and increase tuning quality, we plan to investigate more advanced techniques e.g. spatially distributed statistical significance approach [64], chess rating system-tuning (CRS-tuning) [65], etc. The proposed inter-subsystem LS methods appear to be effective while there are still rooms for improvements. ...
Redundancy allocation is widely used for optimizing system reliability in various fields. Non-homogeneous redundancy allocation in multi-state series-parallel systems has captured much attention in recent years. Due to the large and complex solution space, it is among the most difficult types of redundancy allocation problems. Due to the multimodal landscape of solution neighborhood, ordinary evolutionary algorithms (EAs) could not sufficiently exploit the solution neighborhood and thus usually trap into local optimums. Therefore, local search (LS) becomes a critical attachment to the main EA to achieve better performance. This work proposes two originally designed inter-subsystem LS methods for non-homogeneous redundancy allocation in multi-state series-parallel systems. They provide a well-balanced mechanism to reduce system cost via inter-subsystem operation and ensure system availability. The inter-subsystem local search methods are combined with quantum-inspired evolutionary algorithm (QEA), establishing an efficient memetic algorithm. The proposed memetic algorithm is validated on five benchmark problems with different availability requirements. The statistical tests results indicate that the inter-subsystem LS strategies are significantly better than the LS strategies published in QEA approach proposed by one of the authors of this paper, at small computational costs. The comparisons to various published results show that the proposed memetic algorithm achieves the best published solutions using much fewer fitness evaluations and higher robustness.
Computational intelligence methods have gained importance in several real-world domains such as process optimization, system identification, data mining, or statistical quality control. Tools are missing, which determine the performance of computational intelligence methods in these application domains in an objective manner. Statistics provide methods for comparing algorithms on certain data sets. In the past, several test suites were presented and considered as state of the art. However, there are several drawbacks of these test suites, namely: (i) problem instances are somehow artificial and have no direct link to real-world settings; (ii) since there is a fixed number of test instances, algorithms can be fitted or tuned to this specific and very limited set of test functions; (iii) statistical tools for comparisons of several algorithms on several test problem instances are relatively complex and not easily to analyze. We propose a methodology to overcome these difficulties. It is based on standard ideas from statistics: analysis of variance and its extension to mixed models. This paper combines essential ideas from two approaches: problem generation and statistical analysis of computer experiments.
