Scholarly record
AN ADAPTIVE LIBRARY OF HYBRID APPROXIMATION ALGORITHMS BASED ON MULTI-EXTREMUM FUNCTIONS
Abstract
Approximation technologies have regained much interest in recent years in the context of wide variety of computation techniques and large scale optimization. The purpose of this study is to outline the conceptual framework and architectural design of adaptive library for approximate solutions based on multi-extremum functions. Simulation technology of complex systems includes approximation methods as estimation tools. Despite the fast development of computing power and technology, many problems of multi-extremum optimization are known to be computationally difficult. In this study we describe the adaptation library of algorithms and programs provides the realization of surrogate based modeling technique using hybrid approximation method, in which the goal is to implement sets of special functions differing in possibility of selecting special types of functions in conjunction with fast greedy algorithms or genetic optimization parameters. The adaptation library includes hybrid approximation algorithms, special models, data structures and several complete solutions for simulation various classes of approximation problems. Adaptation to these problems is based on three selection methods of calculation process: procedural, parametric and combination method. More than two hundreds approximate models and algorithms are integrated in this special models library. A problem of the environmental factors impacts on complex systems is used to demonstrate the implementation of adaptive library. We further show how several fast-rate results illustrate the usage of hybrid approximate algorithms. Furthermore, comparative analysis of hybrid algorithms has presented robustness of proposed adaptive library.
Publication Impact Profile
Publication details
References13
Bishop C.M., Pattern Recognition and Machine Learning, Springer, 2007.
Hegde C., Indyk P., Schmidt L., Approximation Algorithms for Model-Based Compressive Sensing, IEEE Transactions on Information Theory, 2015.
Ausiello G., Protasi M., Local search, reducibility and approximability of NP-optimization problems, Information Processing Letters, vol. 64, pp. 73 79, 1995.
Forrester A., Sobester A., Keane A., Engineering Design Via Surrogate Modelling. A Practical Guide, Willey, 2008.
S. Laue., A Hybrid Algorithm for Convex Semidefinite Optimization, Proceedings of the 29th International Conference on Machine Learning, Edinburg, Scotland, UK, pp. 177-184, 2012.
Alexeev D. V., Neural-Network Approximation of Functions of Several Variables, Journal of Mathematical Sciences, vol. 168, No. 1, pp. 5-13, 2010.
Hecht-Nielsen R., Kolmogorov's Mapping Neural Network Existence Theorem, IEEE First Annual International Conference on Neural Networks, San Diego, vol. 3, pp. 11-13, 1987.
Deep K., Das K.N., Quadratic approximation based hybrid genetic algorithm for function optimization, Applied Mathematics and Computation, vol. 203, No 1, pp 86-98, 2008.
Baqais. A., Genetic Algorithm for Function Approximation: an Experimental Investigation, International Journal of Artificial Intelligence and Applications (IJAIA), vol. 7, No. 3, 2016.
Stolbunov V., Nair P.B., Sparse radial Basis Function Approximation with Spatially Variable Shape Parameters, Applied Mathematics and Computation, vol.330, pp. 170-184, 2018.
Kolmogorov A.N., On the Representation of Continuous Functions of Many Variables by Superposition of Continuous Functions of One Variable and Addition, American Mathematics Society, vol. 28, pp. 55-63, 1963.
Petrunina E., Nicolskiy A., Beloglazov A., Pecherskij D., a Successive Multi-Extremum Approximation Algorithms Based on Special Functions in the Enviromental Impacts Researches on Technogenic Systems, 18th International Multidisciplinary Scientific Geoconference SGEM 2018, Sofia, Bulgaria, vol.18, pp. 27-35, 2018.
Erohin A.T., Approximation technique for random processes based on special functions, Geophysics methods for environment monitoring, Russia, pp. 115-125, 1991.
View or Download full articleAccess options
SWS access login
Login as SWS Scientific CommitteeLogin as SWS Scientific PartnerLogin as SWS AuthorAuthors and approved SWS contributors will read and export their own linked papers after identity matching by SWS profile, email and SGEM GlobalID.
For librarian assistance: [email protected]
Purchase Instant Access
- Article can be downloaded after successful payment.
- Article may be used according to SWS library access terms.
- Article cannot be redistributed.
