Home // International Journal On Advances in Intelligent Systems, volume 11, numbers 3 and 4, 2018 // View article
Authors:
Rudolf Neydorf
Ivan Chernogorov
Dean Vucinic
Keywords: search optimization; multi-extreme; method of swarming particles; test functions; irregularity; software.
Abstract:
The paper studies the actual task of developing and setting up the algorithms for multi-extreme objects search optimization. To solve such problems, the heuristic methods are effectively used, in particular applying the swarming particles method. The mathematical base for the modified swarming particles method, which is oriented to solve the multi-extreme search problems, is developed and described in detail. The modified algorithm is applied to the irregular multi-extreme "Epsilon" test functions, considered to be a very difficult test case. The functionality of developed software "EpsilonFunction", which is created to control the generation, editing and investigation of multidimensional multi-extreme test functions is described. Epsilon function is a multiplicative function developed by R. Neydorf, which has fundamental extremes, multidimensionality and isolation in the factor space, which makes possible to completely exclude the influence of the results. The use of these fundamentally new test functions made possible to show that such modified method of swarming particles is suitable for solving the rather complex multi-extreme search problems. The developed software tool "EpsilonFunction" has a wide range of possibilities for developing and researching the test functions that are being created for other related applications. Epsilon-functions have found application in the method of "Cut-Glue" approximation of experimental data, which is at the researhing stage, but has already proved itself as an innovative and effective tool of data approximation.
Pages: 146 to 155
Copyright: Copyright (c) to authors, 2018. Used with permission.
Publication date: December 30, 2018
Published in: journal
ISSN: 1942-2679