Yazar "Toktas, Abdurrahim" seçeneğine göre listele

Listeleniyor 1 - 2 / 2
  • [ X ]
    Öğe
    Adaptive Flower Pollination Algorithm Based on Spatial Dispersal
    (Institute of Electrical and Electronics Engineers Inc., 2020) Ustun, Deniz; Toktas, Abdurrahim
    An adaptive flower pollination algorithm (AFPA) is evolved with respect to the accuracy and stability performance. AFPA is developed by modifying both global and local search operators. The exploration and exploitation abilities of the algorithm are thus improved by integrating a spatial dispersal mechanism adapting with the number of iterations to the global search and the best solution to the local search. Its performance is well corroborated through several benchmark functions with respect to various measuring parameters. AFPA is exhibited the distinguished accuracy and stability performance better than FPA for most benchmark functions. The obtained results show that AFPA has the better global and local search abilities over FPA. © 2020 IEEE.
  • [ X ]
    Öğe
    Pioneer Pareto artificial bee colony algorithm for three-dimensional objective space optimization of composite-based layered radar absorber
    (Elsevier, 2020) Toktas, Abdurrahim; Ustun, Deniz; Erdogan, Nursev
    A three-dimensional objective space (3DOS) optimization strategy using an enhanced multi-objective artificial bee colony (ABC) algorithm for the design optimization of layered radar absorbing material (LRAM) is presented in this study. The multi-objective exploitation ability of ABC is improved with regard to the convergence and diversity by integrating a pioneer Pareto (PP) solution to the onlooker bee phase, which is selected from the Pareto optimal set. Initially, the performance of PP-ABC is successfully verified by a comparison with ABC and the well-known multi-objective counterparts like particle swarm optimization (PSO) and differential evolution (DE) algorithms. The comparison is carried out through five multi-objective benchmark functions with respect to three favorable and reliable multi-objective indicators such as hypervolume (HV), HV ratio and Pareto sets proximity (PSP). The employed three objective functions to be the dimensions of 3DOS are weighted bandwidth-based total reflection coefficient involving sub-reflection waves of a wide oblique incident angular range 0 degrees-75 degrees, the total thickness and the number of layers. By using PP-ABC, a 3D designed LRAM operating at a large frequency band of 2-18 GHz is then designed for synchronously minimizing the three objective vectors by finding out the design variables: thickness and material types. Meanwhile, the material types of the proposed LRAM are optimally picked up from a composite material database with 51 specimens from 9 previously reported studies (51 /9#database). In order to point out the effectiveness of the proposed 3DOS optimization strategy, three LRAMs are also compared with respective reported designs whose material type is selected from a database with 6 specimens (6/1#database). The results show that the proposed LRAMs are hence the global optimal designs in terms of all objective functions thanks to the proposed 3DOS optimization strategy based on PP-ABC. (C) 2020 Elsevier B.V. All rights reserved.