Open AccessUsing Modified Conjugate Gradient Method to Improve SCA
Author(s) -
Ayad Hamad Khalaf,
Ban Ahmed Mitras
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1591/1/012050
Subject(s) - conjugate gradient method , conjugate , construct (python library) , gradient descent , algorithm , convergence (economics) , mathematics , population , conjugate residual method , mathematical optimization , computer science , artificial intelligence , mathematical analysis , artificial neural network , economics , programming language , economic growth , demography , sociology
This research is to improve Sine - Cosine Algorithm (SCA) that is like any other intelligent techniques that encounter some problem such as slow convergence and the dropping in local solution. To overcome these problems. SCA has been developed and improved through three directions, First: Hybrid of SCA with Modified conjugate gradient method (MCG) that has improved through that derivation of parameter of new conjugate factor (β new ) and attest its characteristic such as descent and global to construct improve algorithm called SCA-MCG. The second direction was a hybrid of SCA with classic optimization methods such as conjugate gradient (CG) algorithm to construct SCA-CG, and parallel Tangent (PT) algorithm to construct SCA-PT. Third combining both previous methods, using the Hybrid value with SCA to construct SCA-CG-PT Algorithm of high quality accounts in all directions mentioned above. To improve the initial population which randomly generated by using excellent characteristics of MCG-CG-PT as well as using this improvement as initial population for SCA. Numerical results have proved the efficiency of improved Algorithm and the results was excellent if we compared with SCA. In addition, we got optimum global values for most functions by achieving functions minimum.