OPTIMIZATION BY NEURAL NETWORK ALGORITHM AND ITS APPLICATIONS IN SYSTEM OF NONLINEAR EQUATIONS

• Main Author: Martua Sinambela, Reynold
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/60699
• Institution: Institut Teknologi Bandung

Inspired by Artificial Neural Network (ANN), a metaheuristics optimization algorithm which then so-called Neural Network Algorithm (NNA) is utilized to solve optimization problems. In addition to the global optimization problem, NNA is also can be modified to find all roots of nonlinear system although only one root can be obtained in one program execution. In this thesis, to find all possible roots of nonlinear system in one program execution, it is developed the combination between NNA and clustering technique. By combining both, all roots of nonlinear system can be obtained just in one program execution. Furthermore, this NNA with clustering technique is also can be utilized to map the extreme points of multimodal function by single program execution. Finally, after doing test to all benchmark functions, NNA with clustering technique has been tested successfully such as global optimization problems, finding all root of nonlinear system, and multimodal optimization problems Keywords: optimization, artificial neural network, clustering,