AN INNOVATIVE APPROACH OF ENGINEERING OPTIMIZATION

Abstract

Optimization methods have various applications in the field of science and engineering. The construction, design and maintenance of engineering systems involve decision making both at the technological and the managerial level. In today’s era, engineers strive to achieve optimal solutions with high accuracy and minimal computational time, leveraging the power of emerging artificial intelligence (AI) technologies. The effectiveness of traditional optimization methods can be enhanced by machine learning model trainers. Among the various optimization methods, the quasi-Newton methods are developed and proved to overcome the limitations of other classical methods. Multi-step quasi-Newton methods are the best version of QN method which approximate the current Hessian by using the data from more than one previous step. Several nonlinear problems were solved by using multi-step quasi-Newton method with skipping and modified search direction technique on variable space. In this study, 3-step quasi-Newton methods with different techniques of skipping and modified search direction on gradient space is experimented. On the basis of numerical results, the behavior of proposed strategy is compiled and found the superiority of fixed-point approach.