Khirul Fozi, Naznin Faiqa and Mohd Rodi, Hanis Sofia (2019) Comparisons between Newton and QuasiNewton method in solving unconstrained optimization problems / Naznin Faiqa Khirul Fozi & Hanis Sofia Mohd Rodi. Degree thesis, Universiti Teknologi MARA.
Abstract
Newton and QuasiNewton methods are widely used in solving unconstrained optimization problems. The solution to optimization problems are known as local optimum solutions and global minimum solutions. For Newton method, if the initial points are far from the solution points, it may fail to converge. As an alternative, two QuasiNewton methods which are DavidonFletcherPowell (DFP) and Broyden FletcherGoldfarbShanno (BFGS) methods were developed to overcome this problem. In this research, a comparison was made between Newton and QuasiNewton method to determine the best method in solving unconstrained optimization problems. These methods were tested using six test functions with different initial points and their performance were compared based on the number of iterations, CPU time, and accuracy. This research also discussed about the convergence rate, global convergence and local convergence of the three methods. From numerical results, it has been shown that BFGS method is better compared to the other methods.
Metadata
Item Type:  Thesis (Degree)  

Creators: 


Contributors: 


Subjects:  Q Science > QA Mathematics > Equations Q Science > QA Mathematics > Mathematical statistics. Probabilities Q Science > QA Mathematics > Analysis > Analytical methods used in the solution of physical problems Q Science > QA Mathematics > Instruments and machines > Electronic Computers. Computer Science > Algorithms 

Divisions:  Universiti Teknologi MARA, Terengganu > Kuala Terengganu Campus > Faculty of Computer and Mathematical Sciences  
Programme:  Bachelor of Science (Hons) Computational Mathematics  
Item ID:  39809  
Uncontrolled Keywords:  Newton Method ; DavidonFletcherPowell ; Broyden FletcherGoldfarbShanno  
URI:  http://ir.uitm.edu.my/id/eprint/39809 
Download
Text
39809.pdf Download (190kB) 
Fulltext
ID Number
39809
Others
View in Google Scholar
Actions (login required)
View Item 