13.05.2018 Views

merged

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

The Comparison between Different Versions of<br />

Conjugate Gradient in Minimization Problems<br />

Yan Lai Lee<br />

Supervisor: Dr. Nur Fadhilah Binti Ibrahim<br />

Bachelor of Science (Financial Mathematics)<br />

School of Informatics and Applied Mathematics<br />

Optimization is widely used in many fields such as mathematics, economics and<br />

engineering. One of the useful mathematical technique is the conjugate gradient method.<br />

However, different conjugate gradient methods correspond to different versions of the<br />

parameter β k . The different versions of the parameter have different rate of convergence.<br />

Besides that, the Polak-Ribiere-Polyak (PRP) method is generally regarded to be one of<br />

the most efficient conjugate gradient methods. This research aims to compare the three<br />

different modified PRP method which are β k BZAU , β k SRMI , β k<br />

MRM<br />

with exact line search in<br />

minimization problem, to determine which version is the most efficient conjugate gradient<br />

version as well as to identify the similarities and differences between the versions of the<br />

parameter β k . In this research, the algorithm of conjugate gradient is used to solve<br />

optimization problems and compared by changing the parameter β k which are β k BZAU ,<br />

β k SRMI , β k MRM . The number of iterations and Central Processing Unit (CPU) time per second<br />

for each of the β k will be computed by the MATLAB. The results showed that β k<br />

BZAU<br />

is<br />

the most efficient version of modified PRP Moreover, the similarities and difference<br />

between the versions of the parameter β k were identified after comparing both methods.<br />

906 | UMT UNDERGRADUATE RESEARCH DAY 2018

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!