Computational Study of Kernel - Based Interior - Point Method for LCP

<p>One of mathematical problems, that have many practical applications, is the well-known linear complementary problem (LCP) which consists of finding a certain vector that satisfy a set of linear inequalities and complementary equation. In this thesis, after introducing and analyzing a kernel-based primal-dual interior-point method (IPM) for solving LCP, we consider several, fairly general, eligible kernel functions. We show that the algorithm, with some of those kernel functions, has comparable complexity with the best complexity results obtained in the literature for these type of methods. Three basic implementations of the algorithm in MATLAB were used to conduct a series of numerical tests for different kernel functions, showing promising performance.</p>

Name thesis:degree_name

Master of Science in Mathematics (M.S.)

Level thesis:degree_level

Thesis (restricted to Georgia Southern)

Discipline thesis:degree_discipline

Department of Mathematical Sciences

Year dc:date.available

2015

Author dc:creator

• Katic, Robert

Contributors dc:contributor

• Hua Wang
• Yan Wu

• ETD
• Kernel function
• Primal-dual
• Interior-point method
• Linear complementarity problem
• Numerical Analysis and Computation
• Other Applied Mathematics
• Other Mathematics

Repository record dc:identifier

https://digitalcommons.georgiasouthern.edu/etd/1304

OAI identifier oai:identifier

oai:digitalcommons.georgiasouthern.edu:etd-2371