Skip to content

Bachelor's graduate work on the topic: "Three-component conjugate gradient methods for solving high-dimensional optimization problems". A new optimization method was developed, which demonstrated high performance.

Notifications You must be signed in to change notification settings

egubova/graduation-work

Repository files navigation

Graduation work

date: 06/2021

Bachelor's graduate work on the topic: "Three-component conjugate gradient methods for solving high-dimensional optimization problems". A new optimization method was developed, which demonstrated high performance.

The methods files contain the implementation of my own method and some basic algorithms that are compared with. In the cute_methods file, the methods are implemented for conducting numerical experiments using the pycutest library (it works only on UNIX operating systems). And the quad_methods file contains an implementation of the same methods, but for testing on quadratic optimization problems (Ridge regression).

In the main files, the methods themselves are tested directly, as well as the construction of graphs.

The cute and quad parts of the program work separately from each other.

In the file problem_list.txt contains a list of optimization problems from the CUTEst library, on which numerical experiments were conducted.

The Work_report.pdf file contains a full report on the work performed (93 pages).

About

Bachelor's graduate work on the topic: "Three-component conjugate gradient methods for solving high-dimensional optimization problems". A new optimization method was developed, which demonstrated high performance.

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages