Comparing HOSS with flat method.

<p>Optimizations were carried out using 200 start points for each method, launched on 16 processes on a 128-core server. Solution times were computed individually for each optimization on each process. A-D: Distribution of solution times for HOSS (blue) and flat (orange) methods, for each of t...

Full description

Saved in:
Bibliographic Details
Main Author: Nisha Ann Viswan (20371285) (author)
Other Authors: Alexandre Tribut (20371288) (author), Manvel Gasparyan (14250300) (author), Ovidiu Radulescu (266188) (author), Upinder S. Bhalla (11768249) (author)
Published: 2024
Subjects:
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:<p>Optimizations were carried out using 200 start points for each method, launched on 16 processes on a 128-core server. Solution times were computed individually for each optimization on each process. A-D: Distribution of solution times for HOSS (blue) and flat (orange) methods, for each of the four models. HOSS converges faster. E-H: Distribution of initial cost (blue, green) and final cost (red = HOSS; orange = flat) for the two methods. HOSS produces better costs in most models but in one case (D3 EGFR) the distribution overlaps with the flat method.</p>