•  
  •  
 

Journal of Undergraduate Research

Keywords

speculative approach, parallelization, particle swarm optimization

College

Physical and Mathematical Sciences

Department

Computer Science

Abstract

Optimization problems are ubiquitous in our modern world. Businesses such as Google need to decide where to place ads, scientists need to t models to data, and airlines need to schedule their flights. All of these problems are optimization problems, and optimization techniques can be used to find solutions. Particle swarm optimization (PSO) is a relatively new optimization technique that draws on ideas from the sociology of flocking birds. It was originally developed for use on a single machine, and attempts at modifying the algorithm to run in parallel are still quite preliminary. However, the problems people want to solve are getting bigger, and larger problems require more computation power than is available on any single computer. To effectively solve those problems, optimization algorithms need to be improved to better use parallel clusters of machines. This project addressed that issue.

Download

Share

COinS