|
||||||
|
||||||
| Undergraduate Honours Projects | ||||||
|
Carleton University - School of Computer Science Undergraduate Honours Project Summer 2011 Solving Classical Problems in Computational Molecular Biology Using Distributed Computing and Web Technologies Rakhim Davletkaliyev 
ABSTRACT One of the most promising ways to solve the hardest computational problems nowadays is distributed and parallel computing. The motivation is easy: if problem is too big and hard, divide it into pieces and let several entities solve their piece; then, consolidate the results into the potential answer. Most of personal computers and computing devices today are connected to internet and have web-browser installed, which usually allows to run code on host machine. Yet, this feature is not widely used as a tool to divide and distribute computation. The goal of this project is to create a JavaScript framework that would allow anyone to set a server and a webpage, connecting to which a userâs web-browser will contribute to overall computation. For example, the famous traveling salesman problem can has plenty of potential solutions. With help of degas.js (distributed evolutionary (genetic) algorithm spanned framework) one can define this problem on the server (as array of coordinate points which salesman must visit), define the way to evaluate a solution (as function summing the distance travelled) on the client side. Now, every user that visits the webpage contributes their CPU cycles to generate and send potential solutions to the server. |
||||||
|
