Ant Colony Heuristics in File Compression: An Exploratory Study
Vijesh Mehta · 2003
In compressing files, conventional algorithms use mathematical patterns to reduce file size. It is our thesis to test and see if conventional file compression can be improved by altering the bits in a file. Using Ant Colony algorithms (Dr. Gargano – Ant Colony Heuristics), we were able to create a testing environment for this thesis. Software was developed which puts individual instances of ants in a file, where each ant moves in a predictable pattern while changing the bits in its path. Once the algorithm is finished executing, the altered file is written to disk in order to see if further compression can be achieved using a normal ZIP compression. The results showed minimal changes in file compression. It is probabilistically equally likely to change a bit from 1 or 0. So the starting number of zeros and ones in a file remain somewhat equal to the ending number. In future tests it would probably be prudent to make more intelligent ant motions which move based on understanding zip’s pattern search mechanism.