A modified pheromone dominant ant colony algorithm for computer virus detection

Abstract

This paper proposes an elementary pattern detection approach for viruses propagated through e-mail and address books using the non-uniform pheromone deposition mechanism of ant colony. The local temporary tabu memory has been used to learn the pattern and it can combine known information from past viruses with a type of prediction for future viruses. This is achieved through certain generated test signature of viruses associated with e-mail over landscape. A non-uniform and non-decreasing time function for pheromone deposition and evaporation ensures that subsequent ants who are close enough to a previously selected trial solution will follow the trajectory or test landscape. They are capable to examine gradually thicker deposition of pheromone over the trajectory. It is empirically shown that the proposed modified pheromone learning mechanism can be an alternative approach to detect virus pattern for e-mail messages.