DEFINING VARIATION OPERATOR FOR GRAMMAR REACHABILITY SEARCH BASED VULNERABILITIES DETECTION

Abstract

In population-based search algorithm such as Evolutionary Programming (EP), the search process typically involves seeding population of first generation with randomly generated individuals, selecting parents through fitness evaluation, producing offsprings through variation of parents, and selecting parents and offsprings into next generation of candidate solutions. Obviously, the quality of the variation operator is important in leading the search process towards global optimal solution.  In this paper, a high-quality variation operator is proposed. The proposed variation operator has the capacity to bias search towards optimal solutions while ensuring adequate balance between exploration and exploitation of the search space so as to facilitate discovery of optimal solutions in fewer number of generations. The proposed variation operator was used in our published work named EPSQLiFix. The proposed variation operator demonstrated high performance. Thus, it can as well be applicable in other related problem domains.