MMAS Algorithm using Fuzzy Rules

The real life problems deal with imperfectly specified knowledge and some degree of imprecision, uncertainty or inconsistency is embedded in the problem specification. The well-founded theory of fuzzy sets is a special way to model the uncertainty. The rules in a fuzzy model contain a set of propositions, each of which restricts a fuzzy variable to a single fuzzy value by means of the predicate equivalency. That way, each rule covers a single fuzzy region of the fuzzy grid. The proposed system of this thesis extends this structure to provide more general fuzzy rules, covering the input space as much as possible. In order to do this, new predicates are considered and a Max-Min Ant System is proposed to learn such fuzzy rules. Ant system is a general purpose algorithm inspired by the study of behavior of ant colonies. It is based on cooperative search paradigm that is applicable to the solution of combinatorial optimization problem. In this thesis we consider the combinatorial optimization issue of travelling salesman problem (TSP) which evaluates more generic Fuzzy rules provided by Max-Min Ant System (MMAS). The existing ant colony system (ACS) was a distributed algorithm applied to the travelling salesman problem (TSP). In ACS, a set of cooperating agents called ants cooperate to find good solutions for TSPs (but here, Ants search their path randomly). Ants cooperate using an indirect form of communication mediated by pheromone they deposit on the edges of the TSP problem in symmetric instances. However most of the TSP issues carry both symmetric and asymmetric instances.