A heuristic algorithm for fuzzy control

A fuzzy logic-based heuristic algorithm for fuzzy control is presented. The algorithm achieves optimal control with a dynamic programming principle and heuristics used in a branch-and-bound paradigm. The structure of the system, i.e. states, is described in terms of fuzzy propositions and production rules. The states are taken as linguistics expressions of propositions and the production rules define the antecedent-consequent relationship between two states. The system is further represented by a directed acyclic graph to establish the relationship between states. The dynamic behavior of the system is shown by the proposed control algorithm which determines a new state of the system using an approximate reasoning process in a fuzzy environment. A path is established from a given state to the new state with maximum fuzziness and the least space search. The performance of the proposed algorithm is illustrated with an example