Genetic algorithm (GA) is an optimization technique which is applicable to all the functions that can be evaluated by using fuzzy rule based system. The problems can also be optimized by using mathematical functions such as calculus (derivative, integration etc.), and other nonlinear modeling tools such as neural networks. But the main advantage of fuzzy rule based systems over other methods is their high transparency. The fuzzy rule based system consists of fuzzy if-then rules such as "if X1 is large and X2 is medium, then Z is large". The main problem with the existing fuzzy if-then rules is that as the complexity of the problem increases, the number of rules to define the problem also increases. But this increase in the number of rules is exponential, and not linear. As a result, the memory requirement and the search time also increases. This is the major problem with the earlier systems.
Keywords: Genetic Algorithm, Expert System, Fuzzy Expert System, Rule Based System, Efficient Rule
Based System
[...] Vol No November 1998. C.Z. Jainikow. A genetic algorithm for optimizing fuzzy decision trees. In Proc. Of 6th International Conf. on Genetic Algorithms, pp. 421-428, July 15- Cordon, O., F. Herrera, et al. Recent advances in genetic fuzzy systems, Information sciences, 136(2001). A. Abraham, Evonf: A framework for optimization of fuzzy inference systems using neural network learning and evolutionary computation, in: The 17th IEEE International Symposium on Intelligent Control,ISIC'02, IEEE Press, ISBN 0780376218, pp.327-332, Canada [10]. [11]. References L. B. [...]
[...] Efficient Rule Base Proposed Genetic Algorithm After creating the rule based system, the next step is to optimize the number of rules. The proposed approach is shown in the fig 2 below in the form of the flowchart: Rule Evaluation Rules / Inferences Fuzzy Output Defuzzification Output Output Membership Function Crisp Output Fig Flowchart for Fuzzy Expert System Generate an initial population, call these as chromosomes Attach fitness value to each of these chromosomes the given initial population. Then the probability is assigned to each of these chromosomes by using the following formula: Select initial population according to fitness value Select ith chromosome for mating with probability Pi = Fitness Σ Fitness i = 1 to n Apply crossover operator to chromosomes Xi and Xj Pi = Fitness / Σ Fitness i = 1 to n where Xi is the ith chromosome in the given population for I = n Crossover: Here, two or more chromosomes are selected from the solution space depending on their probability. [...]
[...] Here, we are trying to apply the three genetic operators on the antecedent part of the rule set. Effectively dealing with some knowledge verification issues is the future research work Conclusion An optimization method for fuzzy rule base is proposed, that can be incorporated in fuzzy expert systems. Both representative data and expert knowledge are included in designing fuzzy rules, which include expert knowledge. This approach does not require any human intervention during the optimization phase. The time required is thus dependent on the computer execution speed, but not on human experts, much time can thus be saved, since experts may be geographically dispersed, and their discussions are usually time consuming. [...]
[...] Yes Minimized Rule Base Stop Fig Proposed Efficient Rule Based Genetic System Initial Population The genetic algorithm requires an initial population for its operation, which is nothing but is the collection of feasible solutions. The initial population is selected randomly from the given solution space provided by the different experts in the form of rules and membership functions are also associated with these rules. This is also known as the domain knowledge. This knowledge is updated at each time during the evaluation process of the algorithm. [...]
[...] Booker, D. E. Goldberg, and J. H. Holland: Classifier System and Genetic Algorithms, Artificial Intelligence, Vol No. September 1989. M. Grabisch and J. M. Nicolas: Classification by Fuzzy Integrals: Performance and tests, Fuzzy sets and systems, Vol No. August 1994. [12]. [13]. [14]. J. H. Holland: Adaptation in Natural and Artificial System, University of Michigan Press, in the year 1995. H. Ishibuchi, N. Yamamoto, and H. Tanaka: Selecting Fuzzy if [...]
APA Style reference
For your bibliographyOnline reading
with our online readerContent validated
by our reading committee
