A hybrid approach to optimize fuzzy if-then rules

In this paper we propose a hybrid approach attempting to find optimal value and points for profit and cost functions for which there is no equivalent equation. In this case only vague information as fuzzy if-then rules are accessible. This approach involves five steps: In the first step, fuzzy rules by expert information are collected and in the second step Mamdani inference method is applied to produce a surface called M-surface. As this surface is not a desirable one, so in the third step, data is extracted from the M-surface and they are used in Surrogate Software. Then in the forth step, by using various methods in Surrogate more accurate vision is built for the mentioned function of fuzzy if-then rules. In the final step a Genetic Algorithm is applied to optimize the result model. This approach can be used to optimize any unknown function via finding only a number of fuzzy if-then rules of this function by the aid of experts.