Fuzzy rule base learning through simulated annealing

We study the use of simulated annealing to optimize the membership function s of Takagi-Sugeno rules. The necessary adaptation of simulated annealing i n order to be efficient for this problem is discussed in detail. The conver gence is carefully studied for the test application of the approximation of an analytical function specially built to test the efficiency of the algor ithm. The obtained results are compared with gradient descent optimization results. We point out that simulated annealing is particularly interesting in the case (usual in practical implementations) when there are few rules c ompared to the complexity of the problem. (C) 1999 Elsevier Science B.V. Al l rights reserved.