Type reduction is the work of computing the centroid of a type-2 fuzzy set. At present, most of type reduction
methods have high computational cost. The enhanced Karnik-Mendel(EKM) algorithm can compute the centroid of an
interval type-2 fuzzy set efficiently. However, the initialization of the switch point in the EKM algorithm is not a good
one, and the computation time is long, which makes a limit on the application in real system. In view of these problems,
a novel improved EKM algorithm is developed for improving the EKM algorithm. The proposed algorithm provides two
improvements on the EKM algorithm. Firstly, the initialization conditions of switch points are changed. Then, the method of
searching for switch points is improved, in which can search upward and downward. The number of computations involved
is greatly reduced and type reduction can be done much more efficiently. The simulation results show the effectiveness of
the proposed method.