Type-2 Fuzzy Logic Augmentation of the Imperialist Competitive Algorithm with Dynamic Parameter Adaptation

Abstract

In this paper we propose the utilization of type-2 fuzzy systems for the dynamic adjustment of parameters in the imperialist competitive algorithm (ICA), a type-1 fuzzy system was used as a basis, with decades as the input variable and the beta parameter as the output variable, then it was extended to interval type-2 fuzzy systems, and three variants with triangular, Gaussian and trapezoidal membership functions were performed. The imperialist competitive algorithm is based on the concept of imperialism, where the strongest countries absorb the weakest and make then their colonies. To measure the performance of the proposed method 10 mathematical functions with different number of decades are used and finally, a comparison was made between our variants and the results obtained with the type-1 fuzzy system to observe their behavior in the face of optimization problems.