资 源 简 介
FuzzyGPU is a fuzzy arithmetic library written in CUDA, where basic arithmetic operators have been overloaded using C++.
The library is based on the α-cuts concept and relies on interval and rounded floating point arithmetic optimized for different hardware. The implementation offers two fuzzy classes. The first one uses a traditional lower-upper encoding for intervals and allows to represent any fuzzy number, regardless of the shape of the membership function. The second one utilizes an optimized midpoint-radii encoding for fuzzy numbers having a symmetric membership function, which notably improves computational efficiency. The optimization consists of using a single variable to store all the α-cuts’s common midpoint.