Parallel Algorithm for Numerical Methods Applied to Fractional-order System

Abstract

A parallel algorithm is presented that approximates a solution for fractional-order systems. The algorithm is implemented in CUDA, using the specific GPU capabilities. The numerical methods used are Adams-Bashforth-Moulton (ABM) predictor-corrector scheme and Diethelm’s numerical method. A comparison is done between these numerical methods that adapts the same algorithm for the approximation of the solution.