Numerical Approximations to Fractional Problems of the Calculus of Variations and Optimal Control

TL;DR

This paper reviews numerical methods for fractional calculus of variations and optimal control, focusing on discretization and series expansion approaches to approximate fractional derivatives and solve related problems.

Contribution

It introduces and compares two main approximation techniques for fractional derivatives, providing error bounds and practical examples for solving fractional variational problems.

Findings

Discretization using Grunwald-Letnikov method effectively approximates fractional terms.

Series expansion approximations enable transformation of fractional problems into classical ones.

Numerical examples demonstrate the accuracy and applicability of the proposed methods.

Abstract

This chapter presents some numerical methods to solve problems in the fractional calculus of variations and fractional optimal control. Although there are plenty of methods available in the literature, we concentrate mainly on approximating the fractional problem either by discretizing the fractional term or expanding the fractional derivatives as a series involving integer order derivatives. The former method, as a subclass of direct methods in the theory of calculus of variations, uses finite differences, Grunwald-Letnikov definition in this case, to discretize the fractional term. Any quadrature rule for integration, regarding the desired accuracy, is then used to discretize the whole problem including constraints. The final task in this method is to solve a static optimization problem to reach approximated values of the unknown functions on some mesh points. The latter method, however, approximates fractional problems by classical ones in which only derivatives of integer order are present. Precisely, two continuous approximations for fractional derivatives by series involving ordinary derivatives are introduced. Local upper bounds for truncation errors are provided and, through some test functions, the accuracy of the approximations are justified. Then we substitute the fractional term in the original problem with these series and transform the fractional problem to an ordinary one. Hereafter, we use indirect methods of classical theory, e.g. Euler-Lagrange equations, to solve the approximated problem. The methods are mainly developed through some concrete examples which either have obvious solutions or the solution is computed using the fractional Euler-Lagrange equation.