Evaluating Integrals Using Self-Similarity
As the author states, there are two familiar elementary methods for evaluating definite integrals: 1. Taking a limit of Riemann sums (the method of Archimedes); 2. Finding an antiderivative and using the Fundamental Theorem of Calculus (the method of Newton and Leibniz). This paper introduces the method of self-similarity to obtain the basic formula \[ \int_0^1x^ndx={1\over n+1}, \] and then basic properties of the integral (linearity, additivity, linear change of variable and integration of constants) which are easily derived from the definition of the integral, to obtain integrals of polynomial functions. The basic self-similar identity is \[ \int_0^1f(x)dx={1\over 2}\int_0^1f\biggl({1\over 2}x\biggr)dx+{1\over 2}\int_0^1f\biggl({1\over 2}x+{1\over 2}\biggr)dx. \] Of course this offers no advantage over the other methods for the integration of polynomials, but the remainder of the paper includes applications to more exotic integrals on the interval and integration on fractals. E.g., while \(\int_0^1f(x)dx\) can be interpreted as the expected value of \(f(x)\) with respect to the uniform distribution on \([0,1]\), one can use the method of self-similarity to obtain the expected value of \(f(x)\) with respect to other measures on the interval.
This page was built for publication: Evaluating Integrals Using Self-Similarity
Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2757465)
