birm / calculus.jl

Calculus functions in Julia

Calculus.jl

Introduction

The Calculus package provides tools for working with the basic calculus operations of differentiation and integration. You can use the Calculus package to produce approximate derivatives by several forms of finite differencing or to produce exact derivative using symbolic differentiation. You can also compute definite integrals by different numerical methods.

API

Most users will want to work with a limited set of basic functions:

• derivative(): Use this for functions from R to R
• second_derivative(): Use this for functions from R to R
• Calculus.gradient(): Use this for functions from R^n to R
• hessian(): Use this for functions from R^n to R
• differentiate(): Use this to perform symbolic differentiation
• simplify(): Use this to perform symbolic simplification
• deparse(): Use this to get usual infix representation of expressions

Usage Examples

There are a few basic approaches to using the Calculus package:

• Use finite-differencing to evaluate the derivative at a specific point
• Use higher-order functions to create new functions that evaluate derivatives
• Use symbolic differentiation to produce exact derivatives for simple functions

Direct Finite Differencing

using Calculus

# Compare with cos(0.0)
derivative(sin, 0.0)
# Compare with cos(1.0)
derivative(sin, 1.0)
# Compare with cos(pi)
derivative(sin, float(pi))

# Compare with [cos(0.0), -sin(0.0)]
Calculus.gradient(x -> sin(x[1]) + cos(x[2]), [0.0, 0.0])
# Compare with [cos(1.0), -sin(1.0)]
Calculus.gradient(x -> sin(x[1]) + cos(x[2]), [1.0, 1.0])
# Compare with [cos(pi), -sin(pi)]
Calculus.gradient(x -> sin(x[1]) + cos(x[2]), [float64(pi), float64(pi)])

# Compare with -sin(0.0)
second_derivative(sin, 0.0)
# Compare with -sin(1.0)
second_derivative(sin, 1.0)
# Compare with -sin(pi)
second_derivative(sin, float64(pi))

# Compare with [-sin(0.0) 0.0; 0.0 -cos(0.0)]
hessian(x -> sin(x[1]) + cos(x[2]), [0.0, 0.0])
# Compare with [-sin(1.0) 0.0; 0.0 -cos(1.0)]
hessian(x -> sin(x[1]) + cos(x[2]), [1.0, 1.0])
# Compare with [-sin(pi) 0.0; 0.0 -cos(pi)]
hessian(x -> sin(x[1]) + cos(x[2]), [float64(pi), float64(pi)])

Higher-Order Functions

using Calculus

g1 = derivative(sin)
g1(0.0)
g1(1.0)
g1(pi)

g2 = Calculus.gradient(x -> sin(x[1]) + cos(x[2]))
g2([0.0, 0.0])
g2([1.0, 1.0])
g2([pi, pi])

h1 = second_derivative(sin)
h1(0.0)
h1(1.0)
h1(pi)

h2 = hessian(x -> sin(x[1]) + cos(x[2]))
h2([0.0, 0.0])
h2([1.0, 1.0])
h2([pi, pi])

Prime Notation

For scalar functions that map R to R, you can use the ' operator to calculate derivatives as well. This operator can be used abritratily many times, but you should keep in mind that the approximation degrades with each approximate derivative you calculate:

using Calculus

f(x) = sin(x)
f'(1.0) - cos(1.0)
f''(1.0) - (-sin(1.0))
f'''(1.0) - (-cos(1.0))

Symbolic Differentiation

using Calculus

differentiate("cos(x) + sin(x) + exp(-x) * cos(x)", :x)
differentiate("cos(x) + sin(y) + exp(-x) * cos(y)", [:x, :y])

Numerical Integration

The Calculus package no longer provides routines for univariate numerical integration. Use QuadGK.jl instead.

Credits

Calculus.jl is built on contributions from:

• John Myles White
• Tim Holy
• Andreas Noack Jensen
• Nathaniel Daw
• Blake Johnson
• Avik Sengupta
• Miles Lubin

And draws inspiration and ideas from:

• Mark Schmidt
• Jonas Rauch