Two Integrals
January 11, 2011
The exponential integral appears frequently in the study of physics, and the related logarithmic integral appears both in physics and in number theory. With the Euler-Mascheroni constant γ = 0.5772156649015328606065, formulas for computing the exponential and logarithmic integral are:
Since there is a singularity at Li(1) = −∞, the logarithmic integral is often given in an offset form with the integral starting at 2 instead of 0; the two forms of the logarithmic integral are related by Lioffset(x) = Li(x) – Li(2) = Li(x) – 1.04516378011749278. It is this form that we are most interested in, because the offset logarithmic integral is a good approximation of the prime counting function π(x), which computes the number of primes less than or equal to x:
x | 106 | 1021 |
Lioffset(x) | 78627 | 21127269486616126182 |
π(x) | 78498 | 21127269486018731928 |
If you read the mathematical literature, you should be aware that there is some notational confusion about the two forms of the logarithmic integral: some authors use Li for the logarithmic integral and li for its offset variant, other authors turn that convention around, and still other authors use either notation in either (or both!) contexts. The good news is that in most cases it doesn’t matter which variant you choose.
Your task is to write functions that compute the exponential integral and the two forms of the logarithmic integral. When you are finished, you are welcome to read or run a suggested solution, or to post your own solution or discuss the exercise in the comments below.