Weak approximation for ℚ #
The weak approximation theorem for the rational numbers: a rational number can be found
arbitrarily close (simultaneously) to any prescribed real number and any finite collection of
p-adic numbers. This is the arithmetic input for the high-rank Hasse–Minkowski induction.
The proof is elementary. Fix tolerances, approximate each p-adic coordinate y_p by a
rational c_p, clear denominators with D = ∏ p, (c p).den and use the Chinese remainder
theorem to solve r ≡ D·c_p (mod p^{m}) for a single integer r. Then x = r/D already
matches every p-adic coordinate; adding an integer multiple of N = ∏ p^{m} (divided by D)
preserves the p-adic congruences as long as the multiplier is p-adically integral, and a
multiplier of the form s / L^j with L a prime outside the finite set is p-adically
integral for every p in the set while being dense in ℝ. This gives the real approximation.
The n-th prime is prime, as an instance.
The finite embedding of ℚ into the product of the completions of ℚ at a finite set of
places (which includes ℝ).
Equations
- Rat.finiteEmbedding S x = ((algebraMap ℚ ℝ) x, fun (p : ↥S) => (algebraMap ℚ ℚ_[↑↑p]) x)
Instances For
Elementary lemmas #
Weak approximation #
Given a finite set of places and a point in the product of the completions of ℚ at those places, there exists a rational number that is arbitrarily close to the given point at all those places.
The approximation theorem can be restated as saying that the finite embedding is dense.