Shapiro inequality

In mathematics, the Shapiro inequality is an inequality proposed by Harold S. Shapiro in 1954.

Statement of the inequality

Suppose n is a natural number and x1, x2, …, xn are positive numbers and:

Then the Shapiro inequality states that

∑i=1nxixi+1+xi+2≥n2,

where xn+1 = x1 and xn+2 = x2. The special case with n = 3 is Nesbitt's inequality.

For greater values of n the inequality does not hold, and the strict lower bound is γ ⁠n/2⁠ with γ ≈ 0.9891… (sequence A245330 in the OEIS).

The initial proofs of the inequality in the pivotal cases n = 12 and n = 23 rely on numerical computations. In 2002, P.J. Bushell and J.B. McLeod published an analytical proof for n = 12.

The value of γ was determined in 1971 by Vladimir Drinfeld. Specifically, he proved that the strict lower bound γ is given by ψ(0), where the function ψ is the convex hull of f(x) = e−x and g(x) = 2 / (ex + ex/2). (That is, the region above the graph of ψ is the convex hull of the union of the regions above the graphs of f and g.)

Interior local minima of the left-hand side are always ≥ n / 2.

Similar to Nesbitt’s inequality, there is no corresponding upper bound as the cyclic sum inequality can be made arbitrarily large.

Counter-examples for higher n

The first counter-example was found by Lighthill in 1956, for n = 20:

x20=(1+5ϵ, 6ϵ, 1+4ϵ, 5ϵ, 1+3ϵ, 4ϵ, 1+2ϵ, 3ϵ, 1+ϵ, 2ϵ, 1+2ϵ, ϵ, 1+3ϵ, 2ϵ, 1+4ϵ, 3ϵ, 1+5ϵ, 4ϵ, 1+6ϵ, 5ϵ),

where ϵ is close to 0. Then the left-hand side is equal to 10−ϵ2+O(ϵ3), thus lower than 10 when ϵ is small enough.

The following counter-example for n = 14 is by Troesch (1985):

x14=(0,42,2,42,4,41,5,39,4,38,2,38,0,40) (Troesch, 1985)