Source: Imperial College Mathematics Competition 2018/19 - Round 2
August 7, 2020
college contestsfunction
Problem Statement
Let f:{0,1}n→{0,1}⊆R be a function. Call such a function a Boolean function.
Let ∧ denote the component-wise multiplication in {0,1}n. For example, for n=4, (0,0,1,1)∧(0,1,0,1)=(0,0,0,1).Let S={i1,i2,…,ik}⊆{1,2,…,n}. f is called the oligarchy function over S if f(x)=xi1,xi2,…,xik (with the usual multiplication),where xi denotes the i-th component of x. By convention, f is called the oligarchy function over ∅ if f is constantly 1.(i) Suppose f is not constantly zero. Show that f is an oligarchy function if and only if f satisfies f(x∧y)=f(x)f(y),∀x,y∈{0,1}n.
Let Y be a uniformly distributed random variable over {0,1}n. Let T be an operator that maps Boolean functions to functions {0,1}n→R, such that (Tf)(x)=EY(f(x∧Y)),∀x∈{0,1}n where EY() denotes the expectation over Y. f is called an eigenfunction of T if ∃λ∈R\{0} such that
(Tf)(x)=λf(x),∀x∈{0,1}n(ii) Prove that f is an eigenfunction of T if and only if f is an oligarchy function.