To determine whether there exists a finite set A of positive integers of at least two elements and an infinite set B of positive integers such that any two distinct elements in A+B are coprime, and for any coprime positive integers m,n, there exists an element x in A+B satisfying x≡n(modm), we proceed as follows:
### Proof by Contradiction
Assume, for the sake of contradiction, that such sets A and B exist, with A={a1,a2,…,ak}.
#### Lemma 1
If (A,B) satisfy the conditions of the problem with ∣A∣=k, then there exist infinitely many b∈B such that the number ai+b has at least k distinct prime divisors.
Proof of Lemma 1:
Let m1,m2,…,mk be pairwise coprime numbers with at least k prime divisors each. Let m=m1m2…mk, and let n be chosen such that n≡ai−ai+1(modmi) for all i (indices are modulo k). Since (A,B) satisfy the conditions of the problem, there exists ai,b such that
ai+b≡n(modm).
Therefore,
ai+b≡ai−ai+1(modmi).
Now,
ai+1+b≡0(modmi),
which proves the statement.
#### Main Proof
Let b be an arbitrarily large number satisfying the condition of Lemma 1. Without loss of generality (after suitable relabeling), let the respective ai be a1.
Let p1,p2,…,pk be distinct primes dividing b+a1. Choose m=p1p2…pk, and n such that
n≡ai−ai+1(modpi)
for all 1≤i≤k, where indices are considered modulo k. By the Chinese Remainder Theorem, n corresponds to some residue modulo m. By assuming that pi are large enough (that is, larger than ∣ai+1−ai∣ for all i), we can guarantee (m,n)=1. To get pi large, just increase b if necessary: as the elements of A+B are coprime, only finitely many b′∈B may be such that b′+ai has a "small" prime divisor dividing ∣ai+1−ai∣ for some i.
Now, assume there exists some c∈B such that x=ai+c (1≤i≤n) satisfies x≡n(modm). This, in particular, implies
ai+c≡ai−ai+1(modpi).
This implies
ai+1+c≡0(modpi),
with ai+1+c∈A+B. Since the elements of A+B are pairwise coprime, this implies c=b, with the b described as above. Additionally,
ai+1+b≡0(modp1),
and
a1+b≡0(modp1)
by definition, so ai+1=a1, and therefore i≡0(modk). That is,
a1+b≡n(modm).
Now,
a1+b≡n≡a1−a2(modp1),
and also
a1+b≡0(modp1).
This means that a1≡a2(modp1), which cannot happen for big enough p1. As explained before, we can force the pi to be as large as we want. This, finally, is a contradiction.
Therefore, the answer is: \boxed{\text{No}}.