I'll address the problem for subsets of $\{1,2,\ldots,N-1\}$ where the original problem is for $N=1000$.
First, I'll provide a case for which there is no solution: ie a set $U$ of size $n$ for which there are no $\{a,b,c,d\}\subset U$ with $a+2b+3c=d$.
Let $U=\{N-n,N-n+1,\ldots,N-1\}$. That makes $a+2b+3c\ge 6(N-n)+4$. If $6n< 5N+5$, then $a+2b+3c>N-1$ which ensures no solution for $d$ exists. Thus, if $n<5(N+1)/6$, there is a set $U$ of size $n$ for which no solution exists.
Next, let's assume we have a set $U$ of size $n$. Let $c<b<k$ be the three smallest elements of $U$, and let $U'=U\setminus\{b,c\}$. For $m=2b+3c$, we are looking for $a,d\in U'$ so that $a+m=d$. If there is no such solution, then the union of $U'$ and $U'+m$ must be $2n-4$ distinct numbers ranging from $k$ to $N+m-1$, which requires $2n-4\le N+m-k$, otherwise the pigeonhole principle ensures a collision between $U'$ and $U'+m$.
As $c<b<k$, we have $m=2b+3c\le 5k-8$, so $N+m-k\le N+4k-8$.
If $2n-4>N+4k-8$, we get $2n-4>N+m-k$ which is sufficient to ensure there are $a,d\in U'$ such that $a+m=d$. We know that $k\le N-n+2$, so $N+4k-8\le 5N-4n$. Thus, if $2n-4>5N-4n$, then we have
$$
2n-4>5N-4n\ge N+4k-8\ge N+m-k.
$$
Now, $2n-4>5N-4n \iff 6n>5N+4 \iff 6n\ge 5(N+1)$, so if $n\ge 5(N+1)/6$, there will always be a solution to $a+2b+3c=d$.
Unless I've made a slip-up some place, that should complete the proof.