use std::cmp::min;
use std::io::Read;
const INF: i64 = 1000000000;

fn dijkstra(start: usize, end: usize, graph: &Vec<Vec<(usize, i64)>>) -> i64 {
    let mut dist = vec![INF; graph.len()];
    dist[start] = 0;
    let mut bh = std::collections::BinaryHeap::new();
    // (startからその頂点に辿り着くコスト, 頂点番号)を優先度付きキューで管理
    // BinaryHeapは最大値を返すので、最小コストを得るためにコストは負の数で管理する。
    bh.push((0i64, start));
    while let Some((cost, v)) = bh.pop() {
        if dist[v] < -cost {
            continue;
        }
        for e in graph[v].iter() {
            let nv = e.0;
            let nc = e.1;
            if dist[nv] > -cost + nc {
                dist[nv] = -cost + nc;
                bh.push((-dist[nv], nv));
            }
        }
    }
    dist[end]
}

fn main() {
    let mut s: String = String::new();
    std::io::stdin().read_to_string(&mut s).ok();
    let mut itr = s.trim().split_whitespace();
    let n: usize = itr.next().unwrap().parse().unwrap();
    let s: Vec<i64> = (0..n)
        .map(|_| itr.next().unwrap().parse().unwrap())
        .collect();
    let m: usize = itr.next().unwrap().parse().unwrap();
    let mut graph: Vec<Vec<(usize, i64)>> = vec![Vec::new(); n];
    for _ in 0..m {
        let a: usize = itr.next().unwrap().parse().unwrap();
        let b: usize = itr.next().unwrap().parse().unwrap();
        let c: i64 = itr.next().unwrap().parse().unwrap();
        graph[a].push((b, c));
        graph[b].push((a, c));
    }

    let mut ans: i64 = INF;
    // どの２地点に止まるか。
    for i in 1..n - 1 {
        for j in 1..n - 1 {
            if i == j {
                continue;
            }
            ans = min(
                ans,
                dijkstra(0, i, &graph)
                    + dijkstra(i, j, &graph)
                    + dijkstra(j, n - 1, &graph)
                    + s[i]
                    + s[j],
            );
        }
    }
    println!("{}", ans);
}