为什么基于优先级队列的Dijkstra最短路径算法不能用于负加权图？

Question

pq.insert(s, 0.0);
while (!pq.isEmpty()) {
  int v = pq.delMin();
  for (DirectedEdge e: G.adj(v))
    relax(e);
}

...  

void relax (DirectedEdge e) {
  int v = e.from(), w = e.to();
  if(distTo[w] > distTo[v] + e.weight()) {
    distTo[w] = distTo[v] + e.weight();
    edgeTo[w] = e;


    // once a node is newly relaxed, it is inserted to pq,
    // meaning that even if a node receives a negative (shorter)
    // path late, we won't miss its contribution; pq will still
    // dequeue it to relax its adjacent, am I correct?

    if (pq.contains(w)) pq.decreaseKey(w, distTo[w]);
    else pq.insert(w, distTo[w]);
  }
}

那么Dijkstra的算法为什么不能处理负权图？如果通过上述方法使用pq来存储具有distTo值更改的节点，那么它实际上与SPFA algorithm相同，只是SPFA algorithm可以使用FIFO队列而不是优先级队列？< / p>