בינה מלאכותית

ASTAR(state)
  OPEN (node(state, NIL, 0, h(state), h(state)))
  While OPEN ≠ ( )
    ; Take the next node to process, mark it is closed.
    next pop(OPEN)
    push (next, CLOSE)

    ; If we reached the target, print the way to it and
    ; finish. We knows that it is the shortest way we can find.
    if goalp(next.state) then return(trace(next))

    ; For each son s of the selected node
    for s in succ(next.state)

      ; Calculate the cost of s
      new-cost next.g+cost(next.state, s)

      ; Check if we already found it earlier,
      ; and also we didn't pass it yet (Not in close list)
      old-node find-state(s, OPEN)

      if old-node ≠ {} then

        ; OK, we found it earlier. Let check if we found
        ; better path to that node. If yes, we need to reenter
        ; it to the open list, according to its new cost
        if old-node.g > new-cost then
          old-node.g new-cost; old-node.f old-node.g+old-node.
         old-node.parent next; delete(old-node, OPEN);
          ; Reinsert old node according to new f
          insert(old-node, OPEN, f)
        end

      ; If the node is already in the closed list, and also
      ; we find better cost for it, we should take it
      ; back to the OPEN list
      else old-node find-state(s, CLOSE)
      if old-node ≠ {} then
      if old-node.g > new-cost then
        old-node.g new-cost; old-node.f old-node.g+old-node.
        old-node.parent next; delete(old-node, CLOSE);
        insert(old-node, OPEN, f)
      end

      ; If we are here, it means that we just reached that node
      ; in the first time. insert it to the OPEN list as new node
      else
        insert(node(s, next, new-cost, h(s), h(s)+new-cost), OPEN, f)
      end;
  end;
  return FAIL