平衡二分堆的使用

java.util.Timer是一个用来调度后台任务的类，TimerTask就是一个任务。
里面用了一个平衡二分堆的算法：

/**
 * This class represents a timer task queue: a priority queue of TimerTasks,
 * ordered on nextExecutionTime.  Each Timer object has one of these, which it
 * shares with its TimerThread.  Internally this class uses a heap, which
 * offers log(n) performance for the add, removeMin and rescheduleMin
 * operations, and constant time performance for the the getMin operation.
 */
class TaskQueue {
    /**
     * Priority queue represented as a balanced binary heap: the two children
     * of queue[n] are queue[2*n] and queue[2*n+1].  The priority queue is
     * ordered on the nextExecutionTime field: The TimerTask with the lowest
     * nextExecutionTime is in queue[1] (assuming the queue is nonempty).  For
     * each node n in the heap, and each descendant of n, d,
     * n.nextExecutionTime <= d.nextExecutionTime.
     */
    private TimerTask[] queue = new TimerTask[128];

    /**
     * The number of tasks in the priority queue.  (The tasks are stored in
     * queue[1] up to queue[size]).
     */
    private int size = 0;

    /**
     * Adds a new task to the priority queue.
     */
    void add(TimerTask task) {
        // Grow backing store if necessary
        if (++size == queue.length) {
            TimerTask[] newQueue = new TimerTask[2*queue.length];
            System.arraycopy(queue, 0, newQueue, 0, size);
            queue = newQueue;
        }

        queue[size] = task;
        fixUp(size);
    }

    /**
     * Return the "head task" of the priority queue.  (The head task is an
     * task with the lowest nextExecutionTime.)
     */
    TimerTask getMin() {
        return queue[1];
    }

    /**
     * Remove the head task from the priority queue.
     */
    void removeMin() {
        queue[1] = queue[size];
        queue[size--] = null;  // Drop extra reference to prevent memory leak
        fixDown(1);
    }

    /**
     * Sets the nextExecutionTime associated with the head task to the
     * specified value, and adjusts priority queue accordingly.
     */
    void rescheduleMin(long newTime) {
        queue[1].nextExecutionTime = newTime;
        fixDown(1);
    }

    /**
     * Returns true if the priority queue contains no elements.
     */
    boolean isEmpty() {
        return size==0;
    }

    /**
     * Removes all elements from the priority queue.
     */
    void clear() {
        // Null out task references to prevent memory leak
        for (int i=1; i<=size; i++)
            queue[i] = null;

        size = 0;
    }

    /**
     * Establishes the heap invariant (described above) assuming the heap
     * satisfies the invariant except possibly for the leaf-node indexed by k
     * (which may have a nextExecutionTime less than its parent's).
     *
     * This method functions by "promoting" queue[k] up the hierarchy
     * (by swapping it with its parent) repeatedly until queue[k]'s
     * nextExecutionTime is greater than or equal to that of its parent.
     */
    private void fixUp(int k) {
        while (k > 1) {
            int j = k >> 1;
            if (queue[j].nextExecutionTime <= queue[k].nextExecutionTime)
                break;
            TimerTask tmp = queue[j];  queue[j] = queue[k]; queue[k] = tmp;
            k = j;
        }
    }

    /**
     * Establishes the heap invariant (described above) in the subtree
     * rooted at k, which is assumed to satisfy the heap invariant except
     * possibly for node k itself (which may have a nextExecutionTime greater
     * than its children's).
     *
     * This method functions by "demoting" queue[k] down the hierarchy
     * (by swapping it with its smaller child) repeatedly until queue[k]'s
     * nextExecutionTime is less than or equal to those of its children.
     */
    private void fixDown(int k) {
        int j;
        while ((j = k << 1) <= size) {
            if (j < size &&
                queue[j].nextExecutionTime > queue[j+1].nextExecutionTime)
                j++; // j indexes smallest kid
            if (queue[k].nextExecutionTime <= queue[j].nextExecutionTime)
                break;
            TimerTask tmp = queue[j];  queue[j] = queue[k]; queue[k] = tmp;
            k = j;
        }
    }
}