Autonomous vehicles (AVs) can be stacked behind each other like valet parking due to their self-parking capability to increase land utilization and decrease the parking space allocation. However, such parking arrangements cause blockage and necessitate AVs relocation inside the car-park whenever a blocked AV is summoned. In this paper, we investigate AV parking spot selection based on their arrival and departure times to minimize the number of such relocation movements. We study both full and partial information scenarios in which we have exactly a priori information, and no a priori information about arrival and departure time, respectively. The results show that parking in the spot with the lowest blockage probability can decrease the number of relocation movements. Also, our analysis shows that retrieving vehicles from the rear side does not have a significant impact on the number of relocation movements.

中文翻译：

---

某停车场自动驾驶车辆搬迁问题

自动驾驶汽车 (AV) 可以像代客泊车一样相互堆叠，因为它们具有自动泊车能力，以提高土地利用率并减少停车位分配。然而，这种停车安排会导致阻塞，并且每当呼叫阻塞的 AV 时就需要将 AV 重新定位在停车场内。在本文中，我们研究了基于到达和离开时间的 AV 停车位选择，以最大限度地减少此类搬迁移动的次数。我们研究了完整和部分信息场景，其中我们分别有准确的先验信息和没有关于到达和离开时间的先验信息。结果表明，在堵塞概率最低的位置停车可以减少搬迁次数。还，