中文摘要：

本文针对主题公园游客时空分流的决策支持系统问题进行了研究。首先分析了游客在景区中的一次完整游憩行为，归纳出涉及这一行为的若干要素和状态。其次，为模拟这些要素并实现状态切换，设计了一种基于现场环境的游憩方向决策算法，该算法借助计算机推理技术并综合考虑游客自身的个性化需求、景点设施的运营情况和其他游客的排队位置等环境信息来确定下一个游憩方向。再者，为验证算法的有效性，基于离散事件建模开发了一个计算机仿真系统，并采用虚拟数据进行了实验模拟和比较分析。实验结果表明，与传统的基于最短距离的算法相比，本文算法在游客等待时间、游客游玩景点数量、景点容量利用率和景点平均等待时间4个评价指标上均有不错的表现，表明该决策支持系统具有时空一体化的特性。系统产生的推荐路线在理论上能让游客自动分流并导航到较少拥挤的景点上，既能均衡各景点的接待能力，同时降低游客的等待成本，有助于主题公园的客流时空分布调控。

英文摘要：

The paper is a study of the decision support system concerning personalized route guidance service for tourists in theme parks. It first outlined a single complete recreational behavior of a tourist while visiting at the park, thus summarizing the components and states related to this behavior. To simulate these components and accomplish the switching states, it presented, using computer reasoning technology, a state-of-art model （recreation direction） finding algorithm via contextual environment, which integrated the tourist ＇ s preferences, the availability of the targeted recreation facility （ amusement ride） and the queuing situation of other facilities, so as to get the single-step optimal direction. A computerized simulation system based on discrete event modeling was then implemented and experiment was conducted on an analog dataset as compared with the traditional shortest-path algorithm. The findings show that the proposed model outperforms its competitor in four evaluating indicators including tourist waiting time, amount of non-visited rides, utilization of each ride and average waiting cost on each ride. The validity of this study depends on its temporal-spatial integration by generating an appropriate visiting route for each tourist to follow, which in turn, theoretically, guides him/her to those vacant rides or less congested areas. All these indicate that this work caters to modulating the spatial and temporal distribution of tourist flow in theme parks, where queuing problem generally occurs, by balancing the capacity of each park ride as well as reducing the waiting cost of the tourists.