位于中国西南边陲的西藏自治区，地震活动十分强烈和频繁。在这种海拔高、地震危险性高、道路少的高原山地地区，在整体路网脆弱性的影响下，震后道路的破坏将是限制该地区快速救援的一个重要因素，震后路网的通达情况将会严重影响救援的进度和效率。同时，由于高原山地地区相对平原地区道路更为稀少且可替代性更低，因此在局部路段脆弱性的影响下，高原山地地区路网内的某些路段相较于其他路段而言更为重要，这些关键路段一旦无法通行则可能导致全区域无法通达。因此，大区域范围内整体路网的脆弱性和小区域范围内局部路段的脆弱性息息相关，在震后研究路网脆弱性的时候，不但应该研究整体路网通过本次地震所体现出来的脆弱性——即道路破坏的相关情况，也应该对局部路段的脆弱性进行进一步的评估和分析来找到整体路网内的关键路段，以探讨整体和局部、大范围和小范围脆弱性之间的相互关系。

因此，本研究通过收集研究区相关的包括桥梁、隧道和道路易损性等的承灾体数据、包括震中、震级和震源深度等的地震数据以及包括断层、河流、高程和坡度等的其他灾害相关数据，首先通过模拟计算得到地震的道路破坏结果并与实际地震的道路破坏结果比较以验证可靠性，然后1）针对不同救援阶段和救援车辆的道路需求差异，结合迪杰斯特拉算法和蚁群算法规划总体应急救援方案并验证合并算法的合理性，给出III级响应、IV级响应两种情景下的完善的应急救援方案规划结果；2）构建空间效应模型，采用路网鲁棒性指数法（NRI）分析影响路网内路段重要性的相关因素，再将该方法运用在考虑了实际地震影响的研究区路网中，对路网内关键路段进行甄别与分析。最后结合以上两部分内容综合分析整体路网脆弱性对局部路段的影响和局部路段脆弱性对整体路网的影响，并据此讨论路网脆弱性的空间效应，为完善震后脆弱性评估的相关研究、未来路网脆弱性研究、应急救援路线规划等方面提供更好的指导建议。

研究的主要结论包括：

（1）受到整体路网脆弱性的影响，震后路段会出现不同程度的破坏。通过充分考虑实际震后道路破坏情况并结合应急救援过程中不同阶段特点，分别采用迪杰斯特拉算法以规划工程车辆救援方案、采用蚁群算法以规划医疗物资车辆救援方案；对于需要结合不同阶段特点的转移安置车辆救援方案，则通过结合以上两种算法以充分解决这个问题。而验证发现合并算法在缩放区域大小的情况下对最终规划路线的结果影响较小，因此可以认为本研究所采用的基于道路通达性的高原地区震后应急救援方案结果具有一定的科学性和合理性，对于不同大小区域也有其适用性，能够充分解决震后路网脆弱性空间效应中由于受到整体路网脆弱性影响而出现不同的局部路段破坏后的应急救援方案路线规划问题。在本研究的所得结果中，对应于III级响应情景下规划所得的路径主要由国道G318和G219、省道S209以及县道X214构成；对应于IV级响应情景下规划所得的路径主要由国道G318、省道S209、县道X214以及各乡镇内部道路构成。

（2）受到局部路段脆弱性的影响，路网内的不同路段对整体路网存在不同的影响，空间效应模型能够较好地分析影响路网内路段重要性的相关因素。在这些因素中，路段级别越高、连接的区域范围差异越大、周围路段对其替代性越低、受到地震影响后道路综合破坏等级越低，该路段就越为重要，据此计算得到的日喀则地区关键路段包括：国道G318拉孜县到热萨乡段、国道G219桑桑镇到切热乡段、国道G219、G349、G216切热乡到萨嘎县段、国道G216萨嘎县到吉隆县段、省道S209岗嘎镇到绒辖乡段和国道G219定结县到岗巴县段等。在一个路网中，一次灾害所体现出来的被完全破坏的路段，即从整体路网脆弱性角度出发找到的关键路段和通过NRI值比较得出的数值较高的路段，即从局部路段脆弱性角度出发找到的关键路段尽管会存在一定的重合，但在最关键路段的寻找上，二者仍然存在较大差异。对日喀则地区而言，对整体路网影响最大的国道G318拉孜县到热萨乡段仅能从NRI值的比较中体现。

（3）作为整体的路网由作为局部的路段所组成，整体路网的脆弱性和局部路段的脆弱性同时存在，息息相关，在灾后也会存在相互作用的效应。考虑了整体路网脆弱性对局部路段的影响后，可以提高震后应急救援方案路线规划的客观性和准确性；考虑了局部路段脆弱性对整体路网的影响后，可以快速寻找出路网内的一些关键路段。只有充分的结合大范围整体路网脆弱性对小范围局部路段以及小范围局部路段脆弱性对大范围整体路网这两方面的影响来探究路网脆弱性的空间效应，才能客观准确地分析震后路网的脆弱性。

The Tibet Autonomous Region, which located in the southwestern border of China, seismic activity is very strong and frequent. In this high-altitude, high-earthquake-risk and low road mountainous region, under the influence of the overall road network vulnerability, the destruction of roads after the earthquake will be an important factor limiting the rapid rescue in the region. Accessibility will seriously affect the progress and efficiency of the rescue. At the same time, due to the fact that roads in the plateau and mountainous areas are scarcer and less replaceable than plain areas, under the influence of local road vulnerability, some road sections in the road network in the plateau and mountainous areas are more important than others. Once these critical road sections are impassable, it may lead to the inaccessibility of the entire region. Therefore, the vulnerability of the overall road network in a large area is closely related to the vulnerability of local road sections in a small area. When studying the vulnerability of the road network after an earthquake, you should not only study the vulnerability of the overall road network through the earthquake，which is the relevant situation of road damage, the vulnerability of local road segments should be further evaluated and analyzed to find key road segments in the overall road network to discuss the overall and local, large-scale and small-scale vulnerability interrelationships.

In this study, we collected disaster-related body data including bridges, tunnels, and road fragility, seismic data including epicenters, magnitudes, and focal depths, and other disaster-related data including faults, rivers, elevations, and slopes. First, the road damage results of the earthquake are obtained through simulation calculations and compared with the actual road damage results to verify the reliability, and then 1) for different rescue stages and the road demand differences of rescue vehicles, combining Dijkstra ’s algorithm and ant colony algorithm to plan the overall emergency rescue plan and verifies the rationality of the merged algorithm, and gives the perfect emergency rescue plan results under the two scenarios of Level III response and level IV response scenarios; 2) Constructing a spatial effect model and using the road network robustness index Method (NRI) analyzes the relevant factors that affect the importance of road sections in the road network, and then applies this method to the road network in the study area that takes into account the impact of actual earthquakes to screen and analyze the key road sections in the road network. Finally, combined with the above two parts, comprehensively analyze the impact of the overall road network vulnerability on the local road section and the impact of the local road section vulnerability on the overall road network, and then discuss the spatial effect of the road network vulnerability, to provide better guidance and suggestions for improving research related to post-earthquake vulnerability assessment, future road network vulnerability research, and emergency rescue route planning.

The main conclusions of the study include:

(1) Affected by the vulnerability of the overall road network, different levels of damage will occur after the earthquake. By fully considering the actual road damage after the earthquake and combining the characteristics of different stages in the emergency rescue process, the Dijkstra algorithm is used to plan the engineering vehicle rescue plan, and the ant colony algorithm is used to plan the medical material vehicle rescue plan; The stage-specific transfer and resettlement vehicle rescue solution fully solves this problem by combining the above two algorithms. The verification found that the merge algorithm has little effect on the final planned route results when the area is scaled. Therefore, it can be considered that the results of the post-earthquake emergency rescue plan based on road accessibility in this study are scientific and reasonable. It also has its applicability to areas of different sizes, and can fully solve the problem of route planning for emergency rescue plans after different local road sections are damaged due to the vulnerability of the overall road network vulnerability after the earthquake. In the results obtained in this study, the path corresponding to the planned level III response scenario is mainly composed of the national highways G318 and G219, the provincial road S209 and the county road X214; the path corresponding to the planned level IV response scenario is mainly composed of the national highway G318 , Provincial Highway S209, County Highway X214 and the internal roads of each township.

(2) Affected by the vulnerability of local road segments, different road segments in the road network have different effects on the overall road network. The spatial effect model can better analyze the relevant factors that affect the importance of road segments in the road network. Among these factors, the higher the level of the road segment, the greater the difference in the range of the connected area, the lower the replacement of the surrounding road segments, and the lower the road damage level after the earthquake, the more important the road segment, according to the calculation of Shigatse The key sections of the region include: National Highway G318 from Lazi County to Resa Township, National Highway G219 from Sangsang Town to Chere Township, National Highway G219, G349, G216 from Chere Township to Saga County Section, National Road G216 from Saga County to Jilong County Section, Provincial Highway S209 Gangga Town to Rongxia Township Section and National Highway G219 Dingjie County to Gangba County Section, etc. In a road network, a completely damaged road segment reflected by a disaster, that is, a key road segment found from the perspective of the overall road network vulnerability and a road segment with a higher value obtained by comparing NRI values, that is, a partial road segment is vulnerable Although there will be a certain coincidence of the key sections found from a sexual perspective, there is still a big difference in the search for the most critical sections. For the Shigatse region, the national highway G318 from Lazi County to Resa Township, which has the greatest impact on the overall road network, can only be reflected from the comparison of NRI values.

(3) The road network as a whole consists of partial road sections. The vulnerability of the whole road network and the vulnerability of local road sections exist at the same time, are closely related, and there will also be an interactive effect after the disaster. After considering the impact of the overall road network vulnerability on local road sections, the objectivity and accuracy of the post-earthquake emergency rescue plan route planning can be improved; after considering the impact of the local road section vulnerability on the overall road network, you can quickly find the exit network Some key road sections. Only by fully combining the influence of the vulnerability of the large-scale overall road network on the small-area partial road sections and the vulnerability of the small-area partial road section on the large-scale overall road network to explore the spatial effect of the road network vulnerability, can the earthquake be objectively and accurately analyzed The vulnerability of the back road network.