山区路网震后分阶段应急及资源保障集成优化研究

The montanic road-network is restricted by terrain, comparing with the urban road-network, connectivity is low and the shape like tree or string. The montanic road-network is easy to be affected by the earthquake and then hindering the emergency rescue, which highlights the importance of montanic road-network emergency management after an earthquake. In addition, the post-earthquake montanic road-network emergency management presented apparent phases of features and current researches mainly focus on the period of repair of urban road-network recovery, what’s more, these researches were separated with supply support. However, the road-network emergency management cannot do without timely supply support. Therefore, it is needed to analysis the montanic road-network topological structure and the phases of montanic road-network emergency management features, and then to integrate optimization of post-earthquake montanic road-network phases of emergency management and supply support: 1) in the period of repair, considering time is limited and demand is fuzzy, researches on the problem of integrating optimization of repairing schedule of the montanic road-network and supply support with virtual repair point, traffic control and fuzzy demand; 2) in the phase of restoration, considering the secondary disasters happened suddenly, researches on the problem of integrating optimization of dynamical restoration schedule of montanic road-network and supply support with stochastic demand; 3) in the stage of rehabilitation, the main problem is priority rehabilitation-based montanic road-network resources allocation optimization. Finally, we take the earthquakes were broke-out in China as background, empirical researches on post-earthquake montanic road-network phases of emergency management and supply support, and an assistant decision support system is developed. The researching results have important theoretical and practical significance to improve our country and even the world post-earthquake montanic road-network emergency management level.

山区路网因地形限制，相对城市路网，呈树形或串形，表现为低连通度，易受地震影响阻碍应急救援，凸显山区路网震后应急的重要性。此外，路网震后应急还呈现阶段性特征，当前研究着重震后抢通阶段城市路网的应急且与资源保障是分开研究的，然而，路网震后应急离不开资源保障。故需分析山区路网结构及震后分阶段应急特性，研究山区路网震后分阶段应急及资源保障集成优化问题：1) 抢通阶段，考虑时间紧迫与需求模糊，对虚拟抢修点、交通管制下及模糊需求下的抢修排程及资源保障集成优化问题进行研究；2) 保通阶段，考虑突发次生灾害，研究随机需求下震后保通动态排程及资源保障集成优化问题；3) 灾后重建阶段，基于山区路网灾后重建优先级的资源优化配置问题。最后，以我国地震实际为依据，实证研究我国山区路网震后分阶段应急及资源保障集成优化问题，开发一套辅助决策支持系统。课题成果对提高我国乃至世界山区路网震后应急水平具有重要理论和现实意义。

由于受地理位置、地形地貌等因素的影响，山区路网多呈树突形结构，表现为路网密度低、连通度低、可达性低、可选出行路线少等特点。一旦发生中断，将会严重阻碍震后应急救援工作的开展，造成大量人员伤亡和财产损失。因此，需要对山区受损路网进行恢复。. 然而，震后不同阶段对交通路网需求不同，资源保障也存在差异。因此，本课题将山区路网震后恢复及资源保障集成优化问题划分为三个阶段：（1）震后抢通阶段，通信不畅，受损信息通过遥感、航拍技术获取，在有限时间和资源约束下，考虑路网交通流动态变化情况，建立了山区路网震后抢修排程与应急物资配送集成优化模型，分别设计了遗传算法（GA）、细菌群趋药性算法（BCCOA）以及改进的BCCOA（采用混沌迁移策略，微分进化策略，精度自适应策略及目标函数值归一化策略进行改进）。通过对算法计算效果进行对比分析表明：大规模问题，改进的BCCOA相较于GA，解的质量提升了20%，计算效率提升了20%。（2）震后保通阶段，通信基本恢复，受损信息能够及时获得，为提高抢通路段通行能力，以路网抢修绩效最大为目标，在有限时间和资源约束下，以路网交通流动态变化情况为依据，以OD对通行时间变化衡量路网抢修绩效，分别建立了以路网抢修绩效最大为目标的山区路网震后抢修排程与应急物资配送集成优化模型和山区路网震后多周期恢复问题：抢修排程与应急物流集成优化模型，并设计了混合稳态并行遗传算法（HSSPGA）对模型进行求解。采用GNETGEN产生不同规模的算例对上述模型进行验证。（3）灾后重建阶段，通信完全恢复，对原路网受损路段实施重建。建立了动态交通流下以最大化路网重建累积绩效为目标的山区路网灾后重建排程优化模型，并设计了一种混合遗传算法对模型进行求解，并将动态交通流下的灾后路网重建排程优化模型与非动态交通流下优化模型进行比较，结果表明，动态交通流优化模型的重建累积绩效提升了8%。（4）以汶川大地震为背景，构建了系列算例，对上述模型和算法进行了有效性和可靠性验证。