Integrated Optimization On Shelter Location And Evacuation Routing Problem Under Landslide Disasters

Our country is prone to geological disasters.Major geological disasters that are represented by earthquakes and giant landslides occur frequently,seriously threatening people’s lives.Among them,landslides are the most common,accounting for more than92% of the national disaster occurrence frequency and impact degree.Major landslide geological hazard events have become the focus and difficulty of research in the field of emergency management due to their characteristics of suddenness,strong destructiveness,the urgency of time,high uncertainty,and derivation and evolution.To carry out emergency management of landslide geological disasters,it is not only necessary to continuously improve the identification,monitoring,and early warning capabilities of landslide geological disasters,but also urgent to improve the emergency response capabilities to such disasters and to strengthen the location and layout of shelters,and evacuation optimization.To more effectively respond to complex emergencies,the Ministry of Emergency Management of the People’s Republic of China has been established in 2018 to improve the overall coordination of emergency management.In 2019,the Ministry of Natural Resources of the People’s Republic of China deployed and carried out the demonstration work of identifying hidden dangers of geological disasters based on comprehensive remote sensing technology,and the effect was obvious.However,there are many and wide-ranging earthquakes,landslides,collapses,debris flows,and other geological disasters that have a large number of disaster-causing factors in our country,and 70% of the new geological disasters that occur every year in our country are outside the scope of identified hidden dangers.Landslide geological disaster emergency management still faces a series of challenges,such as the shelter location and evacuation problem in mountainous areas with many hidden disaster points,the shelter location and evacuation problem based on uncertain disaster information,the shelter location and evacuation problem taking into account the evacuation time,and the shelter location and evacuation problem considering the issues of secondary disaster.So,this thesis takes Zigui County in the Three Gorges Reservoir area as the research object to study the shelter location and evacuation route optimization of landslide geological disaster refuge.Shelter location problem and evacuation routing problem are decision-making issues in two different stages.First,considering evacuation routing decisions in the predisaster shelter location and layout stage,and conducting evacuation routing simulation,how to achieve “the shorter the evacuation path”,“the farther between the shelter and the hidden disaster point”,“fewer shelter construction and maintenance costs”,and other goals,and simultaneously obtain a two-stage overall decision-making optimization solution with a determined disaster scenario? Second,taking into account the factors that landslide geological disasters are difficult to accurately predict and the uncertainty of disaster scenarios,how to optimize the shelter location and evacuation routing,and obtain a robust decision-making solution for shelter and evacuation routing problem? Third,considering the uncertainty of landslide geological disasters,how to quantify the evacuation time of large-scale people(non-linear expression of evacuation time),optimize the shelter location and evacuation routing problem,and then obtain a robust shelter location and evacuation routing solution based on the nonlinear expression of evacuation time? Furthermore,based on the above model,how to add the factors of the disaster chain to solve the problem? Moreover,given the characteristics of large-scale landslides in Zigui County with many hidden danger points and high disaster susceptibility,how to conduct a case analysis of shelter location and evacuation routing in Zigui County,and put forward reasonable countermeasures and suggestions.Because of the above problems,the main research contents of this thesis are as follows:(1)Optimization on shelter location and evacuation routing problem under determined disaster scenarios.With the development of science and technology and the continuous efforts of geologists,it is possible to accurately predict the occurrence of some landslide geological disasters.Based on the determined disaster scenarios,the shelter location and evacuation routing problem need to consider the cost of shelter location,evacuation time,the number of shelters,shelter capacity,road flow conservation,and other practical factors.In this regard,this thesis firstly establishes a two-stage mathematical programming model aiming at minimizing the cost of shelter location and evacuation cost(time)and constrained by other realistic factors.Then,set and adjust the parameters of the Gurobi solver to improve the solution efficiency.Finally,the more detailed case data are sorted out and substituted into the model for a solution and sensitivity analysis.The results show that the solution time of the model can be shortened by deleting redundant constraints,determining suitable decision variable value range,and adding effective cutting planes.And that the change of model input parameters will affect the value of the objective function and the shelter location results,but the impact fo the shelter layout is small,that is to say,the positions of the shelters in different shelter location solutions are not the same,but the layout of the shelters is the same.(2)Two-stage robust optimization of shelter location and evacuation route under disaster uncertainty scenarios.Because most of the landslide geological disasters that have occurred are still unable to make accurate pre-disaster predictions,and the shelter location decisions are long-term decisions made before the disaster,the shelter location decision cannot ignore the uncertainty of disasters.Therefore,this thesis firstly developed a two-stage robust optimization(min-maxmin)model for shelter location and evacuation routing problem in the worst disaster scenarios.The first min is used to make the shelter location and its capacity setting in the first stage,max is used to generate the worst disaster scenario,min is used to find the optimal evacuation routing under the worst disaster scenario,that is,maxmin is to find the worst scenario and optimal evacuation paths under worst-case disaster scenarios.Then a column and constraint generation algorithm is constructed for the two-stage robust optimization model,where the main problem is the two-stage optimization model for a given subset of disaster scenarios,and the sub-problem is the second-stage maxmin model for a given shelter location and its capacity setting decision variables.Iteration of the problem to find the optimal solution.Finally,the solution and sensitivity analysis are carried out for the case data in Zigui County.The results show that the two different sub-problem construction methods can both solve the problem and obtain the optimal solution(Gap is less than or equal to 1.53%).And the best robust solution under the disaster uncertainty scenario is different from the optimal solution in the disaster determination scenario.(3)Two-stage robust optimization of shelter location and evacuation routing poblem considering traffic flow.Because the travel speed during evacuation is affected by factors such as the number of crowds and road capacity,which in turn affect the selection of evacuation routing,the impact of road traffic flow needs to be considered in the process of shelter location and evacuation routing optimization.Therefore,this thesis firstly introduces a high-order nonlinear traffic flow expression,that is,the actual traffic time of the road is presented by a high-order nonlinear expression composed of variables such as barrier-free travel time,road capacity,and road flow,and is substituted into the two-stage robust optimization model.Then,the high-order nonlinear expression is represented by the second-order cone through the power reduction operation.Then the dual model of the second-order cone problem was deduced,and it is substituted into the column and constraint generation algorithm.Finally,the mathematical model was solved and the sensitivity analysis was carried out.The research results show that when the road is greatly affected by the environment,the shelter location and evacuation routing are more inclined to the main road(national road and provincial road).(4)Two-stage robust optimization of shelter location and evacuation routing considering the disaster chain.The occurrence of a disaster chain will place higher requirements on shelters and make evacuation more difficult.This thesis firstly assumes that a two-level disaster chain occurs,and constructs a two-stage robust optimization model for shelter location and evacuation routing.Then the constraints of the two-level disaster chain are simplified to obtain a mathematical model with the same structure as(3),and the algorithm in(3)is used to solve and analyze the sensitivity.Finally,summarize and analyze the sensitivity analysis results of the landslide geological disaster shelter location and evacuation routing optimization in all scenarios in this thesis,and select No.44,50,57,65,69,79,80,120,121,123,126 and 129 as the main shelters,also proposed countermeasures related to material storage and evacuation routings,providing a reference for the layout of shelters,the storage of emergency material,and evacuation routing in Zigui County.The research work of this thesis is mainly about the optimization model and algorithm of landslide geological disaster shelter location and evacuation routing.The innovations are:(1)Established a joint optimization model of shelter location and evacuation routing considering the safety of the shelters.(2)Established a two-stage robust optimization model for shelter location and evacuation routing under uncertain disaster scenarios,and constructed a two-step sub-problem-based column and constraint generation algorithm.(3)Established a two-stage nonlinear robust optimization model for shelter location and evacuation routing considering traffic flow.And a column and constraint generation algorithm based on high-order nonlinear solution technology is constructed.(4)A two-stage nonlinear robust optimization model for shelter location and evacuation routing based on disaster chains is proposed.