Scenario Deduction Method For Response Against Urban Cascading Disasters

Scenario is a basic description of the extent,scope and evolution of disasters.It is the data and information of the characteristics of disaster-causing factors(such as rainstorm),related disaster-bearing bodies(such as bridges and culverts),response bodies(such as emergency rescue teams)and disasterpregnant environment(such as water systems around cities)in a selected disaster situation.Different from general disasters,cascading disasters are accompanied by many kinds of disasters,and different disasters have potential correlation,which makes the evolution of scenarios more difficult to predict.In the implementation of scenario deduction,we need to consider a variety of complex situations.For cascading disasters,how to integrate multi-source disaster data,information and knowledge to achieve a comprehensive description and effective deduction of disaster scenarios,so as to serve emergency preparedness and mission planning,is an important issue of concern to emergency decision makers.The existing situation deduction research focuses on the situation deduction method of single disaster or simple disaster chain and its technical support,and pays less attention to the complex scenario situation evolution sc enario under cascading disaster background.In the context of scenario construction,the relationship between scenario elements is still unclear,and the slice description of scenario situation exists universally,which makes it difficult to guarantee the effectiveness of scenario recognition.In big data environment,the relationships among different levels,scales and granularity scenarios are closer.Under this background,this paper pays attention to the cognitive model and deduction methods of scenario situation of cascading disasters in cities,and carries out the following research work.Firstly,the construction of full-view scenario model for urban cascading disaster response is studied.This research mainly solves the problem of lack of global correlation of scenario description information of cascading disasters,divides scenario hierarchy,scale and granularity into three full-view dimensions,establishes scenario panoramic framework of cascading disasters,establishes full-view scenario model through feature association based on the structuring of scenario elements,and further proposes an iterative algorithm for full-view scenario generation,which integrates multi-level disaster scenario data,realizing the revision and perfection of scenario de scription.In addition,this part work summarizes the concept of cascading disaster scenario deduction on the basis of a large number of cascading disaster literature.In terms of the combining evolution,converting evolution and mapping evolution model of cascading disaster scenario situation,the scenario situation deduction is divided into three categories: combining deduction,converting deduction and mapping evolution,and the scenario situation deduction system of cascading disaster is proposed.Secondly,aiming at the problem of scenario evolution in which many kinds of disaster-causing factors coexist(called combin ing evolution in this paper),a scenario combining deduction method of urban cascading disasters is proposed.Based on historical disasters,this method uses information diffusion theory to analyze the probability of association between the combining of scenario elements and the latter scenario elements,that is,the possibility of the latter scenario occurrence,and mining the combining deduction rules.Compared with traditional methods,the proposed method divides the scenario deduction process more carefully.The method is flexible and comprehensive.It can be used in early warning,decision-making and case study to gain time for disaster prevention and response..Thirdly,aiming at the problem of scenario evolution from disaster-bearing body to disaster-causing factor(which is called converting evolution in this paper),a deduction method of scenario converting of urban cascade disasters is proposed.This method further reduces the deduction granularity,studies the more detailed disaster-bearing body,specifies the descriptive features of scenario elements,and makes the results more precise.This method uses machine learning combined with historical cases and network data to model scenario converting network,which can better provide the features of elements in scenario evolution network and support converting deduction.The method uses dynamic Bayesian network and Markov chain Monte Carlo method to calculate the converting probability of scenario situation,which overcomes the shortcomings of static Bayesian network in dynamic scenario situation prediction.Fourthly,aiming at the problem of human intervention scenario situation evolution(called mapping evolution in this paper),a scenario situation mapping deduction method for urban cascade disasters is proposed.This method considers the human dynamic factors in the process of situation evolution,and considers that the uncertainty of information may lead to the failure of the process and result of situation deduction,which is manifested in the lack of guiding significance of deduction results.To solve this problem,based on the principle of inaccurate reduction of scenario mapping deduction,the method deduces scenario mapping deduction of cascading disasters based on mess map,and proposes a mess map analysis method for cascading disaster scenario mapping deduction.The method consists of mess map template construction,uncertainty problem discovery and factor influence chain structure,and mess map construction.Then,for cross-organizational knowledge exchange,an ontology configuration method for scenario deduction is proposed,which consists of three parts: ontology modeling,ontology interconnection and ontology integration.Based on the information support of scenario situation deduction method,the scenario ontology,network ontology and evaluation ontology of scenario scenario deduction are established,which correspond to scen ario knowledge,scenario evolution network knowledge and scenario evolution probability evaluation knowledge respectively.On the basis of ontology modeling,the ontology interconnection principle and ontology deduction method of scenario deduction are proposed to reduce the problems of ontology management and operation in the process of ontology fusion.Ontology configuration method considers the operability of scenario deduction,and it can promote knowledge flow and reduce the semantic conflict caused by cognitive differences among different organizations.It plays an important role in improving the effectiveness of scenario deduction.Finally,in terms of the missing of data,information and knowledge in applying scenario deduction,a reference framework of the application routines of two kinds of scenario deduction application(i.e.training deduction and response deduction)are designed to estabish the complete deduction application chain from data,information to knowledge and finally to deduction value.Based on 18 waterlogging training deduction cases and 15 epimedic response deduction cases,a design science approach is employed to identify the stakeholders in scenario deduction applications and the influence factors of data-based value creation and propose the data-information-knowledge-value chains.Further,the collaborative operation framework and application routine framework of deduction application are proposed from organizational and operational perspectives,respectively,to provide policy suggestions for improving deduction applications.