Printer Friendly

Risk Analysis of Emergency Based on Fuzzy Evidential Reasoning.

1. Introduction

Over the past decades, research on emergency management has gained enormous attention in both academia and practice [1-4]. Due to the complexity of emergency and its severity consequence, it is vital to investigate deeply on the emergency management involving risk analysis, development control, and evolution management [5]. There are various types of emergency, and they make emergency management even more complicated. In China, according to the overall state emergency plan, emergency mainly contains four categories: natural disasters, social security, public health, and technical accidents. Taking earthquakes, a typical natural disaster, as an example, it performs complex on its consequence [6]. The consequence usually appears as a complex disaster chain including casualties, economic loss, fires, landslides, floods, plague, and social panic. Another example is railway project accident [7], which is even harder to cope with for the reason that such accident combines industrial technology and natural situation with social environment factors together. So, many works are carried on by the lens of complexity science [8-10], and a series of positive progress is made [11-13]. Considering the complexity of the emergency and its consequence, it will benefit to emergency management if the decision maker could react as early as possible [14]. Therefore, the risk analysis of emergency is of great significance.

As a critical part of emergency management, risk analysis of emergency is actually a complicated work [15]. Accurate and stable analysis of emergency risk means great on many aspects [16], such as to reduce frequency of occurrence [17], lower the initial loss [18], and control the evolution process effectively [19]. Therefore, it is necessary to propose advanced approach to better solve problems of risk analysis. Taking China emergency management as an example, risk management with characteristics of all-phase, all-stakeholder, and all-type was not informed until 2003. Therefore, to make further progress on risk analysis research, especially on coping with the increasingly complicated risk circumstance, it is suggested to develop more flexible and adaptable analysis approach.

Some researchers notice that emergency risk usually performs poor due to the imperfect information either inadequate or loss [20,21]. To deal with such decision scenarios, evidential theory and evidential reasoning method are introduced [22]. In a basic evidential model, it focuses on solving problem with the characteristic of proposition uncertainty to set uncertainty [23]. This could be realized by contrasting the propositions and sets one by one. Take [THETA] = {[G.sub.1], [G.sub.2],..., [G.sub.n]} as a set, in which the elements are mutually exclusive and exhaustive. The set is so called identification framework. Besides, the mass (m) describes the belief of the framework, which satisfies m([empty set]) = 0, and [[summation].sub.A[subset or equal to][THETA]] m(A) = 1, wherein [empty set] signifies empty set and A stands for any subset of [THETA]. As it develops to evidential reasoning, its advantage of dealing with decision scenarios without complete information is increasingly obvious [24-29]. Therefore, a common knowledge for decision makers as well as researchers is formed that evidential reasoning could solve problems with characteristics of multiattributes.

Recent relevant works on evidential reasoning approach concentrate on the update of evidential reason rule, the implementation of this method in practice, and the integration with other methods. Among the extensive research dedicated to the evidential reasoning, the following works are very representative. Wang and Elhag [30] model the bridge risk by artificial neural network, evidential reasoning, and multiple regression analysis. After a comparison of their modelling mechanisms, they argue the remarkable advantage of evidential reasoning is its ability of modelling quantitative and qualitative data using the distributed modelling framework. Some works focus on updating the evidential rule. Yang and Xu [31] improve the evidential reasoning approach by a new rule considering evidence weights and reliabilities. Liu et al. [32] concern the hesitant fuzzy information and highlight the importance of information processing in emergency management. Some works focus on how to use evidential reasoning to solve the problems. Kong et al. [33] develop the belief rule-based inference methodology using the evidential reasoning approach and help the decision maker to predict trauma outcome. Xu et al. [34] investigate a purely data-driven work and propose a classification method using evidential reasoning. Zhang et al. [35] consider a fuzzy rule and conduct a navigational risk assessment. Some works establish novel analysis framework by integrating evidential reasoning with other approaches. Wang et al. [36] incorporate the analytic hierarchy process with evidential reasoning, by which they help contractors to select appropriate subcontractor. Evidential reasoning in this work contributes to rank the alternative subcontractors. Ng [37] takes evidential reasoning-based approach as a fast-track and objective tool and uses it to rank the available design alternatives. Working with the combination of multiattribute decision and non-life cycle assessment, it is proved that evidential reasoning benefits the evaluation of design alternatives' environmental performances. Kong et al. [38] make a combination of the principal component analysis and the evidential reasoning approach and provide a new framework to assess patient satisfaction.

To sum up, two streams of research are closely related to our work: emergency management, especially risk analysis, and approach that may contribute to risk analysis, especially evidential reasoning. The existing studies have made significant contributions to risk analysis and evidential reasoning. However, it is worth mentioning that little attention was paid to risk analysis of technical accident such as railway project accident. It may be induced by the lack of understanding for the risk analysis of such emergency. Considering the information in emergency is always uncertain and incomplete, we adopt the fuzzy evidential reasoning. Further, to better understand risk attributes, we introduce structural description framework to the fuzzy evidential reasoning. Therefore, we propose an analysis approach by incorporating the structural description framework in fuzzy evidential reasoning. To prove the validity of the model we proposed, we run a case study on railway project accident. Railway project accident is a complex system which may contain transportation accidents, public facility accidents, equipment destruction, environmental pollution, industrial hazard, and economic loss [39]. Furthermore, it is a typical technical accident and is seldom analyzed in emergency research. Therefore, our work is different from previous studies. We propose a novel approach for risk analysis by combining structural description framework with fuzzy evidential reasoning, especially when the information is imperfect either uncertain or incomplete. We firstly establish accident structural description framework, identification framework, and it helps to identify risk attributes rapidly and accurately. Secondly, we construct fuzzy belief structural model and normalize all the information we collected as evidence. Thirdly, the fuzzy evidential reasoning model is used to further processing of risk attributes. Finally, we obtain the risk analysis results, so as to provide solution for risk analysis of emergency.

The remainder of this paper is organized as follows. Section 2 establishes structural description framework for identification of risk attributes. Section 3 illustrates the model, and Section 4 takes certain railway project accident as an example to test the validity of our work. Finally, Section 5 concludes our work and offers directions for the future research.

2. Analysis of Risk Attributes

The accurate identification of emergency risk attributes is the premise of risk analysis. The core object is to identify the general attribute and basic attribute [40]. As there is no universal or uniform model for different emergencies [41], we propose a reasonable analysis framework. We determine the basic risk sources followed by their possible outcome and set them as the evidence. After that, we assign the reasoning rules and transform all the evidence to the general attribute. Till then, we accomplish emergency risk attributes analysis and lay foundation for the construction of fuzzy evidential reasoning model.

2.1. Identification of Emergency Risk Attributes. With regarding to the complicated characteristics of emergency, may be high uncertainty and severity consequence, we ought to consider both internal and external risk attributes of such complex system. To systematically illustrate emergency, it adopts a set form to establish structural description framework as Emergency = {{Emergency Type}, {Key Attribute}, {Secondary Attribute}, {Environmental Attribute}, {Hazard Attribute}} [40]. This framework significantly helps to understand the emergency and its severity as well. Looking back to the framework, it provides a simple way to identify risk attributes, covering the risk source, pattern, trend, and consequence. Therefore, this framework could also be applied to emergency risk attribute identification and correspondingly simplified as Emergency = {{Key Attribute}, {Secondary Attribute}, {Environmental Attribute}}. Among which, key attribute is the risk factor inside the emergency. To some extent, the risk grade (risk value) directly determines the severity of the initial consequences once the event occurs. The secondary attribute also stems from the inside of emergency and usually affects emergency indirectly. The environmental attribute exists outside of emergency, affects the emergency observably and measurably, but not necessarily controllably. For the convenience of further analysis, these three attributes are collectively referred as general attributes. Besides, to illustrate specific emergency risk, each general attribute could be subdivided to basic attribute. That is, the general attributes are extracted to describe the common characters, while the basic attributes are used to highlight the specific status. Then, the identification framework for emergency risk attribute is established as shown in Figure 1.

Considering the relationship among general attributes, basic attributes, and emergency risk grade, we obtain the reasoning rules as follows. If no basic attribute of certain general attribute belongs to certain risk grade, then the general attribute does not belong to this risk grade either. If all the basic attributes of certain general attribute belong to a certain risk grade, then the general attribute also belongs to this risk grade. If the basic attribute of the general attribute belongs to multiple grades, then the general attribute should be allocated to different grades under specific rules. Using the framework shown in Figure 1, we can identify the risk attributes, find the sources for risk, evaluate the occurrence probability, and predict the possible consequences. Once the risk attribute analysis framework is set up, it would support the subsequent modelling research on risk analysis of emergency.

2.2. Fuzzy Belief Model of Emergency. Considering the complicated characteristics of risk attribute, it is necessary to normalize all the risk attributes for further analysis. Usually, it takes grade to describe emergency risk. The grade should contain several standards to assess risk. The standards may cover severity of the emergency, frequency of emergency occurrence, and emergency consequence. Taking CRH brake system failure risk as an example, the risk grade is defined by grade I to grade IV. As to natural disaster risk, it is subdivided to five grades. There are also some emergency risk grades described by the qualitative method. For example, fault mode severity can be classified to insignificant, marginal, critical, and catastrophic; incident frequency can be described as seldom, little, average, usually, and always. Additionally, there are also some quantitative description methods, such as the risk value. The evidence for risk attribute is so diverse that the evidence ought to be normalized before further analysis. Combining the fuzzy set theory and the belief structure model, we establish fuzzy belief structure.

Assume emergency has N fuzzy risk grades and the adjacent two grades may intersect. Given membership functions, we symbol fuzzy numbers for each grade as [FG.sub.n] and further describe them with triangular fuzzy number or trapezoidal fuzzy number. The intersection between [FG.sub.n] and [FG.sub.n+1] is denoted as [FG.sub.n, n+1], as is shown in Figure 2. Then, the emergency fuzzy risk can be illustrated by FG = {[FG.sub.1], [FG.sub.2],..., [FG.sub.n],..., [FG.sub.N]}. Together with the belief model, we obtain the fuzzy belief model as follows:

FBS(E) = {([FG.sub.n], [[beta].sub.n]), n = 1,2,..., N}, (1)

wherein [FG.sub.n] represents fuzzy risk grade, N depicts the total number of grades, and [[beta].sub.n] denotes the belief of emergency risk grade falling on grade [FG.sub.n]. Moreover, we have [[beta].sub.n] [greater than or equal to] 0,0 [less than or equal to] [[summation].sup.N.sub.n=1] [[beta].sub.n] [less than or equal to] 1. Specifically, [[summation].sup.N.sub.n=1] [[beta].sub.n] indicates that the risk attribute information is completely known, while [[summation].sup.N.sub.n=1] [[beta].sub.n] = 0 indicates that people know nothing about risk attribute at all.

3. Model

Based on the aforementioned analysis, the evidential reasoning model would be proposed. The model would be ultimately established through the following four steps. Firstly, risk attributes are identified. By the structural description framework for emergency, we could obtain the general attribute and further get the basic attribute through experience on emergency management. Secondly, information of risk attributes is normalized. We collect the information of all the risk attributes and take them as evidence for further reasoning. Then, we set the normalization rules for evidence both qualitative and quantitative. Thirdly, the general attributes contributing to risk are calculated. The major work in this step is mainly composed of calculation on mass value and correspondingly belief for all the general attributes, as well as allocation of the intersection reliability. Finally, the emergency risk analysis results are obtained. The framework of evidential reasoning model is shown in Figure 3. As the first step has been prepared by risk attribute analysis, we would pursue the model construction from the second step.

31. Conversion of Fuzzy Belief Structure. Evidence of risk attribute has the characteristics of multifeature, multistandard, and multilength, which increases the complexity of risk analysis. Therefore, it is necessary to normalize the evidence of risk attributes before subsequent analysis. To solve this problem, we adopt fuzzy belief structure model. Given fuzzy risk grades FG = {[FG.sub.n], n = 1,2,..., N}, fuzzy belief structure is FBS(E) = {([FG.sub.n], [[beta].sub.n]),n = 1,2,..., N}. Assuming that risk attribute R has evidence with length l, its risk grade is [FG.sup.Rl] = {[], n = 1,2,..., N} and its fuzzy belief structure could be labeled as FBS(Rl) = {([], [[alpha].sub.n]), n = 1,2,..., N}. Regarding that evidence could be illustrated either qualitatively or quantitatively (even with different lengths), we set the processing rules as follows.

If the risk attribute has qualitative evidence with the same length comparing to emergency fuzzy risk grade, i.e., [N.sub.l] = N, then the evidence could be converted directly by

([FG.sup.Rl.sub.n2], [[alpha].sub.n2]) [??] ([FG.sub.n1], [[beta].sub.n1]), [n.sub.1] = [n.sub.2] [member of] N, [n.sub.2] [member of] [N.sub.l]. (2)

If the evidence is qualitative with different lengths comparing to fuzzy risk grade, i.e., N [not equal to] [N.sub.l], then we could convert it to [FG.sub.n1] by belief [[gamma].sub.n1,n2]([n.sub.1] [member of] N, [n.sub.2] [member of] [N.sub.l]). Therefore,

([FG.sup.Rl.sub.n2], [[alpha].sub.n2]) [??] ([FG.sub.n1], [[gamma].sub.n1,n2]) [n.sub.1] [member of] N, [n.sub.2] [member of] [N.sub.l], (3)

wherein 0 [less than or equal to] [[gamma].sub.n1,n2] [less than or equal to] 1 and [[summation].sup.N.sub.1=1] [[gamma].sub.n1,n2] = 1.

If the risk attribute has quantitative evidence and its risk value belongs to fuzzy risk grade [FG.sub.n] and [FG.sub.n+1] with membership [mu][FG.sub.n] and [mu][FG.sub.n+1] separately, then it converts as

[mathematical expression not reproducible]. (4)

By (2)-(4), the basic risk attribute evidence is normalized.

3.2. Fuzzy Evidential Reasoning Model. As aforementioned, we establish a structural description framework consisting of key attributes, secondary attributes, and environmental attributes; we could use this framework, i.e., R = {{KR}, {SR}, {ER}} to describe emergency risk attributes. Each general attribute could be subdivided into some certain basic attributes. From then on, we take KR as an example to construct model. Assume the number of KR's basic attributes is X and get [FBS.sup.KR.sub.x] = {([FG.sub.n], [[beta].sup.x.sub.n])}, n = 1,2,..., N, x = 1,2,..., X after fuzzy belief structure converting. Setting the corresponding weight of [KR.sub.x] as [[omega].sub.x], and [[summation].sup.X.sub.x=1] [[omega].sub.x] = 1, the mass value of each basic risk attributes could be obtained by

[m.sup.x.sub.n] = [[omega].sub.x][[beta].sup.x.sub.n], n = 1,2,..., N, x = 1,2,..., X. (5)

Additionally, there also exists some risk which could not be ascertained due to incomplete or even totally unknown information in practice, which is measured by

[m.sup.x.sub.G] = 1 - [N.summation over (n=1)] [m.sup.x.sub.n] = 1 - [N.summation over (n=1)] [[omega].sub.x][[beta].sup.x.sub.n], n = 1,2,..., N, x = 1,2,..., X. (6)

By normalization weights [[omega].sub.x] to [omega]', we can further process the evidence of KR. The risk evaluation result of KR is FBS(KR) = {(FG.sub.n], [[beta].sup.KR.sub.n])}, n = 1,2,..., N, x = 1,2,..., X. Therefore, we obtain the mass value of KR by using the following equation:

[m.sup.KR.sub.n] = [omega]' {[X.PI over (x=1)] [m.sup.x.sub.n] + [m.sup.x.sub.G]] - [X.PI over (x=1)] [m.sup.x.sub.G]}, n = 1,2,..., N. (7)

The mass value of the intersection between two adjacent risk grades is

[mathematical expression not reproducible]. (8)

The mass value of uncertain risk grades due to imperfect information is

[mathematical expression not reproducible]. (9)

Specially, [omega]' is determined by [[summation].sup.N.sub.n=1] m([G.sub.n]) + [[summation].sup.N-1.sub.n=1] [G.sub.n,n+1] = 1.

Belief for each mass value is

[[beta].sup.KR.sub.n] = [m.sup.KR.sub.n]/1 - [[bar.m].sup.KR.sub.G], n = 1,2,..., N, (10)

[[beta].sup.KR.sub.n,n+1] = [m.sup.KR.sub.n,n+1]/1 - [[bar.m].sup.KR.sub.G], n = 1,2,..., N - 1. (11)

We obtain the primary analysis result by {([FG.sup.KR.sub.n], [[beta].sup.KR.sub.n]), ([FG.sup.KR.sub.n,n+1], [[beta].sup.KR.sub.n,n+1]), ([FG.sup.KR.sub.G], [[beta].sup.KR.sub.G])}. However, the intersection between two grades does not exist in reality. Therefore, it still needs further analysis on ([FG.sup.KR.sub.n,n+1], [[beta].sup.KR.sub.n,n+1]).

3.3. Allocation of Fuzzy Intersection Belief. Considering the risk grade is regulated clear in reality, it is necessary to process the intersection between adjacent risk grades. To insure the analysis is reasonable, we set the reasoning rules as follows. If the intersection [FG.sub.n,n+1] totally belongs to [FG.sub.n] or [FG.sub.n+1], then its belief also belongs to [FG.sub.n] or [FG.sub.n+1]. If the intersection [FG.sub.n,n+1] varies from [FG.sub.n] to [FG.sub.n+1], as shown in Figure 4, and the endpoints are set by [FG.sub.n] and [FG.sub.n+1], then we allocate ([m.sup.KA.sub.n,n+1], [[beta].sup.KA.sub.n,n+1]) within this interval.

As it is shown in Figure 4, the proportion of grade [FG.sub.n] in [S.sub.n,n+1] is [[delta].sup.KR'.sub.n], while the proportion of grade [FG.sub.n+1] is [[delta].sup.KR'.sub.n+1]. [[delta].sup.KR'.sub.n] + [[delta].sup.KR'.sub.n+1] = 1, and they are determined by (12) and (13) separately:

[[delta].sup.KR'.sub.n] = 1/2[(1 - [d.sub.n]/[d.sub.n] + [d.sub.n+1]) + [S.sub.n]/[S.sub.n] + [S.sub.n+1]], (12)

[[delta].sup.KR'.sub.n+1] = 1/2[(1 - [d.sub.n+1]/[d.sub.n] + [d.sub.n+1]) + [S.sub.n]/[S.sub.n] + [S.sub.n+1]]. (13)

Therefore, the belief allocated to grade [FG.sub.n] is

[[beta].sup.KR'.sub.n] = [S.sub.n] + [[delta].sup.KR'.sub.n][S.sub.n,n+1]/[S.sub.n] + [S.sub.n,n+1] + [S.sub.n+1]. (14)

However, the belief allocated to grade [FG.sub.n+1] is

[[beta].sup.KR'.sub.n] = [S.sub.n+1] + [[delta].sup.KR'.sub.n+1][S.sub.n,n+1]/[S.sub.n] + [S.sub.n,n+1] + [S.sub.n+1]. (15)

Then, combining with {([m.sup.KR.sub.n], [[beta].sup.KR.sub.n]), ([m.sup.KR.sub.n,n+1], [[beta].sup.KR.sub.n,n+1]), ([m.sup.KR.sub.G], [[beta].sup.KR.sub.G])}, we get the initial risk analysis result of accident as

{([m.sup.KA.sub.n], [[beta].sup.KA.sub.n] + [[beta].sup.KA'.sub.n]), ([m.sup.KA.sub.G], [[beta].sup.KA.sub.G])}. (16)

Since the key attribute, secondary attribute, and environmental attribute have been differentiated when we identify general risk attribute, it is not necessary to rank the risk attributes like risk analysis paradigm in general.

3.4. Analysis of Fuzzy Evidential Reasoning Results. To understand the risk grade more intuitively, we take a further analysis. It also benefits emergency management, such as release warning signals and formulate emergency response in different stages of the emergency. To solve this problem, we use the expected risk value. As the risk grade and its belief could be obtained by (16), we determine the grade risk values as follows. Assuming the risk value of the highest grade to 1 and the lowest grade to 0, the n - 2 grades among them could be determined by specific accident. The analysis result would be finally obtained by the summation of all the risk values:

[U.sup.KA] = [N.summation over (n=1)] [[beta].sub.n][r.sub.n]. (17)

Based on the analysis of the key attribute, we can analyze the risk of the secondary attribute and environmental attribute as the same way proposed by (1)-(17). With regarding to the expected risk value of KR, SR, and ER, it would exactly help to analyze accident risk. That is, given the information of risk attributes, the fuzzy evidential reasoning model would make sense on estimating to what extent the accident would happen. Besides, this approach also helps to evaluate the severity of the accident consequence.

4. Model Analysis on Case Study

Here, we take railway project accident as an example. Railway project accident is a typical technical accident and seldom researched in previous works. We determine the information and evidence of risk attribute mainly from the following two documents: Technical Code for Risk Management of Railway Construction Engineering (issued by China Railway in 2014), Statistics of Construction Risk (a series of documents declared by certain railway project in 2018).

4.1. Model Solution

4.1.1. Step 1. Using the emergency risk attribute identification framework, we determine the general attributes and their basic attributes. Then, calculate the risk grade, weight, and value of all attributes. Finally, we obtain 21 basic attributes, and their information is summarized as shown in Table 1.

4.1.2. Step 2. Considering the specific condition of railway project in China, we measure the risk grade of railway project accident by {Catastrophic, critical, general, Marginal, insignificant}, which simplifies as FG = {[FG.sub.1], [FG.sub.2], [FG.sub.3], [FG.sub.4], [FG.sub.5]}. With regard to Figure 2, finally, we determine the fuzzy risk grade for railway project accident as in Figure 5.

4.1.3. Step 3. We take further process of basic risk attributes. To simplify but without loss of generality, we omit the process of getting the ultimate evidence from multi-information. Calculation starts from cut slope height, whose risk value is 2; triangular fuzzy number for [FG.sub.1] is (0, 0, 4) and for [FG.sub.2] is (0, 4, 8). Then, we obtain fuzzy belief of cut slope height as {([FG.sub.1], 0.5), ([FG.sub.2], 0.5), ([FG.sub.3], 0), ([FG.sub.4], 0), ([FG.sub.5], 0)}, take (0.5, 0.5, 0, 0, 0) for short. All the basic attribute analysis results are demonstrated in the third column in Table 2. Then, we get the mass value by (5)-(7), and the result is shown in the fourth column in Table 2. Based on equations (8)-(11), we get the primary result on railway project accident risk analysis and it is shown in the fifth column in Table 2. Note that there is still an intersection between the adjacent grades; therefore, we pursue to allocate the intersection by the following steps.

4.1.4. Step 4. By equation (12)-(16), we allocate the intersection belief to the two adjacent grades and get the result on risk evaluation considering risk grade and its belief. As mentioned before, we further calculate the expected risk value for each basic attribute by using equation (17). Finally, we conclude the results of this step and list them in Table 3.

4.2. Results Analysis. We could draw some managerial insight out of the above results. One interesting point is that the risk value of the key attribute is not significantly high comparing with the traditional recognition. We explain this counterintuitive result from the perspective of practice. In reality, people tend to spend enormous money and high effort on this aspect, and it leads to technology progress and effective regulation. Therefore, although the key attribute is vital to railway project accident, the failure rate and risk value is low. It is in line with the development status of China railway project.

Secondary attribute performs relatively average in risk grade and belief, which is basically consistent with the actual situation. Unexpectedly, the belief allocated for the environmental attribute is relatively high. It maybe due to the geological condition which is important to railway project and its operation. As it is shown in the statistics data of railway project accident, major economic losses and casualties are usually caused by environmental attribute.

To some extent, these findings will push managers to rethink the role of traditional experience recognition and related training. Moreover, we also provide reference for risk prevention, early warning, and response of railway project accident. Especially, it would obviously benefit to the railway company with capital constraints.

5. Conclusions

Emergency risk analysis is vital to emergency management; especially, it plays an important role on learning emergency operational mechanism. Aiming at providing a stable approach to analyze emergency risk, this paper proposes an analytical approach incorporating structural description framework and fuzzy evidential reasoning. It starts from the establishment of risk attribute identification, which is taking emergency structural description framework as the foundation. Then, we assess the risk grade of risk attributes by evidential reasoning. By a case study on railway project accident, it is proved that the approach outperforms comparing to current risk management. It could provide theoretical support for the risk prevention and decision-making involving in each stage.

To conclude, our study proposes a novel approach for accident risk analysis by combining structural description framework and fuzzy evidential reasoning, especially when the information is imperfect either uncertain or incomplete. We firstly establish accident structural description framework, identification framework; it helps to recognize risk attributes rapidly and accurately. Secondly, we construct fuzzy belief structural model and normalize all the information. Thirdly, the fuzzy evidential reasoning model is used to combine basic attributes. Finally, we obtain the analysis results, so as to assist the decision-making of emergency risk prevention or response.

The limitations of our work are also obvious due to the model construction. Firstly, the evidential reasoning rule we used here is very basic. It is in premise of the implicit assumption that the fuzzy risk grade is defined clearly. However, the grade may be obscure in reality, and it may induce the fact that the evidence belongs to three or more different grades. Secondly, the basic attributes of our case study are relatively small. We obtain such attributes from deep analysis of daily records of specific railway project and statistics of construction risk, and they are undoubtedly helpful to make risk analysis. Considering the analysis result, especially the analysis of key attribute, it still needs more detailed works on the basic attributes to highlight the critical of the key attribute. Finally, our work merely focuses on the occurrence risk of emergency, indeed risk is always developing. So, we have to further alleviate the difficulty of evidential reasoning application or do the risk analysis taking emergency development or evolution under consideration.

Regarding above limitations and the fact that risk management is the basic of emergency development management and evolution research, it needs further study. As to future work, the research could enrich the evidential reasoning rule by new tools of data mining and expand the attributes' scope from occurrence risk to development risk and evolution risk. Further, the research could enrich the works on risk prevention and early warning. In addition, future work may introduce various constraints to close the reality, especially the resource constraints. After that, it could finally provide great references for relevant managers and help to make scientific and efficient decisions.

Data Availability

All the data generated or analyzed during this study are included in this paper.

Conflicts of Interest

The authors declare that there are no conflicts of interest regarding the publication of this paper.


The first author acknowledges support from the Humanities and Social Science Foundation of Ministry of Education of China (grant no. 17YJC630113) and NSFC (grant nos. 71572115 and 91646117). The second author acknowledges support from NSFC (grant nos. 71702021 and 71802143) and the Fundamental Research Funds for the Central Universities (DUT18RC(4)021).


[1] N. Altay and W. G. Green, "OR/MS research in disaster operations management," European Journal of Operational Research, vol. 175, no. 1, pp. 475-493, 2006.

[2] G. Green and R. Batta, "Review of recent developments in OR/ MS research in disaster operations management," European Journal of Operational Research, vol. 230, no. 2, pp. 201-211, 2013.

[3] L. Zhou, X. Wu, Z. Xu, and H. Fujita, "Emergency decision making for natural disasters: an overview," International Journal of Disaster Risk Reduction, vol. 27, pp. 567-576, 2018.

[4] N. Argyris and S. French, "Nuclear emergency decision support: a behavioural OR perspective," European Journal of Operational Research, vol. 262, no. 1, pp. 180-193, 2017.

[5] S. Thorvaldsdottir and R. Sigbjornsson, "Disaster-function management: basic principles," Natural Hazards Review, vol. 15, no. 1, pp. 48-57, 2014.

[6] Y. Zhang, W. G. Weng, and Z. L. Huang, "A scenario-based model for earthquake emergency management effectiveness evaluation," Technological Forecasting and Social Change, vol. 128, pp. 197-207, 2018.

[7] C. W. Yang, Z. H. Li, X. Y. Guo, W. Y. Yu, J. Jin, and L. Zhu, "Application of BP neural network model in risk evaluation of railway construction," Complexity, vol. 2019, Article ID 2946158, 12 pages, 2019.

[8] M.-S. Chang, Y.-L. Tseng, and J.-W. Chen, "A scenario planning approach for the flood emergency logistics preparation problem under uncertainty," Transportation Research Part E: Logistics and Transportation Review, vol. 43, no. 6, pp. 737-754, 2007.

[9] J. Ren, Y. Yusuf, and N. Burns, "A decision-support framework for agile enterprise partnering," The International Journal of Advanced Manufacturing Technology, vol. 41, no. 12, pp. 180-192, 2009.

[10] V. A. Banuls, M. Turoff, and S. R. Hiltz, "Collaborative scenario modeling in emergency management through cross-impact," Technological Forecasting and Social Change, vol. 80, no. 9, pp. 1756-1774, 2013.

[11] S. Luna and M. J. Pennock, "Social media applications and emergency management: a literature review and research agenda," International Journal of Disaster Risk Reduction, vol. 28, pp. 565-577, 2018.

[12] T. B. Vu and I. Noy, "Regional effects of natural disasters in China: investing in post-disaster recovery," Natural Hazards, vol. 75, no. S2, pp. 111-126, 2015.

[13] M. Kumasaki, M. King, M. Arai, and L. Yang, "Anatomy of cascading natural disasters in Japan: main modes and linkages," Natural Hazards, vol. 80, no. 3, pp. 1425-1441, 2016.

[14] Z. Y. Wang, X. D. Liu, and S. T. Zhang, "A new decision method for public opinion crisis with the intervention of risk perception of the public," Complexity, vol. 2019, Article ID 9527218, 14 pages, 2019.

[15] S. D. Guikema, "Natural disaster risk analysis for critical infrastructure systems: an approach based on statistical learning theory," Reliability Engineering & System Safety, vol. 94, no. 4, pp. 855-860, 2009.

[16] T. Christensen, O. Andreas Danielsen, P. Laegreid, and L. H. Rykkja, "Comparing coordination structures for crisis management in six countries," Public Administration, vol. 94, no. 2, pp. 316-332, 2016.

[17] X. Lu and L. Xue, "Managing the unexpected: sense-making in the Chinese emergency management system," Public Administration, vol. 94, no. 2, pp. 414-429, 2016.

[18] M. A. Cole, R. J. R. Elliott, T. Okubo, and E. Strobl, "Predisaster planning and post-disaster aid: examining the impact of the great East Japan earthquake," International Journal of Disaster Risk Reduction, vol. 21, pp. 291-302, 2017.

[19] P. J. Ren, Z. S. Xu, and J. Gu, "Assessments of the effectiveness of an earthquake emergency plan implementation with hesitant analytic hierarchy process," International Journal of Information Technology & Decision Making, vol. 15, no. 6, pp. 1367-1389, 2016.

[20] S. Gupta, M. K. Starr, R. Z. Farahani, and N. Matinrad, "Disaster management from a POM perspective: mapping a new domain," Production and Operations Management, vol. 25, no. 10, pp. 1611-1637, 2016.

[21] M. Setbon, J. Raude, C. Fischler, and A. Flahault, "Risk perception of the "mad cow disease" in France: determinants and consequences," Risk Analysis, vol. 25, no. 4, pp. 813-826, 2005.

[22] J.-B. Yang, "Rule and utility based evidential reasoning approach for multiattribute decision analysis under uncertainties," European Journal of Operational Research, vol. 131, no. 1, pp. 31-61, 2001.

[23] J. B. Yang and D. L. Xu, "On the evidential reasoning algorithm for multiple attribute decision analysis under uncertainty," IEEE Transactions on Systems, Man, and Cybernetics--Part A: Systems and Humans, vol. 32, no. 3, pp. 289-304, 2002.

[24] J. B. Yang, Y. M. Wang, D. L. Xu, and K. S. Chin, "The evidential reasoning approach for MADA under both probabilistic and fuzzy uncertainties," European Journal of Operational Research, vol. 171, no. 1, pp. 309-343, 2006.

[25] D.-L. Xu, "An introduction and survey of the evidential reasoning approach for multiple criteria decision analysis," Annals of Operations Research, vol. 195, no. 1, pp. 163-187, 2012.

[26] M. Sonmez, J. B. Yang, and G. D. Holt, "Addressing the contractor selection problem using an evidential reasoning approach," Engineering Construction and Architectural Management, vol. 8, no. 3, pp. 198-210, 2001.

[27] Y.-M. Wang, J.-B. Yang, and D.-L. Xu, "Environmental impact assessment using the evidential reasoning approach," European Journal of Operational Research, vol. 174, no. 3, pp. 1885-1913, 2006.

[28] M. Zhou, X.-B. Liu, J.-B. Yang, and C. Fang, "Group evidential reasoning approach for MADA under fuzziness and uncertainties," International Journal of Computational Intelligence Systems, vol. 6, no. 3, pp. 423-441, 2013.

[29] A. Akhoundi and S. Nazif, "Sustainability assessment of wastewater reuse alternatives using the evidential reasoning approach," Journal of Cleaner Production, vol. 195, no. 10, pp. 1350-1376, 2018.

[30] Y. M. Wang and T. Elhag, "A comparison of neural network, evidential reasoning and multiple regression analysis in modelling bridge risks," Expert Systems with Applications, vol. 32, no. 3, pp. 336-348, 2007.

[31] J.-B. Yang and D.-L. Xu, "Evidential reasoning rule for evidence combination," Artificial Intelligence, vol. 205, pp. 1-29, 2013.

[32] X. Liu, Z. Wang, and S. Zhang, "A new methodology for hesitant fuzzy emergency decision making with unknown weight information," Complexity, vol. 2018, Article ID 5145348, 12 pages, 2018.

[33] G. Kong, D. L. Xu, J. B. Yang et al., "Belief rule-based inference for predicting trauma outcome," Knowledge-Based Systems, vol. 95, pp. 35-44, 2016.

[34] X. Xu, J. Zheng, J. B. Yang, D. L. Xu, and Y. W. Chen, "Data classification using evidence reasoning rule," Knowledge-Based Systems, vol. 116, pp. 144-151, 2017.

[35] D. Zhang, X. Yan, J. Zhang, Z. Yang, and J. Wang, "Use of fuzzy rule-based evidential reasoning approach in the navigational risk assessment of inland waterway transportation systems," Safety Science, vol. 82, pp. 352-360, 2016.

[36] G. Wang, F. Cetindere, A. Damci, and B. N. Bingol, "Smart home subcontractor selection using the integration of AHP and evidential reasoning approaches," Procedia Engineering, vol. 164, pp. 347-353, 2016.

[37] C. Y. Ng, "Evidential reasoning-based Fuzzy AHP approach for the evaluation of design alternatives' environmental performances," Applied Soft Computing, vol. 46, pp. 381-397, 2016.

[38] G. L. Kong, L. L. Jiang, X. F. Yin et al., "Combining principal component analysis and the evidential reasoning approach for healthcare quality assessment," Annals of Operations Research, vol. 271, no. 2, pp. 679-699, 2018.

[39] Y. Lei, C. Huang, and Y. Wu, "Operational risk assessment for international transport corridor: a case study of China-Pakistan economic corridor," Discrete Dynamics in Nature and Society, vol. 2019, Article ID 5730746, 7 pages, 2019.

[40] Y. J. Li, X. J. Qiao, and X. C. Sun, "Study on emergency structural description framework," Journal of University of Electronic Science and Technology of China (Social Sciences Edition), vol. 5, no. 1, pp. 28-33, 2013.

[41] X. Xu, X. Yin, and X. Chen, "A large-group emergency risk decision method based on data mining of public attribute preferences," Knowledge-Based Systems, vol. 163, pp. 495-509, 2019.

Xiaojiao Qiao [ID] (1) and Dan Shi [ID] (2)

(1) School of Management, Tianjin University of Technology, Tianjin, China

(2) School of Business, Dalian University of Technology, Panjin, LiaoNing Province, China

Correspondence should be addressed to Dan Shi;

Received 26 July 2019; Revised 12 October 2019; Accepted 1 November 2019; Published 22 November 2019

Guest Editor: Baltazar Aguirre-Hernandez

Caption: Figure 1: Emergency risk attribute identification framework.

Caption: Figure 2: Fuzzy risk grade.

Caption: Figure 3: Framework of fuzzy evidential reasoning model.

Caption: Figure 4: Fuzzy intersection reliability allocation.

Caption: Figure 5: Fuzzy risk grade of railway project accident.
Table 1: Railway project accident risk analysis and evident definition.

General attribute                  Basic attribute

                                   Cut slope height
                                  Bank slope height
                           Special construction conditions
Key attribute             Special construction technologies
                                 Deep foundation pit
                               Existing railway effects

                                   Project schedule
                                    Function loss
                                  Project investment
                                 Environment protect
Secondary attribute        Social status along the project
                            Locomotive depot configuration
                                Vehicle configuration
                                  Terminal location
                                  Project interface
                              Infrastructure relocation
                                    Residual risk

                                  Weather condition
Environmental attribute           Underground water
                                 Geological condition
                                    Stuff quality

General attribute                           Risk consequence

                              Slide slump, blocks peeling off, roadbed
                                Embankment settlement, slope collapse
                               Casualty, economic loss, postproject
Key attribute              Project failure, economic loss, derivative
                                  project failure, postproject risks
                           Structural design failure, settlement
                                  deformation, unexpected accidents
                            Technological bottleneck, roadbed
                                   deterioration, schedule delay

                             Go over budget, alleviate function,
                                      interface project delay
                                         Below the market demand
                                Project abandon, poor quality, social
Secondary attribute           Destroy environment like cultural relics
                              and natural habitats. Social conflict
                                           and welfare
                             Social conflict, go over budget, schedule
                              Operational risk, social investigation
                                           Operational risk
                                 Project abandon, infrastructure
                                  Project isolation, project failure
                                Go over budget, schedule delay, group
                                        Unexpected loss

                                   Project duration, operational risk
Environmental attribute     Settlement, roadbed deformation, water
                            Roadbed risk, social conflict, project
                              abandon, karst water and mud inrush
                                         disaster, bombing
                               Casualty, operational risk, project

General attribute            Risk       Risk    Risk
                            grade      weight   value

                          [FG.sub.1]    0.2       2
                          [FG.sub.3]    0.1       9
                          [FG.sub.2]    0.2       6
Key attribute             [FG.sub.1]    0.2       1
                          [FG.sub.2]    0.2       7
                          [FG.sub.4]    0.1      16

                          [FG.sub.4]    0.1      15
                          [FG.sub.3]    0.05      9
                          [FG.sub.4]    0.1      11
Secondary attribute       [FG.sub.1]    0.15      1
                          [FG.sub.5]    0.05     17
                          [FG.sub.4]    0.05     14
                          [FG.sub.5]    0.05     18
                          [FG.sub.3]    0.05     14
                          [FG.sub.3]    0.2       9
                          [FG.sub.2]    0.15      6
                          [FG.sub.3]    0.05     --

                          [FG.sub.3]    0.15      9
Environmental attribute   [FG.sub.5]    0.25     18
                          [FG.sub.3]    0.25      7
                          [FG.sub.2]    0.35      3

Table 2: Analysis results of risk attributes (I).

General attribute           Basic attribute

                            Cut slope height
                           Bank slope height
                    Special construction conditions
Key attribute             Special construction
                          Deep foundation pit
                        Existing railway effects
                            Project schedule
                             Function loss
                           Project investment

                          Environment protect
                        Social status along the
Secondary                       project
attribute                   Locomotive depot
                         Vehicle configuration
                           Terminal location
                           Project interface
                       Infrastructure relocation
                             Residual risk

                           Weather condition
Environmental               Underground water
attribute                 Geological condition
                              Staff quality

General attribute            Belief                       Mass

                       (0.5, 0.5, 0, 0, 0)         (0.1, 0.1, 0, 0, 0)
                         (0, 0, 1, 0, 0)            (0, 0, 0.1, 0, 0)
                       (0, 0.5, 0.5, 0, 0)         (0, 0.1, 0.1, 0, 0)
Key attribute        (0.3333, 0.6667, 0, 0,        (0.0666, 0.1333, 0,
                               0)                         0, 0)
                     (0.875, 0.125, 0, 0, 0)      (0.1750,0.025,0,0, 0)
                         (0.0, 0, 1, 0)             (0.0, 0, 0.1, 0)

                      (0, 0, 0,0.625,0.375)         (0, 0, 0, 0.0625,
                         (0, 0, 1, 0, 0)           (0, 0, 0.05, 0, 0)
                     (0, 0, 0.1667, 0.8333,          (0, 0, 0.0167,
                               0)                      0.0833, 0)
                       (0.5, 0.5, 0, 0, 0)        (0.075, 0.075, 0, 0,
                     (0, 0, 0, 0.875, 0.125)        (0, 0, 0, 0.0438,
Secondary                                                0.0625)
attribute               (0.0, 0, 0.6667,            (0, 0, 0, 0.0333,
                             0.3333)                     0.0167)
                        (0.0, 0, 0.8889,            (0, 0, 0, 0.0444,
                             0.1111)                     0.0056)
                     (0, 0, 0.4286, 0.5714,         (0, 0, 0, 0.0214,
                               0)                        0.0286)
                         (0, 0, 1, 0, 0)            (0, 0, 0.2, 0, 0)
                     (0, 0.8571, 0.1429, 0,        (0, 0.1286, 0.0214,
                              0, 0)                       0, 0)
                         (0, 0, 0, 0, 0)             (0, 0, 0, 0, 0)

                         (0, 0, 1, 0, 0)           (0, 0, 0.15, 0, 0)
Environmental           (0, 0, 0, 0.3333,           (0, 0, 0, 0.0833,
attribute                    0.6667)                     0.1667)
                      (0, 0.25, 0.75, 0, 0)        (0, 0.0625, 0.1875,
                                                          0, 0)
                    (0.8571, 0.1429, 0, 0, 0)     (0.3, 0.05, 0, 0, 0)

General attribute                     Primary result

                          {([FG.sub.1], 0.0241),
Key attribute             ([FG.sub.1,2], 0.0136), ([FG.sub.2], 0.1906),
                          ([FG.sub.2,3], 0.1742), ([FG.sub.3], 0.1814),
                          ([FG.sub.3,4], 0.0258), ([FG.sub.4], 0.048),
                          ([FG.sub.4,5], 0.0127), ([FG.sub.5, 0)}

Secondary attribute       {([FG.sub.1], 0.1002),
                          ([FG.sub.1,2], 0.0106), ([FG.sub.2], 0.1057),
                          ([FG.sub.2,3], 0.059), ([FG.sub.3], 0.1067),
                          ([FG.sub.3,4], 0.0871), ([FG.sub.4], 0.1323),
                          ([FG.sub.4,5], 0.0144), ([FG.sub.5], 0.1632)}

                          {([FG.sub.1], 0.1123),
Environmental attribute   ([FG.sub.1,2], 0.0344), ([FG.sub.2], 0.1031),
                          ([FG.sub.2,3], 0.0357), ([FG.sub.3], 0.1708),
                          ([FG.sub.3,4], 0.0258), ([FG.sub.4], 0.0746),
                          ([FG.sub.4,5], 0.0322), ([FG.sub.5], 0.0096)}

Table 3: Analysis results of risk attributes (II).

                          [FG.sub.1]   [FG.sub.2]   [FG.sub.3]

Key attribute               0.0319       0.2805       0.2764
Secondary attribute         0.1057       0.1395       0.1787
Environmental attribute     0.1295       0.1392       0.2015

                          [FG.sub.4]   [FG.sub.5]   Expected risk value

Key attribute               0.0672       0.0064           0.5742
Secondary attribute         0.1831       0.1704           0.5649
Environmental attribute     0.1036       0.0257           0.4151
COPYRIGHT 2019 Hindawi Limited
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2019 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Research Article
Author:Qiao, Xiaojiao; Shi, Dan
Date:Dec 31, 2019
Previous Article:Master-Slave Composite Vibration Control of a Mobile Flexible Manipulator via Synchronization Optimization of Observation and Feedback.
Next Article:Chaos in Duopoly Games via Furstenberg Family Couple.

Terms of use | Privacy policy | Copyright © 2022 Farlex, Inc. | Feedback | For webmasters |