CN101359347A - Railway emergency plan modelling approach based on stochastic Petri net - Google Patents

Abstract

The invention discloses a modeling method of the railway emergency plan based on the random Petri net, which relates to the railway emergency plan processing technical field; the technical proposal includes that the type and rank of the current railway emergency plan are analyzed; the emergency plan is classified and prepared to realize the template of the emergency plan content; the common contents in various accident rescues in the emergency plan are extracted; the emergency plan is classified into flows through the preparation of the common contents; the emergency plan flows are processed with qualitative analysis and quantitative analysis based on the Petri net theory, so as to establish the Petri models for the classified emergency flows; a plurality of single models are combined to an integral model through the interrelation of the models; the integral model is processed with performance equivalent simplification; finally, the simplified simple models are combined to obtain the SPN model with integral simplification. The SPN model established by the modeling method is adopted to realize the systematical, digital and intelligent management of the railway emergency plan.

Description

Modeling method based on the railway emergency plan of stochastic Petri net

Technical field

The invention belongs to the railway emergency plan processing technology field, especially a kind of modeling method of the railway emergency plan based on stochastic Petri net.

Background technology

There are problems in China railways emergency preplan system construction at present, and show: to emergency preplan, especially the attention degree of manipulation type prediction scheme is generally not enough, work out perfect inadequately, execution is strong inadequately; The technical support means of emergency preplan are very backward, and existing technology platform has become the bottleneck of the prediction scheme generation effectiveness of having worked out in a large number.

And in the existing railway emergency plan management system, the emergency preplan that has comprised a large amount of variety classeses, grade, the storage mode of these prediction scheme data and file is backward simultaneously, mode of management is unreasonable, searches search inconvenience, revise to upgrade being not easy, and resource sharing is poor.And utilize stochastic Petri net (Stochastic Petri Net:SPN) theory to carry out modeling for the railway emergency plan flow process, can improve the availability of data and information, and then realize systematization, digitizing and the intellectuality of emergency preplan, reach the purpose of optimum management.Such research method still belongs to blank in China's Related Research Domain.

Summary of the invention

The objective of the invention is to, of a great variety at the emergency preplan that exists in the railway emergency plan management system, grade is complicated, mode of management is unreasonable, problems such as retrieval inconvenience and resource sharing difference, proposes a kind of method that realizes emergency preplan systematization, digitizing, intelligent management.

Technical scheme of the present invention is: a kind of modeling method of the railway emergency plan based on stochastic Petri net is characterized in that said method comprising the steps of:

The type and the grade of step 1, the existing railway emergency plan of analysis are divided prediction scheme and arrangement, realize the templating of emergency preplan content;

Step 2, extract the general character content in all kinds of accident rescues in overall prediction scheme, comprehensive prediction scheme and the station section prediction scheme, and prediction scheme is divided into flow process on the basis that templating handles in that the emergency preplan content is carried out by putting these general character contents in order;

Step 3, based on the stochastic Petri net theory, the prediction scheme flow process is carried out qualitative analysis and quantitative test, thereby sets up the Petri model for respectively the prediction scheme flow process of dividing in the step 2;

The Petri model of setting up in step 4, the integration step 3 by the contact between the model, connects into a block mold with numerous single models;

Step 5, the block mold that step 4 is set up carry out performance abbreviation of equal value;

After step 6, the abbreviation, thereby naive model connected the SPN model obtain overall simplification.

In the described step 1, the templating of emergency preplan content is based on the existing emergency preplan template of railway, and at dissimilar prediction schemes, similar stencil design goes out different concrete template frameworks.

In the described step 3, qualitative analysis is to utilize the stochastic Petri net theory, uses accessibility, boundedness and the security of T_ invariant and Markov process analysis model.

In the described step 3, quantitative test be by Markov process transfer rate matrix and fuzzy mathematics theory solved the storehouse utilization factor and system's average delay time of busy probability, storehouse institute idle probability, system's transition.

In the step 5, it is to four kinds of the most basic structures in the Petri net that the block mold that step 4 is set up carries out performance abbreviation of equal value---in proper order, concurrent, selection and circulation carry out time performance abbreviation of equal value; Its method is that each basic structure is regarded as a subnet, represents with equivalent time transition respectively, and this timed transition has the time expectation that equates with original subnet, thereby obtains the SPN model of a simplification.

Effect of the present invention is: in conjunction with the characteristics of railway emergency plan procedure, utilize the modeling method of stochastic Petri net, set up the stochastic Petri pessimistic concurrency control from the angle of information processing for the railway emergency plan flow process, realized railway emergency plan systematization, digitizing, intellectualized management; Simultaneously, use the stochastic Petri pessimistic concurrency control can analyze data, this decision support for the train operation decision maker provides foundation, thereby makes the railway emergency plan system have more science.

Description of drawings

Fig. 1 is the modeling procedure synoptic diagram of railway emergency plan.

Fig. 2 refines general character content synoptic diagram on prediction scheme early warning template basis.

Fig. 3 utilizes the theoretical stochastic Petri network model synoptic diagram of setting up of stochastic Petri net.

Fig. 4 simplifies the sequential organization synoptic diagram.

Fig. 5 simplifies concurrent structural representation.

Fig. 6 simplifies the choice structure synoptic diagram.

Fig. 7 simplifies the loop structure synoptic diagram.

Embodiment

Below in conjunction with accompanying drawing and embodiment the present invention is described in further detail.

With reference to the modeling schematic flow sheet of the railway emergency plan shown in the accompanying drawing 1, the modeling flow process of railway emergency plan is specific as follows:

Step 101, by investigation and analysis on the spot, existing railway emergency plan has polytype;

Step 102, through abstract and standardization processing prediction scheme being divided and put in order is several large forms;

Step 103 is respectively that prevention early warning template, information report and submit template, response classification template, organizational structure's template, emergency guarantee template, later stage to dispose template etc.;

Step 104, on the cope match-plate pattern basis, with these templates be subdivided into comprehensively, two types of the sections of station; With the Qinghai-Tibet Railway line is example, and the comprehensive pre-warning template of prevention early warning template comprises " Qinghai-Tibet Railway company off-gauge goods shipping accident emergency preplan " again, " Ge La of Qinghai-Tibet Railway company section strong wind weather safe train operation emergency preplan ", " Ge La of Qinghai-Tibet Railway company section snowfall weather safe train operation emergency preplan ", " Qinghai-Tibet Railway company railway (road) traffic injuries and deaths accident emergency prediction scheme ", " Qinghai-Tibet Railway company public health emergency emergency preplan ", " Qinghai-Tibet Railway company burst passenger flow, the late emergency preplan of passenger train ", " Qinghai-Tibet Railway company Dangerous Goods Transport accident emergency prediction scheme ", " Qinghai-Tibet Railway network and information security accident emergency prediction scheme ", " railway geologic hazard emergency preplan ".The section of station early warning template comprises " the overall emergency preplan of Qinghai-Tibet Railway public security bureau Emergent Public Events ", " the overall emergency preplan of station, Xining Emergent Public Events ".On the basis of these type templates, extract general character content in all kinds of accident rescues;

Step 105, and arrangement general character content will rescue prediction scheme and be divided into several big flow processs,

Step 106 is respectively that flow process, rescue later stage disposal flow process etc. are disposed in early warning flow process, expert decision-making flow process, emergency management and rescue and rescue; Utilize the correlation theory of stochastic Petri net, the accessibility of use T_ invariant and Markovian process qualitative analysis model, boundedness, security etc., the correlation theory of use Markovian process transfer rate matrix and fuzzy mathematics; The quantitative test storehouse utilization factor and system's average delay time of busy probability, storehouse institute idle probability, system's transition, thereby set up the Petri model for respectively these flow processs;

Step 107 is integrated these models, by the contact between the model, numerous single models is connected into a block mold;

Step 108, utilize the four kind basic structures of time performance equivalence to Petri net---in proper order, selection, concurrently simplify, be connected with circulation, thus acquisition overall simplification model;

Step 109, final mask is set up successfully.

Shown in accompanying drawing 2, on prediction scheme early warning template basis, refine general character content synoptic diagram, the general character content that extracts on prediction scheme early warning template basis is an example with the Qinghai-Tibet Railway in the above-mentioned steps 104, comprises step: 201, have an accident; 202, accident is received a crime report; 203, alert reports and passes on, and 204, simultaneously will be in station intersegmental part issue accident information, 205, the segment base layer early warning of taking to stand accordingly response.

Shown in Figure 3 is to utilize the theoretical stochastic Petri network model synoptic diagram of setting up of stochastic Petri net.106 set up the Petri model for single flow process set by step; And 107 integrate these models set by step, by the contact between the model, numerous single models are connected into a block mold.Below illustrate method for establishing model: shown in Fig. 3 the early warning flow process, the expert decision-making flow process, emergency management and rescue are disposed flow process with rescue, and the rescue later stage is disposed the whole Petri model of flow process.Wherein, storehouse institute implication explanation: p ₁: warning message; p ₂: the information of receiving a crime report; p ₃: alert information reporting information; p ₄: disposes of information in advance; p ₅: dispose ending message in advance; p ₆: the casual-clearing station section releases news; p ₇: the section of station response ending message; p ₈: on-the-spot real-time analysis information; p ₉: the emergent feedback information of the section of station; p ₁₀: alert information; p ₁₁: casual-clearing station section early warning information; p ₁₂: the crash analysis object information; p ₁₃: the expert info library information; p ₁₄: the expert confers information; p ₁₅: prediction scheme information; p ₁₆: the prediction scheme recalls information; p ₁₇: the prediction scheme update information; p ₁₈: prediction scheme is determined information; p ₁₉: scheme is determined information; p ₂₀: the scheme update information; p ₂₁: scheme implementation information; p ₂₂: the accident investigation personnel information that puts in place; p ₂₃: the personnel of the emergency leading group information that puts in place; p ₂₄: implementation information is carried out in emergency management and rescue; p ₂₅: state of affairs appreciation information; p ₂₆: expert's feedback information; p ₂₇: the rescue ending message; p ₂₈: expression later stage disposes of information; p ₂₉: expression rehabilitation information; p ₃₀: represent incident disposes of information concerning foreign affairs; p ₃₁: expression survey feedback information; p ₃₂: the expression archive information; p ₃₃: expression assessment ending message; p ₃₄: expression accident investigation reporting information; p ₃₅: the expression letter reports ending message; p ₃₆: expression emergency management and rescue overall process information.Transition implication explanation: t ₁: accident alarming; t ₂: accident is received a crime report; t ₃: alert is passed on; t ₄: the casual-clearing station section is disposed in advance; t ₅: basic-level stations and depots emergency response (guaranteeing and the Police Command Center information feedback); t ₆: the alert analysis begins; t ₇: the field accident information analysis; t ₈: the expert holds a conference or consultation; t ₉: the prediction scheme analysis; t ₁₀: prediction scheme is called; t ₁₁: scheme generates; t ₁₂: scheme is determined; t ₁₃: do not need road bureau's scope emergency management and rescue; t ₁₄: start emergency preplan; t ₁₅: state of affairs assessment; t ₁₆: implement the rescue scheme; t ₁₇: the emergency management and rescue task termination; t ₁₈: the expression later stage is disposed; t ₁₉: expression is disposed and is finished; t ₂₀: the expression performance evaluation; t ₂₁: the presentation of events filing; t ₂₂: expression investigation information reports; t ₂₃: expression is emergent to be finished.

It should be noted that (1) storehouse p of institute ₄, p ₆, p ₁₀At transition t ₃Can have Tuo Ken (the willing implication of holder is the trigger message of each incident in the rescue flow process) after implementing simultaneously, i.e. transition t ₃After the enforcement, Police Command Center is in to alert analysis and prediction scheme analysis, and issue accident information in the casual-clearing station section will be disposed in advance simultaneously and stand organizes station section personnel actively to speedily carry out rescue work, and makes accident infringement and influence reach minimum as far as possible.(2) p of storehouse institute ₈Expression casual-clearing station section in emergency response must and Police Command Center between carry out real-time information alternately, this not only helps carrying out smoothly of the section of station emergency response work, improve rescue efficiency, and can and whether enlarge the emergent decision-making foundation that provides for expert decision-making, guarantee the correctness of solution formulation.(3) p of storehouse institute ₉Expression station segment information feedback, promptly the casual-clearing station section will provide up-to-date enforcement multidate information for station section emergency command office in emergency response.(4) transition t ₄, t ₅, t ₆It is concurrency relation, after casual-clearing station section dispatching office or emergent section send to emergency office of road bureau with warning message, emergency office will organize the expert personnel to carry out work such as field accident analysis, expert are held a conference or consultation, prediction scheme analysis immediately, here with the refinement of expert decision-making flow process, in order that make things convenient for theoretical research, three aspect work (transition t in fact ₄, t ₅, t ₆) there is not obvious limit, generally all be to intersect to carry out.(5) as can be known, transition t by Petri net principle ₁₁Enforcement need the p of storehouse institute ₁₂, p ₁₄, p ₁₈, p ₂₀Have identical holder and agree, and the p of storehouse institute ₂₀Had Tuo Ken when the initialization of net, therefore final rescue scheme generates need have the crash analysis object information simultaneously, the expert confers information and prediction scheme is determined information, i.e. the p of storehouse institute ₁₂, p ₁₄, p ₁₈All have Tuo Ken, this modelling thought is more realistic.(6) transition t ₁₃, t ₁₄Be conflict relationship, represent or need road bureau's scope emergency management and rescue, perhaps do not need road bureau's scope emergency management and rescue, two kinds of situations can only be selected one.(7) p of storehouse institute ₅, p ₇, p ₂₅, p ₂₇Must have simultaneously and hold in the palm when agreeing transition t ₁₇Just can implement.Here it is to be noted the p of storehouse institute ₂₅, p ₂₇Be not only transition t ₁₅, t ₁₆Output storehouse institute, but also be transition t ₁₃Output storehouse institute, main cause is not need the emergency management and rescue (t of road bureau ₁₃Enforcement) under the situation, transition t ₁₇The prerequisite of implementing is the p of input magazine institute equally ₂₅, p ₂₇Tuo Ken must be arranged, so the p of storehouse institute ₂₅, p ₂₇Be transition t ₁₃The output storehouse mainly be correctness, rationality in order to guarantee model.

Be illustrated in figure 4 as and simplify the sequential organization synoptic diagram.In step 108 simplified model process, need simplify order Petri net basic structure.Short-cut method is as follows: by n timed transition t ₁, t ₂... t _nIn the sequential organization of forming, the delay time of establishing this n the transition of connecting is a n mutually independent random variables, and to obey parameter respectively be λ ₁, λ ₂..., λ _nExponential distribution, i.e. the average delay time of n transition is respectively 1/ λ ₁, 1/ λ ₂..., 1/ λ _n, the delay time of equivalent time transition t is 1/ λ, then

$\frac{1}{\mathrm{\λ}}=\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\frac{1}{{\mathrm{\λ}}_i}$

In addition, be stochastic variable Y if establish the delay time of equivalent time transition t, further proof Y obeys n rank hyperexponential distribution, and its probability density function is:

$F_Y\left(y\right)=\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{m}_i{\mathrm{\λ}}_i{e}^{-{\mathrm{\λ}}_{i}y},y>0$

Wherein $m_i=\underset{j=1,j\≠i}{\overset{n}{\mathrm{\Π}}}\frac{{\mathrm{\λ}}_j}{{\mathrm{\λ}}_j-{\mathrm{\λ}}_i},1\≤i\≤n$

Be illustrated in figure 5 as and simplify concurrent structural representation.In step 108 simplified model process, need simplify concurrent Petri net basic structure.Short-cut method is as follows: by n timed transition t ₁, t ₂... t _nIn the concurrent structure of forming, the delay time of establishing this n transition in parallel is a n mutually independent random variables, and to obey parameter respectively be λ ₁, λ ₂..., λ _nExponential distribution, i.e. the average delay time of n transition is respectively 1/ λ ₁, 1/ λ ₂..., 1/ λ _n, the average delay time of equivalent time transition t is 1/ λ, then:

$\frac{1}{\mathrm{\λ}}=\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\frac{1}{{\mathrm{\λ}}_i}-\underset{i=1}{\overset{n-1}{\mathrm{\Σ}}}\underset{j=i+1}{\overset{n}{\mathrm{\Σ}}}\frac{1}{{\mathrm{\λ}}_i+{\mathrm{\λ}}_j}+\underset{i=1}{\overset{n-2}{\mathrm{\Σ}}}\underset{j=i+1}{\overset{n-1}{\mathrm{\Σ}}}\underset{k=j+1}{\overset{n}{\mathrm{\Σ}}}\frac{1}{{\mathrm{\λ}}_i+{\mathrm{\λ}}_j+{\mathrm{\λ}}_k}+\·\·\·+{(-1)}^{n-1}\frac{1}{\underset{1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\λ}}_i}$

In addition, be stochastic variable Y if establish the delay time of equivalent time transition t, then stochastic variable Y obeys hyperexponential distribution.(referring to accompanying drawing 4 explanations)

Be illustrated in figure 6 as and simplify the choice structure synoptic diagram.In step 108 simplified model process, need simplify selecting Petri net basic structure.Short-cut method is as follows: n timed transition t ₁, t ₂... t _nDelay time be n mutually independent random variables, and to obey parameter respectively be λ ₁, λ ₂..., λ _nExponential distribution, i.e. the average delay time of n transition is respectively 1/ λ ₁, 1/ λ ₂..., 1/ λ _n, carry out transition t _iProbability be α _i, 1≤i≤n, and $\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\α}}_i=1,$ The average delay time of equivalent time transition t is 1/ λ, then:

$\frac{1}{\mathrm{\λ}}=\underset{i=1}{\overset{n}{\mathrm{\Σ}}}\frac{{\mathrm{\α}}_i}{{\mathrm{\λ}}_i}$

In addition, be stochastic variable Y if establish the delay time of equivalent time transition t, then stochastic variable Y obeys hyperexponential distribution.(referring to accompanying drawing 4 explanations)

Be illustrated in figure 7 as and simplify the loop structure synoptic diagram.In step 108 simplified model process, need simplify cycle P etri net basic structure.Short-cut method is as follows: establish timed transition t ₁And t ₂Delay time be stochastic variable, obeying parameter respectively is λ ₁And λ ₂Exponential distribution, i.e. t ₁The average delay time be 1/ λ ₁, t ₂The average delay time be 1/ λ ₂, as transition t ₁In the time of can triggering, the probability of its actual triggering is α, and the average delay time of equivalent time transition t is 1/ λ, then:

$\frac{1}{\mathrm{\λ}}=\frac{1}{1-\mathrm{\α}}\·(\frac{\mathrm{\α}}{{\mathrm{\λ}}_1}+\frac{1}{{\mathrm{\λ}}_2})$

In addition, the delay time of equivalent time transition t is actually the weighted mean value of some cascaded structure average delay times, therefore should obey hyperexponential distribution.

Above-mentioned Fig. 4, Fig. 5, Fig. 6, Fig. 7 are the simplification processes to single basic structure, after the simplification, the naive model connection need be obtained the SPN model of overall simplification.Short-cut method is as follows: defining basic SPN (ESPN) is the proper subclass of SPN, but recursive definition is: the SPN that (1) only contains a timed transition is ESPN; (2) be ESPN with the resulting SPN of any time transition among any replacement ESPN in four kinds of basic structure.Thought based on chromatographic analysis, ESPN is carried out the performance basic skills of analyzing of equal value is: find out four kinds of basic structures among the ESPN earlier, each basic structure is all regarded a subnet as, utilize the performance equivalence formula to carry out abbreviation then, the ESPN model that obtains simplifying repeats said process to simplified model again, till can not continuing simplification, obviously, the complexity of this method with the growth of model complexity linear growth.The final SPN model that connects four kinds of basic structures obtains overall simplification SPN model.

By said method, can on existing railway emergency plan system basis, utilize the stochastic Petri net correlation theory to set up the SPN model of railway emergency plan.Because the SPN model has the characteristics of stochastic Petri net, the formulation mode of its existing strictness, avatars mode is intuitively also arranged, this is just for solving in the existing railway emergency plan system, and of a great variety, problems such as grade is complicated, mode of management is unreasonable, retrieval inconvenience and resource sharing difference provide condition; Simultaneously, its powerful data analysis function also provides help for train operation decision maker's decision support.

The above; only for the preferable embodiment of the present invention, but protection scope of the present invention is not limited thereto, and anyly is familiar with those skilled in the art in the technical scope that the present invention discloses; the variation that can expect easily or replacement all should be encompassed within protection scope of the present invention.Therefore, protection scope of the present invention should be as the criterion with the protection domain of claim.

Claims (5)

1, a kind of modeling method of the railway emergency plan based on stochastic Petri net is characterized in that said method comprising the steps of:

The type and the grade of step 1, the existing railway emergency plan of analysis are divided prediction scheme and arrangement, realize the templating of emergency preplan content;

Step 2, extract the general character content in all kinds of accident rescues in overall prediction scheme, comprehensive prediction scheme and the station section prediction scheme, and prediction scheme is divided into flow process on the basis that templating handles in that the emergency preplan content is carried out by putting these general character contents in order;

Step 3, based on the stochastic Petri net theory, the prediction scheme flow process is carried out qualitative analysis and quantitative test, thereby sets up the Petri model for respectively the prediction scheme flow process of dividing in the step 2;

The Petri model of setting up in step 4, the integration step 3 by the contact between the model, connects into a block mold with numerous single models;

Step 5, the block mold that step 4 is set up carry out performance abbreviation of equal value;

After step 6, the abbreviation, thereby naive model connected the SPN model obtain overall simplification.

2, the modeling method of a kind of railway emergency plan based on stochastic Petri net according to claim 1, it is characterized in that in the described step 1, the templating of emergency preplan content is based on the existing emergency preplan template of railway, at dissimilar prediction schemes, similar stencil design goes out different concrete template frameworks.

3, the modeling method of a kind of railway emergency plan based on stochastic Petri net according to claim 1, it is characterized in that in the described step 3, qualitative analysis is to utilize the stochastic Petri net theory, uses accessibility, boundedness and the security of T_ invariant and Markov process analysis model.

4, the modeling method of a kind of railway emergency plan based on stochastic Petri net according to claim 1, it is characterized in that in the described step 3, quantitative test be by Markov process transfer rate matrix and fuzzy mathematics theory solved the storehouse utilization factor and system's average delay time of busy probability, storehouse institute idle probability, system's transition.

5, the modeling method of a kind of railway emergency plan based on stochastic Petri net according to claim 1, it is characterized in that in the described step 5 that it is to four kinds of the most basic structures in the Petri net that the block mold that step 4 is set up carries out performance abbreviation of equal value---in proper order, concurrent, selection and circulation carry out time performance abbreviation of equal value; Its method is that each basic structure is regarded as a subnet, represents with equivalent time transition respectively, and this timed transition has the time expectation that equates with original subnet, thereby obtains the SPN model of a simplification.

Images