A method and computer program product for simulating catastrophic events based on Latin hypercube sampling method

Through the catastrophe event sampling method based on the Latin hypercube method, the problem of inaccurate simulation of low-probability events in Monte Carlo sampling is solved, and the accuracy of catastrophe event simulation is improved, making it closer to the actual situation.

CN113344711BActive Publication Date: 2024-11-12CHINA REINSURANCE (GROUP) CORPORATION
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202110390960.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-04-12
Publication Date
2024-11-12
Estimated Expiration
2041-04-12

AI Technical Summary

Technical Problem

When the number of sampling times is small in catastrophe event simulation, the simulation of low-probability events is not accurate enough, resulting in the aggregation phenomenon and distribution deviation, affecting the accuracy of the overall loss calculation.

Method used

The catastrophe event sampling method based on the Latin hypercube method is adopted. By obtaining and sorting the event frequency, the sampling interval value is determined, and the interval value is sampled in sequence and processed in an out-of-order manner to generate a simulated catastrophe event sequence to improve the sampling accuracy of low-probability events. sex.

Benefits of technology

The sampling accuracy of low-probability events is improved, making the simulated catastrophe events happen closer to the actual situation, and reducing the distribution deviation of the results.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113344711B_ABST
    Figure CN113344711B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of catastrophe insurance business, and in particular to a catastrophe event sampling method based on the Latin hypercube method and a computer program product, the method comprising a parameter acquisition step, a sorting step, a sampling interval value determination step, a sampling step, and a random order step; the computer program product of the present invention implements all the steps of the method of the present invention when executed by a processor. The present invention improves the sampling accuracy of low-probability events, and the simulated occurrence of catastrophe events is closer to the actual situation.
Need to check novelty before this filing date? Find Prior Art

Description

[0001] The present invention relates to the field of catastrophe insurance business, and in particular to a catastrophe event sampling method based on the Latin hypercube method and a computer program product. Background Art

[0002] In the process of calculating the loss of catastrophic events and pricing insurance, the system must first simulate the catastrophic events generated by the catastrophic model in the time domain according to their frequency of occurrence, that is, to generate the occurrence of catastrophic events within a given time period. The catastrophic events generated by the catastrophic model contain a variety of information such as the physical characteristics of the event itself and the losses caused to the insurance subject. Among them, the annual average frequency of the event is one of the most important information fields in the event simulation process. The event sampling process obeys the law of large numbers, that is, the same experiment is repeated many times. When the number of experimental samples is large enough, the sample mean and the true (population) mean are sufficiently close. The Monte Carlo sampling method is a simple random sampling method based on this law. It refers to the traditional technique of sampling from a probability distribution using random numbers or pseudo-random numbers. This method can also be used for catastrophic event simulation.

[0003] When using the Monte Carlo method to sample catastrophic events, the frequency interval of the event in which the sampling falls is determined by the random number method. When the number of simulations is large, the actual sampling results of this method can be close to the annual frequency of the event itself. However, when the number of samplings is small, the actual sampling results may be quite different from the frequency distribution of each event. The Monte Carlo sampling technique is completely random, that is, within the range of the input distribution, the sample can fall anywhere. Of course, the sample is more likely to be drawn from the distribution area with a high probability of occurrence. When the number of iterations is small, the sampling will have the problem of aggregation. Values ​​outside the distribution range are not considered as samples, and their impact on the results is not included in the simulation output. When the distribution contains low-probability results, the aggregation becomes particularly obvious and may have a large impact on the results. It is important to consider the impact of these low-probability results. Since the frequency of catastrophic events is often indeed very small, and the smaller the frequency of the event, the greater its loss may be, and the greater the impact on the overall loss, the simulation of low-probability events is even more important. Summary of the invention

[0004] In view of the technical problems existing in the prior art, the technical solution of the present invention proposes a method for simulating catastrophic events based on the Latin hypercube sampling method, including the following steps of simulating catastrophic event sequences:

[0005] The parameter acquisition step acquires the event frequency and event identifier of the simulated catastrophe event;

[0006] The sorting step is to sort the simulated catastrophe events according to the event frequency to obtain the corresponding sorting event frequency and sorting event identifier;

[0007] The step of determining the sampling interval value is to divide the sum of the sorting event frequencies by the sampling number N to obtain the sampling interval value δ each time;

[0008] The sampling step samples the sorting event frequency in sequence at intervals of δ, and records the sorting event identifier corresponding to each sampling;

[0009] The shuffling step shuffles all the extracted sorted event identifiers to obtain a simulated catastrophic event sequence.

[0010] On the other hand, a computer program product is provided, comprising a computer program, characterized in that the processor executes the steps of any method for simulating catastrophic events based on the Latin hypercube sampling method.

[0011] The beneficial effect of the present invention is that the sampling accuracy of low-probability events is improved, and the simulated occurrence of catastrophic events is closer to the actual situation. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] Figure 1 . A line segment diagram of a sampling step in one embodiment;

[0013] Figure 2 . A simulated catastrophe event sequence obtained in one embodiment;

[0014] Figure 3 .A flow chart of an implementation method. DETAILED DESCRIPTION

[0015] Some embodiments of the present invention relate to a method for simulating a catastrophic event based on a Latin hypercube sampling method, including the following steps of simulating a catastrophic event sequence:

[0016] The parameter acquisition step acquires the event frequency and event identifier of the simulated catastrophe event;

[0017] The sorting step is to sort the simulated catastrophe events according to the event frequency to obtain the corresponding sorting event frequency and sorting event identifier;

[0018] The step of determining the sampling interval value is to divide the sum of the sorting event frequencies by the sampling number N to obtain the sampling interval value δ each time;

[0019] The sampling step samples the sorting event frequency in sequence at intervals of δ, and records the sorting event identifier corresponding to each sampling;

[0020] The shuffling step shuffles all the extracted sorted event identifiers to obtain a simulated catastrophic event sequence.

[0021] The step of obtaining parameters also includes: obtaining the total cumulative frequency WT of the sorting event frequency; the sampling step includes:

[0022] When the current sorting event frequency is between two adjacent cumulative frequencies WTn, the corresponding sorting event identifier is recorded until N sorting event identifiers are extracted.

[0023] During the first sampling, any event frequency within the range of greater than 0 and less than δ is selected and recorded as the initial value of the current sorting event frequency.

[0024] The method for simulating the time series of catastrophic events also includes the following steps:

[0025] The parameter acquisition step acquires the time occurrence frequency and time mark of the simulated catastrophic event; the sorting step sorts the simulated catastrophic event according to the time occurrence frequency to obtain the corresponding sorting time occurrence frequency and sorting time mark;

[0026] Determine the sampling interval value step by dividing the total frequency of sorting time by the number of samples N to obtain the sampling interval value δ for each time 时间 ;

[0027] The sampling step sorts the simulated catastrophic events according to δ 时间 Sampling is performed in sequence at intervals of N, and the sorting time labels corresponding to each sampling are recorded;

[0028] The shuffling step shuffles all the extracted sorted time tags to obtain a simulated catastrophic event occurrence time series.

[0029] The step of obtaining parameters also includes: obtaining the total cumulative frequency WT of the sorting time occurrence frequency 时间 ;

[0030] The sampling steps include:

[0031] The frequency of occurrence of the current sorting time is between the two adjacent cumulative frequencies WT 时间 When the number of sorting timestamps is between 0 and n, the corresponding sorting time identifiers are recorded until N sorting time identifiers are extracted.

[0032] Calculating the sampling number comprises the following steps:

[0033] The value obtaining step obtains the simulation year number Y;

[0034] The probability calculation step is to calculate P(k) in sequence according to formula I,

[0035]

[0036] Where P(k) represents the probability of k catastrophic events occurring in a year, k represents the number of catastrophic events, k = 0, k = 1, k = 2, ..., λ is the sum of the frequencies of the sorted events, and e is a natural constant;

[0037] The steps for calculating the number of years are as follows: Q(k) is calculated in sequence according to formula II:

[0038] Q(k)=P(k)×Y Formula II

[0039] Where Q(k) represents the number of years in which k catastrophic events occur; k represents the number of catastrophic events, k = 1, k = 2, ..., and Y represents the number of simulation years;

[0040] Repeat the calculation steps and compare the Q(k) calculated each time with the critical value. When Q(k) is greater than or equal to the critical value and less than 1, When Q(k) is less than λ×Y, Q(k)=1 is set. When Q(k) is less than the critical value, Q(k)=0 is set and the step of calculating the number of years stops.

[0041] Record said k as m;

[0042] The summation step calculates N according to formula III,

[0043]

[0044] Where N is the number of samples, k is the number of times a catastrophic event occurs, k = 1, ..., m, and Q(k) is the number of years in which k catastrophic events occur.

[0045] The method for simulating the time series of catastrophic events also includes a time conversion step:

[0046] According to Q(k), we can obtain the frequency distribution of k in the Y years;

[0047] Scramble the frequency distribution to obtain a yearly frequency sequence;

[0048] Represent the annual occurrence frequency sequence according to the year identification sequence of each event;

[0049] According to the year mark of each event, the time mark of the simulated catastrophic event is marked as the converted time mark according to the corresponding year.

[0050] The method for calculating the time occurrence frequency comprises the following steps:

[0051] Obtaining a seasonal parameter, wherein the seasonal parameter is a frequency ratio of a catastrophic event occurring in each season of a year;

[0052] Obtaining a time frequency weight according to the seasonal parameter;

[0053] The time frequency of the simulated catastrophic event is obtained by the time frequency weight.

[0054] The disordering method comprises the following steps:

[0055] Get the number of elements in the shuffled sequence and record it as X;

[0056] Generate X random numbers from 1 to X;

[0057] The shuffled sequence is arranged in the order of the random numbers and output.

[0058] The original data of the simulated catastrophic event is in the form of a data table, each column of the data table records the values ​​of multiple attributes of the simulated catastrophic event, each row records the values ​​of multiple attributes of a specific catastrophic event, and the data table has enough rows.

[0059] The term "sequence" refers to a data structure that contains multiple data items. The sequence contains multiple data items (also called members) arranged in order. The members can be accessed by indexing. They can be stored in arrays, but are not limited to arrays. The address of each element in the array corresponds to the order of the elements. The sampling results are as follows: Figure 2 shown.

[0060] The term “event identification” refers to the naming of different catastrophic events, including but not limited to the naming form of ID.

[0061] The term "correspondence" refers to the relationship between data located in the same row in the original data table of catastrophic events, for example, an event identification ID corresponds to a unique event frequency.

[0062] The term "sorted event frequency" refers to the event frequency after the original data table of simulated catastrophic events is rearranged in the order of event frequency values ​​from large to small or from small to large.

[0063] The term "cumulative frequency WT" refers to the sequence obtained by successively accumulating and summing the occurrence frequencies in the sorted event frequencies. WT is an increasing sequence, and WT[n]-WT[n-1] is the nth occurrence frequency in the sorted event frequencies.

[0064] The term "time mark" refers to the specific time when the catastrophic event occurs, including but not limited to day mark, week mark, month mark, and quarter mark.

[0065] The term "frequency of occurrence in time" includes, but is not limited to, the frequency of occurrence in units of days, weeks, months, quarters, and years.

[0066] The term “seasonal parameter” refers to the ratio of the frequency of catastrophic events occurring in each month of the year.

[0067] The term "random number" includes pseudo-random numbers generated by a computer.

[0068] The following is combined with Figure 1-3 The present invention is continued to be explained.

[0069] In some implementations, the main steps of simulating a catastrophic event using the Latin hypercube sampling method are as follows:

[0070] 1. Sort the events in descending order of frequency;

[0071] For example, there are four catastrophic events, with IDs 28001, 28002, 28003, and 28004, and their annual occurrence frequencies are 0.1, 0.05, 0.07, and 0.03, respectively. In this step, the events need to be arranged from high to low in terms of frequency. The event sequence after arrangement is:

[0072] 28001, 28003, 28002, 28004.

[0073] 2. Grid segmentation of events based on the number of samples:

[0074] For example, if the simulation needs to extract 2500 events from the four events 28001, 28002, 28003, and 28004, then in this step, the events need to be divided into 2500 parts. Before dividing, the sum of the frequencies of all events must be calculated first. The sum of the frequencies of these four events is 0.25. If it is divided into 2500 parts, the frequency of each part is 0.0001.

[0075] 3. Determine the sampling results within each grid:

[0076] Step 2: Divide the total frequency 0.25 into 2500 grids, each grid is 0.0001. The frequency accumulation of the four events after sorting is as follows: Figure 1 shown.

[0077] When sampling each grid, the median of each grid is taken. For example, the median of the first grid is 0.00005, which is between 0 and 0.03. The event extracted from the first grid is 28004. The median of the second grid is 0.00015, which is still between 0 and 0.03. The event extracted from the second grid is still 28004. Similarly, when the median is greater than 0.03 and less than or equal to 0.08, the grid corresponds to event 28002. When the median is greater than 0.08 and less than or equal to 0.15, the grid corresponds to event 28003. When the median is greater than 0.15 and less than or equal to 0.25, the grid corresponds to event 28001. According to this method, a specific event sequence is generated as follows: Figure 2 shown.

[0078] The term "grid" means, Figure 1Each small line segment in represents a frequency.

[0079] The term "meshing events" refers to Figure 1 The total frequency is divided according to the length of the line segment represented by the shown.

[0080] Some embodiments of the present invention further relate to a computer program product, comprising a computer program, characterized in that when the computer program is executed by a processor, the steps of any one of the methods described in claims 1-9 are implemented.

[0081] A computer program implementation process of an implementation method is as follows:

[0082] The input parameters of the implementation method include the number of simulation years numTrials, the event ID array eventId, the event frequency array eventFreq, and the seasonality parameter seasonality.

[0083] The output results include the event occurrence time sequence simuTrialArray and the event ID sequence simuEventIdArray.

[0084] Among them, seasonality is a string of 12 numbers, which represents the probability ratio of the event occurring in different months, for example, 1,1,1,1,1,1,2,2,2,2,2,2,2, which means that the probability of the event occurring in July to December is twice the probability of the event occurring in January to June. The event occurrence time series simuTrialArray and the event ID series simuEventIdArray are one-to-one corresponding. The event occurrence time series is a series of integer sequences. For example, 5 means that the event occurred on the fifth day of the first year, and 366 means that the event occurred on the first day of the second year. If the first numbers of simuTrialArray and simuEventIdArray are 2 and 28003 respectively, it means that event 28003 occurred on the second day of the first year.

[0085] The overall process of computer implementation is as follows:

[0086] 1. Add all the numbers in eventFreq to get the sum of the frequencies of all events, totalFreq:

[0087] double totalFreq=eventFreq.Sum()

[0088] 2. Calculated by using the distribution function of the Poisson distribution:

[0089] Where λ = totalFreq, calculate the probabilities p for k = 0, k = 1, k = 2... respectively. p × numTrials is the number of years with the corresponding number of catastrophic events. For example, when k = 1, if the calculated result is 0.44 and numTrials = 100000, then there are 44000 years with only one catastrophic event occurring in these years. When 0.4 < P × numTrials < 1 and the total number of occurrences has not reached λ × numTrials, let P × numTrials = 1. When P × numTrials ≤ 0.4, let P × numTrials = 0, and the calculation stops.

[0090] 3. Use a random array to randomly shuffle the generated annual occurrence number sequence to obtain the catastrophic event occurrence number sequence for each year. For example, the annual occurrence number sequence numEventsPerYear for 100000 years is:

[0091] 2, 1, 3, 3, 1, 0, 0, 0, ……, 1, 0, 0, 2

[0092] 4. Sum all the numbers in numEventsPerYear to obtain the total number of events occurred:

[0093] int numEvents = numEventsPerYear.Sum();

[0094] 5. Calculate the specific event occurrence ID sequence simuEventIdArray using eventFreq, eventId, and numEvents;

[0095] 6. Calculate the event occurrence time sequence simuTrialArray using seasonality and numEventsPerYear;

[0096] Among them, the program method flow for calculating simuEventIdArray is as follows:

[0097] 5.1 Sort eventFreq and eventId in descending order according to eventFreq to obtain the corresponding sortedEventFreq and sortedEventId;

[0098] 5.2 Calculate the total probability sum totalWt of all events through sortedEventFreq;

[0099] 5.3 Obtain the cumulative frequency WT of each event by accumulation, that is:

[0100] WT[0] = sortedEventFreq[0]

[0101] WT[n]=WT[n-1]+sortedEventFreq[n]

[0102] 5.4 Calculate the size of the segmentation grid value:

[0103]

[0104] 5.5 Let myWt = delta / 2, and perform the first sampling. If

[0105] myWt≤WT[i], myWt>WT[i-1], then the sampling event corresponding to the grid is sortedEventId[i], and traverse WT[n] at the same time, and then myWt is added with delta each time to sample each grid;

[0106] 5.6 Repeat step 5.5 until numEvents times are added, generating the event sequence myOut;

[0107] 5.7 Use a random array to randomly shuffle the generated event sequence to obtain the specific event ID sequence simuEventIdArray.

[0108] The program method flow for calculating simuTrialArray is as follows:

[0109] 6.1 The time series eventTrialData of the event occurrence is obtained based on the annual occurrence frequency sequence numEventsPerYear. For example, if numEventsPerYear is 2, 3, 0, ..., 1, it proves that the first two events occurred in the first year, and the next three events occurred in the second year, then eventTrialData is 1, 1, 2, 2, 2, ...

[0110] 6.2 Based on seasonality, we get the probability weight of each month. Assuming that each year has 365 days, the probability of occurrence of each day is the probability weight of the month divided by the number of days in the month. For example, if the first day is in January and the weight of January is 1, the probability of occurrence of the first day is 1 / 31. If the 34th day is in February and the weight of February is 2, the probability of occurrence of the 24th day is 2 / 28. And so on, we get the probability of occurrence of each day in the year dailyRate

[0111] 6.3 Use dailyRate, daily sequence dailyInYear(1,2,3,……,365) and numEvents to calculate the daily sequence aDailyTrial of each event. The calculation method is the same as the program method flow for calculating simuEventIdArray.

[0112] 6.4 Use eventTrialData and aDailyTrial to calculate the time series information of each event according to the formula aDailyTrial[i]=trailArray[i]*365+aDailyTrial[i]

[0113] 6.5 Use the random array to randomly shuffle the generated time series to obtain the random time series simuTrialArray where the events occur.

[0114] It should be pointed out that the present invention is not limited to a specific computer program implementation language.

[0115] In some other embodiments, a computer program product and device can be implemented to simulate catastrophic events based on the Latin hypercube sampling method. The device includes at least one processor and a memory storing instructions. When the instructions are executed by at least one processor, the method described in any of the above embodiments is performed.

[0116] A computer program (which may also be referred to or described as a program, software, software application, module, software module, script or code) may be written in any form of programming language, including compiled or interpreted languages ​​or declarative or procedural languages, and the computer program may be deployed in any form, including as a stand-alone program or as a module, component, subroutine or other unit suitable for use in a computing environment. A computer program may, but need not, correspond to a file in a file system. A program may be stored in a portion of a file that stores other programs or data, for example, one or more scripts stored in the following: in a markup language document; in a single file dedicated to a related program; or in multiple collaborative files, for example, files storing one or more modules, subroutines or code portions. A computer program may be deployed to execute on one or more computers, the computers being located in one place or distributed to multiple locations and interconnected by a communication network.

[0117] The processes and logic flows described in this specification can be performed by one or more programmable computers executing one or more computer programs to perform functions by operating on input data and generating output.

[0118] A computer suitable for implementing a computer program includes and exemplarily can be based on a general-purpose microprocessor or a special-purpose microprocessor or both of the above-mentioned processors, or a central processing unit of any other kind. Typically, the central processing unit will receive instructions and data from a read-only memory or a random access memory or both. The main elements of a computer are a central processing unit for running or executing instructions and one or more memory devices for storing instructions and data. Typically, a computer will also include or be operably coupled to receive data from one or more large-capacity storage devices for storing data or transfer data to a large-capacity storage device, or receive and transfer both, the large-capacity storage being, for example, a magnetic disk, a magneto-optical disk, or an optical disk. However, a computer does not necessarily have such a device.

[0119] Computer readable media suitable for storing computer program instructions and data include all forms of nonvolatile memory, media and memory devices, including, as examples: semiconductor memory devices, such as EPROM, EEPROM and flash memory devices; magnetic disks, such as internal hard disks or removable disks; magneto-optical disks; CD-ROM and DVD-ROM disks. The processor and memory may be supplemented by or incorporated into a special purpose logic circuit.

[0120] To transmit interactions with a user, embodiments of the subject matter described in this specification may be implemented on a computer having: a display device, such as a CRT (cathode ray tube) or LCD (liquid crystal display) monitor, for displaying information to the user; and a keyboard and a pointing device, such as a mouse or a trackball, with which the user can transmit input to the computer. Other kinds of devices may also be used to transmit interactions with a user; for example, feedback provided to the user may be any form of sensory feedback, such as visual feedback, auditory feedback, or tactile feedback; and input from the user may be received in any form, including acoustic input, voice input, or tactile input. In addition, a computer may interact with a user by sending documents to a device used by the user and receiving documents from the device; for example, by sending a web page to a web browser on a user's client device in response to a request received from the web browser.

[0121] Specific implementations of the subject matter have been described. Other implementations are within the scope of the following claims. For example, the activities recited in the claims can be performed in a different order and still achieve the desired results. As an example, the processes described in the accompanying drawings do not necessarily require the specific order or sequential order shown in order to achieve the desired results. In certain implementations, multitasking and parallel processing may be advantageous.

Claims

1. A method for simulating catastrophic events based on Latin hypercube sampling method, characterized in that: The method for simulating a catastrophic event sequence comprises the following steps: The parameter acquisition step acquires the event frequency and event identifier of the simulated catastrophe event; The sorting step is to sort the simulated catastrophe events according to the event frequency to obtain the corresponding sorting event frequency and sorting event identifier; The step of determining the sampling interval value is to divide the sum of the sorting event frequencies by the sampling number N to obtain the sampling interval value δ each time; The sampling step samples the sorting event frequency in sequence at intervals of δ, and records the sorting event identifier corresponding to each sampling; The shuffling step shuffles all the extracted sorted event identifiers to obtain a simulated catastrophe event sequence; The step of obtaining parameters further includes: obtaining the total cumulative frequency WT of the sorting event frequency; The sampling steps include: When the current sorting event frequency is between two adjacent cumulative frequencies WTn, the corresponding sorting event identifier is recorded until N sorting event identifiers are extracted; During the first sampling, any event frequency within the range of greater than 0 and less than δ is recorded as the initial value of the current sorting event frequency; The method for simulating the time series of catastrophic events also includes the following steps: The parameter acquisition step acquires the time occurrence frequency and time mark of the simulated catastrophic event; the sorting step sorts the simulated catastrophic event according to the time occurrence frequency to obtain the corresponding sorting time occurrence frequency and sorting time mark; Determine the sampling interval value step by dividing the total frequency of sorting time by the number of samples N to obtain the sampling interval value δ for each time 时间 ; The sampling step sorts the simulated catastrophic events according to δ 时间 Sampling is performed in sequence at intervals of N, and the sorting time labels corresponding to each sampling are recorded; The shuffling step shuffles all the extracted sorting time tags to obtain the simulated disaster event occurrence time series; The step of obtaining parameters also includes: obtaining the total cumulative frequency WT of the sorting time occurrence frequency 时间 ; The sampling steps include: The frequency of occurrence of the current sorting time is between the two adjacent cumulative frequencies WT 时间 n, record the corresponding sorting time identifier until N sorting time identifiers are extracted; and calculating the sampling number includes the following steps: The value obtaining step obtains the simulation year number Y; The probability calculation step is to calculate P(k) in sequence according to formula I, Where P(k) represents the probability of k catastrophic events occurring in a year, k represents the number of catastrophic events, k = 0, k = 1, k = 2, ..., λ is the sum of the frequencies of the sorted events, and e is a natural constant; The steps for calculating the number of years are as follows: Q(k) is calculated in sequence according to formula II: Q(k)=P(k)×Y Formula II Where Q(k) represents the number of years in which k catastrophic events occur; k represents the number of catastrophic events, k = 1, k = 2, ..., and Y represents the number of simulation years; Repeat the calculation steps and compare the Q(k) calculated each time with the critical value. When Q(k) is greater than or equal to the critical value and less than 1, When Q(k) is less than λ×Y, Q(k)=1 is set. When Q(k) is less than the critical value, Q(k)=0 is set and the step of calculating the number of years stops. Record said k as m; The summation step calculates N according to formula III, Where N is the number of samples, k is the number of times a catastrophic event occurs, k = 1, ..., m, and Q(k) is the number of years in which k catastrophic events occur.

2. The method according to claim 1, characterized in that The method for simulating the time series of catastrophic events also includes a time conversion step: According to Q(k), we can obtain the frequency distribution of k in the Y years; Scramble the frequency distribution to obtain a yearly frequency sequence; Represent the annual occurrence frequency sequence according to the year identification sequence of each event; According to the year mark of each event, the time mark of the simulated catastrophic event is marked as the converted time mark according to the corresponding year.

3. The method according to claim 2, wherein the method for calculating the time occurrence frequency comprises the following steps: Obtaining a seasonal parameter, wherein the seasonal parameter is a frequency ratio of a catastrophic event occurring in each season of a year; Obtaining a time frequency weight according to the seasonal parameter; The time frequency of the simulated catastrophic event is obtained by the time frequency weight.

4. The method according to any one of claims 1 to 3, characterized in that: The disordering method comprises the following steps: Get the number of elements in the shuffled sequence and record it as X; Generate X random numbers from 1 to X; The shuffled sequence is arranged in the order of the random numbers and output.

5. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to claim 4 are implemented.