Task-oriented group guarantee demand hierarchy model establishment method

Abstract

The invention discloses a task-oriented grouping guarantee demand hierarchy model establishment method, which is used for establishing a task-oriented grouping guarantee demand hierarchy model by applying a multi-layer network system evolution thought according to grouping and layering characteristics of guarantee demand prediction in a task. By analyzing the'task-system 'structural relationship of equipment and considering the characteristics of multi-stage tasks, multi-level requirements and multi-level task success rate, an equipment guarantee demand level model and a task success rate model are established, and guarantee demand prediction algorithms of two levels of single loading and marshalling are designed.

Description

Technical Field

[0001] The present invention relates to the technical field of demand forecasting, and in particular to a method for establishing a hierarchical model of grouped support requirements for tasks. Background Art

[0002] During the usage stage, the prediction of equipment support requirements is to analyze and predict the possible support requirements of the equipment during the mission based on the equipment status monitoring and fault diagnosis information, the results of the comprehensive evaluation of the technical status, and the characteristics of the facing combat and training missions. Usually, influencing factors such as mission requirements, equipment characteristics, system structure, and environmental conditions need to be considered to predict the resource requirements of the equipment during the mission. The prediction of equipment support requirements has gone through single-equipment single-resource prediction, single-equipment batch-resource prediction, and batch-equipment batch-resource prediction, generating a wealth of demand prediction methods, such as time series analysis, Cronston, Bootstrap, Bayesian and other methods. However, these methods have common deficiencies: (1) Based on historical samples, the case of zero and small samples is not considered; (2) Based on single-resource demand prediction, the resource correlation is not considered; (3) Focusing on equipment demand prediction, the demand differences at the equipment system level are not considered.

[0003] To address these deficiencies, in terms of zero and few-shot learning, Reference [1] selected wavelet transform to extract the trend and detail features of the historical demand data of spare parts according to the operation and maintenance demand characteristics of spare parts in the wind power industry, selected an appropriate time window to transform the time series into a time series pair in the form of supervised learning, and proposed a prediction algorithm GA-dwtSVR that combines genetic algorithm and support vector regression to conduct research on the demand prediction method of few-shot spare parts in the wind power industry. Reference [2] addressed the problem that it is difficult to accurately predict the remaining useful life of rotating machinery under few-shot data. By integrating convolutional autoencoder and Weibull distribution to construct a data-driven degradation behavior model of rotating machinery, a method for constructing a health factor of rotating machinery and predicting the remaining useful life driven by digital twin was proposed. Reference [3] addressed the problem of few fault data and unclear demand patterns of high-tech equipment and materials, and proposed a demand prediction method for exponential equipment and materials under the condition of time-truncated few-shot samples. Reference [4] proposed the zero and few-shot learning problems, pointing out that there are two methods for the zero-shot problem, namely embedded and generative methods, and two methods for the few-shot problem, namely based on improving the priority data utilization rate and based on data generation. In terms of resource correlation, Reference [5] considered the correlation of resource responsiveness in space and time and conducted research on load demand response decision-making. Reference [6] addressed the problem that the current combat demand of aviation ammunition only considers the total quantity and ignores the demand structure. It studied the comprehensive prediction of combat demand from both the total quantity and the structure, and constructed a total combat demand prediction model and a demand structure model for aviation ammunition respectively. In terms of demand hierarchy, References [7] and [8] considered the differences in prediction levels and believed that hierarchical demand prediction is a multi-variable time series prediction problem, which maps a set of incoherent predictions to coherent predictions through prediction coordination to meet a given set of linear constraints. However, the methods in the above-mentioned references still have the following deficiencies: (1) only modeling for the above single deficiency, (2) insufficient mining of demand characteristics oriented to tasks, and (3) the model is too complex and has poor applicability in the practice of task-oriented demand prediction. Summary of the Invention

[0004] Aiming at the above-mentioned shortcomings of the existing technologies, the present invention proposes a method for establishing a hierarchical model of grouped support requirements oriented to tasks. Based on the requirement of task success rate, taking the "task-system" structural relationship of equipment, the hierarchical model of equipment support requirements, and the task success rate model as the basis, it analyzes the characteristics of multi-stage tasks, multi-level requirements, multi-level task success rates, and multi-stage task success rates, considers influencing factors such as failure probability, system structure, system scale, and task time, and proposes a grouped support requirement prediction algorithm at two levels, namely single equipment and formation. Through the prediction algorithm, task-oriented demand prediction is realized.

[0005] To achieve the above object, the present invention provides a method for establishing a hierarchical model of grouped support requirements oriented to tasks, which includes the following steps:

[0006] Step 1: Divide the task into multiple stages, and define the equipment support requirement problem oriented to the task according to the structural relationship of equipment and components in different stages;

[0007] Step 2: For the equipment support requirement problem oriented to the task, determine the equipment support requirement hierarchy structure, and based on this hierarchy structure, construct an equipment support requirement hierarchy model;

[0008] Step 3: Based on the equipment support requirement hierarchy model, construct a mission success rate model;

[0009] Step 4: Use the single-equipment level support requirement prediction algorithm to solve the mission success rate model, and screen out the resource combinations that meet the requirements and their probabilities;

[0010] Step 5: Use the formation level support requirement prediction algorithm to solve the mission success rate model, and screen out the resource combinations that meet the requirements and their probabilities;

[0011] Step 6: Conduct statistical analysis on all the requirement solutions that meet the requirements obtained in S4 and S5, and finally obtain the support requirement resource solution.

[0012] Furthermore, the equipment support requirement problem oriented to the task defined in Step 1 is:

[0013] 1) Entity state change

[0014] Describe the influence relationship between the entity states of each level of the equipment entity by using the idea of system evolution:

[0015]

[0016] Among them, i, j, and k represent different entities; respectively represent the states of entity i and entity j at the α-th layer, represents the state of entity k at the β-th layer; represents the relationship strength between entity i and j at the α-th layer; represents the relationship strength between entity i and entity k at the β→α layer;

[0017] In formula (1), In the formula, fda i represents a i resource status, and qc, ce, tr, es, sc, and ac respectively represent a i related quality characteristics, combat equipment, mission requirements, equipment status, system status, and environmental conditions;

[0018] 2) Dynamic change relationship strength of the interaction between different entities

[0019]

[0020] Furthermore, the hierarchical structure of equipment support requirements described in step 2 is divided into an equipment formation layer, an equipment layer, a subsystem layer, and a unit component layer from top to bottom. Among them, A is defined as a certain equipment formation in the region, and A i is defined as the i-th equipment in the formation, and A ij is defined as equipment A i 's j-th subsystem, and A ijk is defined as the k-th unit component of equipment subsystem A ij .

[0021] Furthermore, when constructing the hierarchical model of equipment support requirements in step 2, based on the hierarchical structure of equipment support requirements, requirement modeling is carried out from bottom to top. The specific steps include:

[0022] Step 2.1: Requirement modeling at the unit component level;

[0023] Step 2.1.1: At the unit component level, FDA ijk is defined as the quantity of resource requirements for A ijk to restore the functional state, and FDA ijk = λA ijk *t*FNum ijk , where λA ijk represents the failure probability of A ijk , and FNum ijk represents the resource vector required for a single failure of unit component A ijk ; among them,

[0024]

[0025] In the formula, m is the type of personnel resource, M is the total number of personnel resource types, n is the type of spare part resource, and N is the total number of spare part resource types. represents the quantity of the m-th personnel resource required for the k-th unit component of the j-th subsystem of the i-th equipment in a certain formation, represents the quantity of the n-th personnel resource required for the k-th unit component of the j-th subsystem of the i-th equipment in a certain formation;

[0026] Step 2.1.2: Define PFDA ijk as the quantity of personnel resources for A ijk to restore the functional state, and define SFDA ijk as the quantity of spare part resources for A ijk to restore the functional state, and

[0027] Step 2.1.3: Define MDA ijk as A ijkThe quantity of resource requirements to meet the task requirements, i.e., the decision variable to be determined at this level, and MDA ijk = {PMDA ijk , SMDA ijk}, where PMDA ijk is the demand vector for the required personnel resources, and SMDA ijk is the demand vector for the required spare part resources, and SMDA ijk is the demand vector for the required spare part resources;

[0028] Step 2.2: Subsystem-level requirement modeling;

[0029] Step 2.2.1: At the subsystem level, represent FDA ij as the quantity of resource requirements for A ij to restore the functional state. FDA ij = {PFDA ij , SFDA ij}, where PFDA ij represents the quantity of personnel resources required for A ij to restore the functional state, and SFDA ij represents the quantity of spare part resources required for A ij to restore the functional state, and:

[0030]

[0031] k = 1, 2, … K;

[0032] Step 2.2.2: Represent MDA ij as the quantity of resource requirements for A ij to meet the task requirements, i.e., the decision variable to be determined at this level, and MDA ij = {PMDA ij , SMDA ij}, where PMDA ij is the demand vector for the required personnel resources, and SMDA ij is the demand vector for the required spare part resources;

[0033] Step 2.2.3: Calculate MDA ij through the following formula:

[0034] MDA ij = λA ij * t * ∑ k α k MDA ijk

[0035] where λA ij represents the failure probability of A ij , and α kDenote the weight of the k-th unit component;

[0036] Step 2.3: Equipment hierarchical requirement modeling;

[0037] Step 2.3.1: At the single-equipment level, represent FDA i as A i The quantity of resource requirements for restoring the functional state, FDA i ={PFDA i , SFDA i}, where PFDA i represents the quantity of personnel resource requirements for A i to restore the functional state, and SFDA i represents the quantity of spare part resource requirements for A i to restore the functional state, and:

[0038]

[0039] SFDA i ={SFDA i1 ,…, SFDA ij ,…SFDA iJ}

[0040] j = 1, 2, … J;

[0041] Step 2.3.2: Represent MDA i as A i The quantity of resource requirements for achieving the task requirements, that is, the decision variable to be solved at this level, and MDA i ={PMDA i , SMDA i}, where PMDA i is the requirement vector of the required personnel resources, and SMDA i is the requirement vector of the required spare part resources;

[0042] Step 2.3.3: Calculate MDA i through the following formula:

[0043] MDA i = λA i * t * ∑ j β j MDA ij

[0044] where λA i represents the failure probability of A i , and β j represents the weight of the j-th subsystem;

[0045] Step 2.4: Formation hierarchical requirement modeling;

[0046] Step 2.4.1: At the grouping level, represent FDA as the quantity of resource requirements for A to restore its functional state, where FDA = {PFDA, SFDA}. Here, PFDA represents the quantity of personnel resource requirements for A to restore its functional state, and SFDA represents the quantity of spare part resource requirements for A to restore its functional state, and:

[0047] PFDA = {PFDA1, … PFDA i , … PFDA I}

[0048] SFDA = {SFDA1, … SFDA i , … SFDA I}

[0049] i = 1, 2, … I;

[0050] Step 2.4.2: Represent MDA as the quantity of resource requirements for A to achieve the task requirements, that is, the decision variable to be solved at this level, and MDA = {PMDA, SMDA}. Here, PMDA is the requirement vector of the required personnel resources, and SMDA is the requirement vector of the required spare part resources;

[0051] Step 2.4.3: Calculate MDA through the following formula:

[0052] MDA = λA * t * ∑ i γ i MDA i

[0053] where λA represents the failure probability of A, and γ i represents the weight of the i-th equipment.

[0054] Furthermore, in Step 3, build task success probability models for unit components, subsystems, equipment, groupings, and multi-stage tasks respectively, to calculate the success rates of unit components, subsystems, equipment, groupings, and multi-stage tasks respectively, specifically including:

[0055] 1) Resource and state relationship model at the unit component level

[0056]

[0057] where length(MDA ijk -FDA ijk ≥0) is to count the number of elements in MDA ijk -FDA ijk that are greater than or equal to 0, and ndims(FDA ijk ) is the dimension of FDA ijk ;

[0058] 2) Resource and state relationship model at the subsystem level

[0059] σ ij = [σ ij1 ,…σ ijk ,…σ ijK *SA ij *[σ ij1 ,…σ ijk ,…σ ijK T

[0060] Among them, [σ ij1 ,…σ ijk ,…σ ijK is the state row vector of A ijk , [σ ij1 ,…σ ijk ,…σ ijK T is the state column vector of A ijk , SA ij is the adjacency matrix of the structure of A ij , that is, the association relationship of A ij in A ijk . If there is an association, it is represented by 1; if there is no association, it is represented by 0;

[0061] 3) Single equipment resource and state relationship model

[0062] σ i = [σ i1 ,…σ ij ,…σ iJ *SA i *[σ i1 ,…σ ij ,…σ iJ T

[0063] Among them, [σ i1 ,…σ ij ,…σ iJ is the state row vector of A ij , [σ i1 ,…σ ij ,…σ iJ T is the state column vector of A ij , SA i is the adjacency matrix of the structure of A i , that is, the association relationship of A i in A ij . When there is an association, it is 1; when there is no association, it is 0;

[0064] 4) Equipment group resource and state relationship model

[0065] σ = [σ1,…σ i ​​​​,…σ I *SA*[σ1,…σ i ,…σ I T

[0066] i = 1, 2, …I

[0067] Among them, [σ1, …σ i ,…σ I is the state row vector of A i , [σ1, …σ i ,…σ I T is the state column vector of A i , SA i is the adjacency matrix of the A structure, that is, the association relationship of A in A i . If there is a relationship, it is represented by 1, and if there is no relationship, it is represented by 0.

[0068] Furthermore, the single - installation level guarantee requirement prediction algorithm includes the following steps:

[0069] Step 4.1: Obtain the failure probabilities ferrytask_λA ijk , attacktask_λA ijk , returntask_λA ijk of unit component A during the ferry stage, attack stage, and return stage respectively; the resource vectors ferrytask_FNum ijk required for the l - th single failure of unit component A during the ferry stage, attack stage, and return stage; the adjacency matrices SA ijk , SA ijkl , SA of each hierarchical system structure, the overall mission success rate requirement TSR ijkl , and the mission success rate requirements TSR ijkl of the ferry stage, attack stage, and return stage; ij , SA i , SA, TSR total task as well as the mission success rate requirements TSR ferry task of the ferry stage, attack stage, and return stage; attack task TSR return task ;

[0070] Step 4.2: Initialize MDA ijk , and set the current iteration number id and the total number of iterations num;

[0071] Step 4.3: Calculate the unit - component - level FDA ijk during the ferry, attack, and return stages respectively;

[0072] Step 4.4: Calculate the subsystem - level MDA during the ferry stage respectively​​ij and σ ij and single - unit - level MDA i and σ i ;

[0073] Step 4.5: Judge whether σ i is equal to 1. If it is equal, go to Step 4.6; otherwise, go to Step 4.12;

[0074] Step 4.6: Enter the attack phase, and recalculate the subsystem - level MDA ij and σ ij and single - unit - level MDA i and σ i ;

[0075] Step 4.7: Judge whether σ i is equal to 1. If it is equal, go to Step 4.8; otherwise, go to Step 4.12;

[0076] Step 4.8: Enter the return phase, and recalculate the subsystem - level MDA ij and σ ij and single - unit - level MDA i and σ i ;

[0077] Step 4.9: Judge whether σ i is equal to 1. If it is equal, go to Step 4.12, and increment the task success count by 1;

[0078] Step 4.10: Calculate the task success probability;

[0079] Step 4.11: If the task success probability is less than the sum of the task success rate requirements for each different phase, return to Step 4.2, and adjust the MDA in sequence according to the resource types ijk , with the multiple of the resource group required for each adjustment of the unit component increased by 1; otherwise, output the MDA ijk and calculate the MDA i ;

[0080] Step 4.12: Statistically analyze the MDA that meets the requirements i , and obtain the resource combination and its probability, which is the demand prediction result.

[0081] Furthermore, the grouping - level support demand prediction algorithm includes the following steps:

[0082] Step 5.1: Respectively obtain the failure probabilities ferrytask_λA ijk of unit component A during the ferry phase, attack phase, and return phase ijk , attacktask_λA ijk , returntask_λAijk ; Unit component A ijk The resource vectors ferrytask_FNum required for the first type of single failure during the ferry stage, attack stage, and return stage ijkl , attacktask_FNum ijkl , returntask_FNum ijkl ; The adjacency matrix SA of the system structure at each level ij , SA i , SA, the overall mission success rate requirement TSR total task And the mission success rate requirements TSR for the ferry stage, attack stage, and return stage ferry task , TSR attack task , TSR return task ;

[0083] Step 5.2: Calculate the unit component-level FDA for the ferry, attack, and return stages respectively ijkl ;

[0084] Step 5.3: Initialize MDA ijk , set the current iteration number id and the total number of iterations num;

[0085] Step 5.4: Enter the ferry stage and calculate the subsystem-level MDA ij and σ ij and the single-unit-level MDA i and σ i ;

[0086] Step 5.5: Calculate the MDA and σ of the formation resource requirements;

[0087] Step 5.6: Judge whether σ i is equal to 1. If it is equal, enter Step 5.7; otherwise, enter Step 5.13;

[0088] Step 5.7: Enter the attack stage and recalculate the subsystem-level MDA ij and σ ij and the single-unit-level MDA i and σ i ;

[0089] Step 5.8: Judge whether σ i is equal to 1. If it is equal, enter Step 5.9; otherwise, enter Step 5.13;

[0090] Step 5.9: Enter the return stage and recalculate the subsystem-level MDA ij and σ ij and the single-unit-level MDA i and σ i ;

[0091] Step 5.10: Determine σ i whether it is equal to 1. If it is equal, proceed to Step 5.13 and increment the task success count count by 1;

[0092] Step 5.11: Calculate the task success probability;

[0093] Step 5.12: If the task success probability is less than the sum of the task success rate requirements for each different stage, return to Step 5.2 and adjust the MDA sequentially according to the resource types ijk , and increase the multiple of the resource group required for each adjustment unit component by 1; otherwise, output the MDA ijk and calculate the MDA;

[0094] Step 5.13: Statistically analyze the MDA that meets the requirements to obtain the resource combination and its probability, which is the demand prediction result.

[0095] Therefore, the present invention adopts the above-mentioned method for establishing a hierarchical model of grouped support requirements for tasks, and has the following beneficial effects:

[0096] First, for the problems that the premise assumptions of traditional demand prediction models are that resources are independent of each other, there are certain historical samples, and the upper-level requirements are obtained by summarizing the lower-level requirements, while in actual tasks, resources show a correlation relationship, often with zero or small samples, and the hierarchical requirements are related to the failure probability and system structure. The present invention constructs a hierarchical model of grouped support requirements. This model has few parameters and is easy to operate. Through this model, the actual task requirements can be more accurately described, and the prediction efficiency can be improved;

[0097] Second, according to the model proposed by the present invention, when the task success rate requirement is relatively high, general methods such as adopting a parallel structure and reducing the task duration are required. At the same time, according to different system structures and failure characteristics, the demand for the resource group of relevant unit components is increased; when the task success rate requirement is not high, methods such as adopting a series-parallel structure and adjusting the scale can be used, that is, the guiding conclusions for adjusting the participating unit components, subsystems, equipment, etc. are required, so as to be able to guide the actual demand configuration task.

[0098] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Description of the Drawings

[0099] Figure 1 is a structural relationship diagram of "task - system".

[0100] Figure 2 is a hierarchical structure diagram of equipment support requirements.

[0101] Figure 3 is a basic model of system state.

[0102] Figure 4They are three basic structures (series, parallel, series-parallel) in the system.

[0103] Figure 5 They are the task-related system and component structures.

[0104] Figure 6 They are the comparison results of the task success rates of the case system structure and different guarantee requirement schemes of the parallel system structure.

[0105] Figure 7 They are the comparison results of the task success rates for different guarantee requirements under different failure rates.

[0106] Figure 8 They are the comparison results of the task success rates under different scales.

[0107] Figure 9 They are the comparison results of the task success rates at different task times.

[0108] Figure 10 They are the comparison results of the task success rates under different resource groups.

[0109] Figure 11 They are the distribution of guarantee requirement schemes in 11 different situations. Specific implementation manners

[0110] In the description of the present invention, it should also be noted that unless otherwise clearly specified and limited, these embodiments are only used to illustrate the present invention and not to limit the scope of the present invention. In addition, it should be understood that after reading the content taught by the present invention, those skilled in the art make various changes or modifications to the present invention, and these equivalent forms also fall within the scope defined by the appended claims of this application.

[0111] The present invention believes that the equipment guarantee requirements oriented to tasks have the characteristics of grouping and stratification. Resources have the characteristic of grouping. For example, in a certain task, when a pressure sensor fails, two spare parts, namely a pressure sensor and a gasket, are required. When there is only a single spare part, the failure cannot be repaired, and the single spare part has no practical significance. Resources have the characteristic of hierarchy. For example, the known failure probability and resource requirements at the lower-level unit component level, and the resource requirements at the upper-level subsystem level are the comprehensive results of the summary of the lower-level resource requirements and the subsystem failure probability. The failure probabilities at each level are different, and their resource requirements are also different. At the same time, the system structures at each level are different, and the ways of summarizing resource requirements are also different. For example, if the lower level is a series structure, the upper-level requirements are the summary of all series component requirements; if the lower level is a parallel structure, the upper-level requirements are the summary of the minimum working parallel component requirements. Based on grouping and stratification, the result of guarantee requirement prediction should be the guarantee resources of a piece of equipment or a set of equipment or a formation;

[0112] 1. Construction of the grouped guarantee requirement hierarchical model

[0113] (1) Problem description of equipment support requirements for tasks

[0114] The state of equipment entities has multi-layer network characteristics, and the influence relationships between entity states at each level can be described by the idea of system evolution.

[0115]

[0116] Equation (1) is used to represent the change of entity state, and i, j, and k respectively represent entities, such as formations, single equipment, subsystems, unit components, etc. Among them, represents the state of entities i and j at the α-th layer, represents the state of entity k at the β-th layer, and this state refers to the entity availability; represents the relationship strength between entities i and j at the α-th layer, which is the association relationship between entities under the influence of horizontal combat task relationships and corresponds to task modeling; represents the relationship strength between entities i and k at the β→α layer, which is the association relationship between entities under the influence of vertical equipment function relationships. Here, it is the functional structure relationship and corresponds to the adjacency matrix of i;

[0117]

[0118] Equation (2) is used to represent the dynamic change relationship strength of interactions. It can be seen from Equation (2) that the horizontal association relationship between entities is closely related to the entity state.

[0119] The above two formulas only represent the change relationship and are not differentials in the mathematical sense.

[0120] In Equation (1), Among them, fda i represents a i resource status, and qc, ce, tr, es, sc, ac respectively represent a i related quality characteristics, combat equipment, task requirements, equipment status, system status, and environmental conditions. This invention mainly focuses on the influence of fda i on , and both fda i and are specifically manifested in four levels: unit components, subsystems, single equipment, and formations. Qc, ce, tr, es, sc, ac are all used as constraint conditions, specifically reflected in failure rate, equipment structure, mission success rate, functional status, resource quantity, and mission phase differences, etc.

[0121] Next, since the combat mission is divided into multiple stages, the required functional relationships in each stage determine which equipment systems are involved, and the required structural relationships of each equipment determine which components are involved. Thus, after the combat mission is clarified, when and how each equipment and component participate are also clarified. Therefore, construct the "mission-system" structural relationship of the equipment, as Figure 1 shown, the mathematical expression of this "mission-system" structural relationship is formulas (1)-(2).

[0122] (2) Hierarchical Modeling of Equipment Support Requirements

[0123] Based on the above-mentioned equipment support requirement problem constructed for the mission-oriented equipment support requirements, the hierarchical structure of equipment support requirements is as Figure 2 shown, where A represents a certain equipment formation in the area, A i represents the i-th equipment in the formation, A ij represents the j-th subsystem of equipment A i , and A ijk represents the k-th unit component of equipment subsystem A ij . The equipment support requirements consist of the required quantity and its probability. Within a unit time, the support requirements of the upper-level entity are obtained by multiplying the demand quantity vector of the lower-level associated entity by the failure probability of the upper-level entity. The specific steps for hierarchical modeling of equipment support requirements include:

[0124] ① Hierarchical Requirement Modeling of Unit Components

[0125] At the unit component level, FDA ijk represents the required quantity of resources for A ijk to restore its functional state, and FDA ijk = λA ijk *t*FNum ijk , where λA ijk represents the failure probability of A ijk , and FNum ijk represents the resource vector required for a single failure of unit component A ijk .

[0126] m is the type of personnel resource, M is the total number of personnel resource types, n is the type of spare part resource, N is the total number of spare part resource types. Here represents the required quantity of the m-th type of personnel resource for the k-th unit component of the j-th subsystem of the i-th equipment in a certain formation, and the spare part resource represents the required quantity of the n-th type of personnel resource for the k-th unit component of the j-th subsystem of the i-th equipment in a certain formation;

[0127] Denote PFDA ijk as representing A ijkThe number of personnel resources to restore the functional state, SFDA ijk Denote A ijk The number of spare part resources to restore the functional state. MDA ijk Denote A ijk The number of resource requirements to meet the task requirements, that is, the decision variable to be solved at this level. MDA ijk ={PMDA ijk , SMDA ijk}, where PMDA ijk is the demand vector of the required personnel resources, and SMDA ijk is the demand vector of the required spare part resources.

[0128] ② Sub - system level requirement modeling

[0129] At the sub - system level, FDA ij Denote A ij The number of resource requirements to restore the functional state, FDA ij ={PFDA ij , SFDA ij}, PFDA ij Denote A ij The number of personnel resources requirements to restore the functional state, SFDA ij Denote A ij The number of spare part resources requirements to restore the functional state. Personnel and spare parts are of the same professional category within the sub - system. When summarizing the requirements upwards, the types of personnel do not increase, only the quantity increases, and the types of spare parts increase with the quantity accumulated. Resource grouping is reflected at the unit component level, and the total demand is composed of the demand packages of unit components as the basic demand combination.

[0130] MDA ij Denote A ij The number of resource requirements to meet the task requirements, that is, the decision variable to be solved at this level. MDA ij ={PMDA ij , SMDA ij}, where PMDA ij is the demand vector of the required personnel resources, and SMDA ij is the demand vector of the required spare part resources.

[0131] MDA ij =λA ij *t*∑ k α k MDA ijk , where, λA ij Denote A ij The failure probability of, α k Denote the weight of the k - th unit component.

[0132] λA in the formula ij is calculated as follows: Taking three unit components as an example, their failure probabilities are λ1, λ2, and λ3 respectively, and λ is the failure probability of the system composed of them. In the parallel structure, λ = λ1 * λ2 * λ3; in the series structure, λ = 1 - (1 - λ1) * (1 - λ2) * (1 - λ3); in the series-parallel structure (λ2 and λ3 are in parallel and then in series with λ1), λ = 1 - (1 - λ1) * (1 - λ2 * λ3). Multiple unit components are calculated in this way successively, and the calculation method of the failure probability λ in the subsequent ③ equipment-level requirement modeling is the same as this.

[0133] Based on the above unit component layer A ijk The resource requirement quantity FDA for restoring the functional state ijk to obtain FDA ij :

[0134]

[0135] k = 1, 2, … K

[0136] ③ Equipment-level requirement modeling

[0137] At the single-equipment level, FDA i represents the resource requirement quantity for restoring the functional state of A i FDA i = {PFDA i , SFDA i}, where PFDA i represents the personnel resource requirement quantity for restoring the functional state of A i , and SFDA i represents the spare part resource requirement quantity for restoring the functional state of Ai. From the subsystem to the single equipment, each subsystem is of a different specialty, the types of personnel increase, the quantity accumulates, the types of spare parts increase, and the quantity accumulates.

[0138] MDA i represents the resource requirement quantity for A i to achieve the task requirements, that is, the decision variable to be solved at this level. MDA i = {PMDA i , SMDA i}, where PMDA i is the requirement vector of the required personnel resources, and SMDA i is the requirement vector of the required spare part resources.

[0139] MDA i = λA i * t * ∑ j β j MDA ij where, λA i represents Ai The failure probability, β j represents the weight of the j-th subsystem.

[0140]

[0141] SFDA i ={SFDA i1 ,…,SFDA ij ,…SFDA iJ}

[0142] j = 1, 2, … J

[0143] ④ Modeling of the requirements at the grouping level

[0144] At the grouping level, FDA represents the quantity of resource requirements for A to restore its functional state, FDA = {PFDA, SFDA}, where PFDA represents the quantity of personnel resource requirements for A to restore its functional state, and SFDA represents the quantity of spare part resource requirements for A to restore its functional state. From a single unit upwards to the grouping, among the personnel specialties and the quantities of spare parts in each single unit, there are both the same departments and different departments. At this time, the quantities of the same specialties are accumulated, the types of different specialties increase, and the quantities are accumulated. The quantities of the same spare parts are accumulated, the types of different spare parts increase, and the quantities are accumulated.

[0145] MDA represents the quantity of resource requirements for A to meet the task requirements, that is, the decision variable to be solved at this level. MDA = {PMDA, SMDA}, where PMDA is the demand vector for the required personnel resources, and SMDA is the demand vector for the required spare part resources.

[0146] MDA = λA * t * ∑ i γ i MDA i , where λA represents the failure probability of A, and γ i represents the weight of the i-th equipment.

[0147] The present invention only considers the following situations:

[0148] Based on the resource requirements of a single unit, considering the grouping and packaging of single-unit resources, calculate the resource requirements of the grouping:

[0149] PFDA = {PFDA1, … PFDA i ,… PFDA I}

[0150] SFDA = {SFDA1, … SFDA i ,… SFDA I}

[0151] i = 1, 2, … I.

[0152] (3) Modeling of the mission success probability

[0153] The system state is usually determined by the performance distribution and structure function of all internal components. Before obtaining the system performance distribution, it is often necessary to obtain the performance distribution of each component. The task success state of a complex system is determined by the available states and structure function of all unit components, subsystems, and single installations in the complex hierarchical structure. As Figure 3 shown in the system, it can be described as follows: The system consists of b1,…b n ,…b N The functional relationship of b1,…b n ,…b N is represented by the adjacency matrix S, and σ n is the available state of b n , σ is the success state of S, σ n ={1,0}, n = 1,2…,N, and we can get: σ = [σ1,…σ n ,…σ N *S*[σ1,…σ n ,…σ N T . When σ>0, σ = 1. Taking the example of a system with three elements, S is defined as follows: In series: S = [0,1,0; 0,0,1; 0,0,0]; in parallel: S = [1,0,0; 0,1,0; 0,0,1]; in series-parallel: S = [0,1,1; 0,0,0; 0,0,0], and its structure is as Figure 4 shown. In the present invention, N≤3 is taken. When N>3, it is split into combinations within 3 and then calculated according to this.

[0154] Based on this, the success rates of unit components, subsystems, equipment, formations, and multi-stage tasks are calculated respectively. Among them, σ ijk , σ ij , σ i , σ respectively represent the states of A ijk , A ij , A i , A, that is, the task availability. At the single installation and formation levels, the task availability is the task success rate.

[0155] It can be seen from this that for unit components, subsystems, equipment, formations, and multi-stage tasks, it is necessary to construct the following task success probability models respectively:

[0156] ① Modeling of the task success probability of unit components

[0157] Construct a model of the relationship between resources and states at the unit component level:

[0158]

[0159] Among them, length(MDA ijk -FDA​ijk ≥ 0) is to count the number of elements in MDA ijk -FDA ijk that are greater than or equal to 0. ndims(FDA ijk ) is the dimension of FDA ijk .

[0160] ② Subsystem mission success probability modeling

[0161] Construct a relationship model between subsystem-level resources and states:

[0162] σ ij = [σ ij1 , … σ ijk , … σ ijK * SA ij * [σ ij1 , … σ ijk , … σ ijK T

[0163] Among them, [σ ij1 , … σ ijK , … σ ijK is the state row vector of A ijk , [σ ij1 , … σ ijk , … σ ijK T is the state column vector of A ijk , SA ij is the adjacency matrix of the A ij structure, that is, the association relationship of A ij in A ijk . If there is a relationship, it is represented by 1; if there is no relationship, it is represented by 0.

[0164] ③ Equipment mission success probability modeling

[0165] Construct a relationship model between a single equipment's resources and states:

[0166] σ i = [σ i1 , … σ ij , … σ iJ * SA i * [σ i1 , … σ ij , … σ iJ T

[0167] Among them, [σ i1 , … σ ij , … σi J] is the state row vector of A ij , [σ i1 , … σ ij,…σ iJ T is the state column vector of A, SA ij for A i is the adjacency matrix of the A i structure, that is, the A in A i in A ij correlation relationship. It is 1 when there is a relationship and 0 when there is no relationship.

[0168] ④ Modeling of the success probability of the grouping task

[0169] Construct a model of the relationship between equipment grouping resources and states:

[0170] σ = [σ1,…σ i ,…σ I *SA*[σ1,…σ i ,…σ I T

[0171] i = 1,2,…I

[0172] Among them, [σ1,…σ i ,…σ I is the state row vector of A i , [σ1,…σ i ,…σ I T is the state column vector of A i for A i SA is the adjacency matrix of the A structure, that is, the A in A i correlation relationship. It is represented by 1 when there is a relationship and 0 when there is no relationship.

[0173] ⑤ Modeling of the success probability of multi-stage tasks

[0174] The present invention only considers the task success rate under a determined combat mission. When the mission is determined, the equipment system participating in the mission is determined, and the system structure is also determined. Assuming that the tasks in the navigation, attack, and return stages are all successful, the total mission is considered successful, that is, there is σ totaltask = σ ferrytask *σ attacktask *σ returntask , σ totaltask = {1,0}, 1 indicates success and 0 indicates failure. σ ferry task , σ attack task , σ return task , σ total task respectively represent the mission success rates of the navigation stage, attack stage, return stage, and overall.

[0175] 2. Support requirement prediction algorithm

[0176] (1) Single-equipment level support requirement prediction algorithm​​​

[0177] Verify whether the required mission success rate is achieved according to the possible combination requirement scenarios, and conduct statistical analysis on all the requirement scenarios that meet the requirements, that is, obtain the safeguard requirements. When the number of unit components is large, considering the number of resource types again, the requirement scenarios will have a combinatorial explosion. When calculating the combinations according to resource groups, the computational workload of the safeguard requirements is greatly reduced.

[0178] Algorithm 1: Prediction of Single-Equipment Grouping Support Requirements

[0179]

[0180]

[0181] As can be seen from the above, the prediction algorithm for single-equipment grouping support requirements includes the following steps:

[0182] Step 1: Obtain the failure probabilities ferrytask_λA ijk in the ferry stage, attack stage, and return stage of unit component A ijk , attacktask_λA ijk , returntask_λA ijk ; the resource vectors ferrytask_FNum ijk required for the l-th single failure of unit component A ijkl , attacktask_FNum ijkl , returntask_FNum ijkl in the ferry stage, attack stage, and return stage; the adjacency matrix SA ij of each hierarchical system structure, SA i , SA, the overall mission success rate requirement TSR total task and the mission success rate requirements TSR ferry task , TSR attack task , TSR return task ;

[0183] Step 2: Initialize MDA ijk , set the current iteration number id and the total number of iterations num;

[0184] Step 3: Calculate the unit component-level FDA ijk in the ferry, attack, and return stages respectively;

[0185] Step 4: Calculate the subsystem-level MDA ij and σ ij and the single-equipment-level MDA i and σ i in the ferry stage respectively;

[0186] Step 5: Determine σ i whether it is equal to 1. If it is equal, go to Step 6; otherwise, go to Step 12;

[0187] Step 6: Enter the attack phase and recalculate the subsystem-level MDA ij and σ ij and the single-unit-level MDA i and σ i ;

[0188] Step 7: Determine σ i whether it is equal to 1. If it is equal, go to Step 8; otherwise, go to Step 12;

[0189] Step 8: Enter the return phase and recalculate the subsystem-level MDA ij and σ ij and the single-unit-level MDA i and σ i ;

[0190] Step 9: Determine σ i whether it is equal to 1. If it is equal, go to Step 12 and increment the task success count by 1;

[0191] Step 10: Calculate the task success probability;

[0192] Step 11: If the task success probability is less than the sum of the task success rate requirements for each different phase, return to Step 2 and adjust the MDA in sequence according to the resource types ijk , increasing the multiple of the resource group required for each adjustment unit component by 1; otherwise, output the MDA ijk and calculate the MDA i ;

[0193] Step 12: Statistically analyze the MDA that meets the requirements i , obtaining the resource combination and its probability, which is the demand prediction result.

[0194] (2) Prediction Algorithm for the Guarantee Requirements at the Grouping Level

[0195] Algorithm 2: Prediction of the Guarantee Requirements for Grouping

[0196]

[0197] As can be seen from the above, the prediction algorithm for the guarantee requirements at the grouping level is basically the same as the prediction algorithm for the guarantee requirements at the single-unit grouping level. The differences are as follows: First, Step 6 for calculating the grouping-level MDA and σ is added after Step 5; second, the last step statistically analyzes the MDA that meets the requirements, rather than the MDA i .

[0198] The specific steps of this prediction algorithm for the guarantee requirements at the grouping level are as follows:

[0199] Step 1: Obtain unit component A respectively ijk The failure probabilities ferrytask_λA ijk , attacktask_λA ijk , returntask_λA ijk in the ferry stage, attack stage, and return stage; The resource vector ferrytask_FNum ijk required for the l-th single failure of unit component A ijkl , attacktask_FNum ijkl , returntask_FNum ijkl in the ferry stage, attack stage, and return stage; The adjacency matrix SA of each hierarchical system structure ij , SA i , SA, the overall mission success rate requirement TSR total task and the mission success rate requirements TSR ferry task , TSR attack task , TSR return task ;

[0200] Step 2: Calculate the unit component-level FDA ijk in the ferry, attack, and return stages respectively;

[0201] Step 3: Initialize MDA ijk , and set the current iteration number id and the total number of iterations num;

[0202] Step 4: Enter the ferry stage and calculate the subsystem-level MDA ij and σ ij and the single-unit-level MDA i and σ i ;

[0203] Step 5: Consider the requirements at the formation level and calculate the formation-level MDA and σ;

[0204] Step 6: Judge whether σ i is equal to 1. If it is equal, enter Step 7; otherwise, enter Step 13;

[0205] Step 7: Enter the attack stage and recalculate the subsystem-level MDA ij and σ ij and the single-unit-level MDA i and σ i ;

[0206] Step 8: Judge whether σ i is equal to 1. If it is equal, enter Step 9; otherwise, enter Step 13;

[0207] Step 9: Enter the return phase and recalculate the subsystem-level MDA ij and σ ij and the single-unit-level MDA i and σ i ;

[0208] Step 10: Determine whether σ i is equal to 1. If it is, enter Step 13 and increment the task success count by 1;

[0209] Step 11: Calculate the task success probability;

[0210] Step 12: If the task success probability is less than the sum of the task success rate requirements for each different stage, return to Step 2 and adjust the MDA successively according to the resource types, ijk with the multiple of the required resource group for each adjustment of the unit component increased by 1; otherwise, output the MDA ijk and calculate the MDA;

[0211] Step 13: Statistically analyze the MDA that meets the requirements to obtain the resource combination and its probability, which is the demand prediction result.

[0212] Embodiment

[0213] The present invention proposes a method for establishing a hierarchical model of group support requirements for tasks. Taking the task of predicting support requirements for a certain task as an example, the method proposed by the present invention is verified. Specifically, in a certain task, a certain formation has 3 types of equipment, and the task-related systems and component structures of each equipment are as Figure 5 shown. The task phase division and the working subsystems of each phase are shown in Table 1, the unit component-level failure probabilities and resource requirements are shown in Tables 2 - 4, the adjacency matrices of the system structures at each level are shown in Tables 5 - 6, and the group-level group support requirements are calculated when the task success rate in the attack phase (1 hour) is not less than 99.8%.

[0214] Table 1 Task Phases and Working Subsystems

[0215]

[0216] Table 2 Unit Component-Level Failure Probabilities

[0217]

[0218] Table 3 Personnel Requirements for Unit Component-Level Failures

[0219]

[0220] Table 4 Spare Part Requirements for Unit Component-Level Failures

[0221]

[0222] Table 5 Sub-system level structure adjacency matrix

[0223]

[0224] Table 6 Single unit and formation level structure adjacency matrix

[0225]

[0226] In this embodiment, only the calculation results of the support resource requirements at the formation level are given. As shown in Table 7 and Table 8, there is a one-to-one correspondence between the personnel group and the spare parts group.

[0227] Table 7 Calculation results of formation level personnel requirements

[0228]

[0229] Table 8 Calculation results of formation level spare parts requirements

[0230]

[0231] Based on the resource group required when the unit component fails, 63 demand combination schemes are taken, and the mission success rate is as Figure 6 shown. The average mission success rate when the resources required by the unit component correspond to different multiples of the resource group is shown in Table 9. When the support demand is Scheme 1 - 15, the resources required by A11, A21, and A31 are 0 times the corresponding resource group, and the mission success rate is 0; when the support demand is Scheme 16 - 31, the resources required by A11, A21, and A31 are 1 time the corresponding resource group, and the average mission success rate is 33.53%; when the support demand is Scheme 32 - 47, the resources required by A11, A21, and A31 are 2 times the corresponding resource group, and the average mission success rate is 41.32%; when the support demand is Scheme 48 - 63, the resources required by A11, A21, and A31 are 3 times the corresponding resource group, and the average mission success rate is 52.83%. As the resources required by A11, A21, and A31 corresponding to the resource group increase, the impact on the mission success rate gradually increases and the effect is obvious, indicating that the more resources they require, the better. When the resources required by A11, A21, A31 and A13, A23, A33 are 0 times the corresponding resource group, the mission success rate is 0, indicating that the resources they require must be available. When the resources required by A13, A23, A33 correspond to 1 - 3 times the resource group, the change in the mission success rate is not obvious, indicating that the resources they require mainly focus on solving the problem of availability. When the resources required by A12, A22, A32 correspond to 0 - 3 times the resource group, the average mission success rate is relatively stable and the change is small, indicating that the resources they require are related to the mission success rate requirement. When the mission success rate requirement is high, the demand is high; when the mission success rate requirement is low, the demand is low.

[0232] Table 9 Average mission success rate when the resource group is at different multiples

[0233] 0 times 1 time 2 times 3 times A11, A21, A31 0 33.53％ 41.32％ 52.83％ A12, A22, A32 35.25％ 25.89％ 23.39％ 30.13％ A13, A23, A33 0 45.26％ 38.61％ 47.01％

[0234] The experimental results of the above cases are analyzed from the following aspects:

[0235] (1) Influence of system structure

[0236] The mission success rates of the case system structure and different safeguard requirement solutions of the parallel system structure are as Figure 6 shown, indicating that the more redundant the system structure is, the lower the safeguard requirement is.

[0237] (2) Influence of failure rate

[0238] The mission success rates of different safeguard requirements under the case failure rate, full normal distribution, full exponential distribution, and full uniform distribution failure rates are as Figure 7 shown. It shows that in this case, the influences of A11, A21, A31 and A13, A23, A33 have nothing to do with the failure rate, and different failure rates have little influence on the safeguard requirement solutions.

[0239] (3) Influence of scale

[0240] In the case, the number of single installations involved is 3, the number of subsystems involved in each single installation is 3, and the number of unit components involved in each subsystem is 3, which is defined as small scale (3-3-3). Based on this, the medium scale (3-6-3) is defined. The number of subsystems involved is doubled, and the parameters are the same as those of the small scale. The subsystem structure is parallel (since only the structural expressions of less than 3 unit components are given in the text, the parallel structure that is easy to calculate is selected here). Furthermore, the larger scale (3-6-6) is defined. The number of single installations involved is doubled, and the parameters are the same as those of the medium scale. The single installation structure is series-parallel, as shown in Table 10. The corresponding calculation results are as Figure 8 shown. Generally speaking, as the scale increases, the number of safeguard requirement solutions that meet the requirements increases. In the medium scale, the mission success rates corresponding to the safeguard requirement solutions 7-16 are not zero because the structure of the subsystems involved is a parallel structure.

[0241] Table 10 Definition of three scales

[0242]

[0243]

[0244] (4) Influence of mission time

[0245] When the mission time increases, the number of failures increases, and the support requirements increase accordingly. Taking the attack phase as an example, when the mission durations are 1h, 5h, and 10h respectively, the support requirement solutions that meet the requirements shift to the right, that is, the requirements increase. The mission success rate impact curves of the requirement solutions in the two cases of 5h and 10h are the same, but the amplitude decreases, indicating that after the mission lasts for a certain period of time, the impact characteristics of the same requirement solution on the mission success rate are the same qualitatively but different quantitatively. The corresponding calculation results are as Figure 9 shown.

[0246] (5) Impact of the resource group on the mission success rate

[0247] The impact of the resource group on the mission success rate is consistent in the attack phase and the entire mission phase. The impacts of A11, A21, A31 and A13, A23, A33 are the same as the previous analysis. In the results, the mission success rate of multiple phases increases significantly because: in the case, the number of participating unit components and subsystems in the ferry phase and the return phase is small. For the same support requirement solution, the mission success rates of these two phases are high. From the perspective of the entire mission phase, at the same success rate, the support requirements in the attack phase decrease. Therefore, under the same support requirement solution, the mission success rate of multiple phases is higher than that of the single attack phase. The corresponding calculation results are as Figure 10 shown.

[0248] (6) Comprehensive analysis

[0249] Considering 11 cases, as shown in Table 11, the calculation results are as Figure 11 shown. The results show that: (1) The mission success rate is closely related to the failure probability, system structure, system scale, mission time, etc. The impact characteristic of the failure rate is that the uniform distribution is large and other distributions are small. The impact characteristic of the system structure is that the parallel connection has a large impact and the series-parallel connection has a small impact. The impact characteristic of the system scale is that the structure has a large impact and the scale has a small impact. The impact characteristic of the mission time is that it is large in the initial stage and small in the later stage; (2) The support requirement solution is closely related to the failure probability, system structure, system scale, mission time, etc. The impact characteristic of the failure probability is that the support requirements are small for the mixed distribution and the uniform distribution. The impact characteristic of the system structure is that the requirements are stable under the parallel system structure. The impact characteristic of the system scale is that the requirements fluctuate little under different scales. The impact characteristic of the mission time is that the longer the mission duration, the greater the fluctuation of the support requirements.

[0250] Table 11 Definition of various cases

[0251]

[0252] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions of the present invention or make equivalent replacements, and these modifications or equivalent replacements cannot make the modified technical solutions deviate from the spirit and scope of the technical solutions of the present invention.

[0253] References:

[0254] [1] Zhang Jian, Zhao Kunhai, Wu Xiuli, et al. Research on the demand prediction method of wind power spare parts based on small sample data [J]. Development & Innovation of Machinery & Electrical Products, 2024, 37(02): 14-18.

[0255] [2] Zhang Cheng, Ma Ziwei, Liu Bin, et al. Remaining life prediction of small sample rotating machinery driven by digital twin [J]. Journal of Xi'an Jiaotong University, 2023, 57(12): 168-178.

[0256] [3] Wang Tiening, Wu Longtao, Yang Fan. Demand prediction of timed truncated small sample exponential equipment based on Bayesian method [J]. Journal of Academy of Armored Force Engineering, 2017, 31(04): 29-34.

[0257] [4] Liu Junfu, Cen Jian, Huang Hankun, et al. Review of zero small sample rotating machinery fault diagnosis [J]. Computer Engineering and Applications, 1-14.

[0258] [5] Song Jie, Chen Zhenyu, Yang Yang, et al. Research on load demand response decision-making considering resource correlation and uncertainty [J]. Electric Power Construction, 2019, 40(06): 132-138.

[0259] [6] Chen Yang, Xu Xiaoshuang, Zhao Liangliang, et al. A method for predicting the demand of aviation ammunition in wartime [J]. Journal of Army Engineering University, 2022, 1(02): 80-86.

[0260] [7] Hyndman, R.J., & Athanasopoulos, G. (2021) Forecasting: principles and practice, 3rd edition, OTexts: Melbourne, Australia.

[0261] [8]Wickramasuriya, S.L. (2024). Probabilistic Forecast Reconciliation under the Gaussian Framework. Journal of Business & Economic Statistics, 42(1), 272–285.

[0262] https: / / doi.org / 10.1080 / 07350015.2023.2181176.

Claims

1. A method for establishing a hierarchical model of group support requirements oriented to tasks, characterized in that, It includes the following steps: Step 1: Divide the task into multiple stages, and define the equipment support requirement problems oriented to the task according to the structural relationships of the equipment and components in different stages; Step 2: For the equipment support requirement problems oriented to the task, determine the hierarchical structure of equipment support requirements, and based on this hierarchical structure, construct a hierarchical model of equipment support requirements; Step 3: Based on the hierarchical model of equipment support requirements, construct a task success rate model; Step 4: Use the single-equipment hierarchical support requirement prediction algorithm to solve the task success rate model, and screen out the resource combinations that meet the requirements and their probabilities; Step 5: Use the formation hierarchical support requirement prediction algorithm to solve the task success rate model, and screen out the resource combinations that meet the requirements and their probabilities; Step 6: Conduct statistical analysis on all the demand solutions obtained in S4 and S5 that meet the requirements, and finally obtain the resource solution for support requirements.

2. The method for establishing a task-oriented hierarchical model of grouped support requirements according to claim 1, characterized in that, The equipment support requirement problems oriented to the task defined in Step 1 are: 1) Changes in entity states Adopt the idea of system evolution to describe the influence relationships between the entity states at all levels of the equipment entity: where i, j, and k represent different entities; respectively represent the states of entity i and entity j at the α-th layer, represents the state of entity k at the β-th layer; represents the relationship strength between entities i and j at the α-th layer; represents the relationship strength between entities i and k at the β→α layer; In formula (1) wherein, fda i represents a i resource status, and qc, ce, tr, es, sc, and ac respectively represent a i related quality characteristics, combat equipment, mission requirements, equipment status, system status, and environmental conditions; 2) The intensity of the dynamic change relationship of the interaction between different entities 3. The method for establishing a task-oriented hierarchical model of grouped support requirements according to claim 1, wherein, The hierarchical structure of equipment support requirements described in Step 2 is divided into an equipment formation layer, an equipment layer, a subsystem layer, and a unit component layer from top to bottom. Among them, A is defined as a certain equipment formation in the region, and A i is defined as the i-th equipment in the formation, and A ij is defined as equipment A i 's j-th subsystem, and A ijk is defined as the k-th unit component of equipment subsystem A ij .

4. The method for establishing a task-oriented hierarchical model of grouped support requirements according to claim 2, characterized in that When constructing the hierarchical model of equipment support requirements in Step 2, based on the hierarchical structure of equipment support requirements, demand modeling is carried out successively from bottom to top. The specific steps include: Step 2.1: Demand modeling at the unit component level; Step 2.1.1: In the unit component layer, define FDA ijk as A ijk to be the number of resource requirements for restoring the functional state, and FDA ijk = λA ijk *t*FNum ijk , where λA ijk represents the failure probability of A ijk , FNum ijk represents the resource vector required for a single failure of unit component A ijk ; among them, Where m is the type of human resource, M is the total number of human resource types, n is the type of spare part resource, and N is the total number of spare part resource types. represents the required quantity of the m-th type of human resource for the k-th unit component of the j-th subsystem of the i-th equipment in a certain group. represents the required quantity of the n-th type of human resource for the k-th unit component of the j-th subsystem of the i-th equipment in a certain group; Step 2.1.2: Define PFDA ijk as A ijk which is the number of personnel resources to restore the functional state, and define SFDA ijk as A ijk which is the number of spare part resources to restore the functional state, and Step 2.1.3: Define MDA ijk as A ijk to achieve the resource requirement quantity of the task requirement, that is, the decision variable to be solved at this level, and MDA ijk ={PMDA ijk , SMDA ijk}, where PMDA ijk is the demand vector of the required personnel resources, and SMDA ijk is the demand vector of the required spare part resources, and SMDA ijk is the demand vector of the required spare part resources; Step 2.2: Demand modeling at the subsystem level; Step 2.2.1: At the subsystem level, represent the FDA ij as A ij The number of resource requirements to restore the functional state, FDA ij ={PFDA ij , SFDA ij}, where PFDA ij represents the number of personnel resource requirements for A ij to restore the functional state, and SFDA ij represents the number of spare part resource requirements for A ij to restore the functional state, and: Step 2.2.2: Represent MDA ij as A ij to represent the quantity of resource requirements for task requirements, i.e., the decision variable to be solved at this level, and MDA ij ={PMDA ij , SMDA ij}, where PMDA ij is the requirement vector of required personnel resources, and SMDA ij is the requirement vector of required spare part resources; Step 2.2.3: Calculate MDA using the following formula ij :[[]]END]] MDA ij = λA ij *t*∑ k α k MDA ijk Among them, λA ij represents the failure probability of A ij , and α k represents the weight of the k-th unit component; Step 2.3: Demand modeling at the equipment level; Step 2.3.1: In the single installation layer, the FDA i is represented as A i The quantity of resource requirements for restoring the functional state, FDA i ={PFDA i , SFDA i}, where PFDA i represents the quantity of personnel resource requirements for restoring the functional state of A i , SFDA i represents the quantity of spare part resource requirements for restoring the functional state of A i and: Step 2.3.2: Express MDA i as A i to achieve the quantity of resource requirements for the task requirements, that is, the decision variable to be solved at this level, and MDA i ={PMDA i , SMDA i}, where PMDA i is the demand vector for the required personnel resources, and SMDA i is the demand vector for the required spare parts resources; Step 2.3.3: Calculate MDA using the following formula i :[[]]END]] MDA i = λA i *t*Σ j β j MDA ij Among them, λA i represents the failure probability of A i , and β j represents the weight of the j-th subsystem; Step 2.4: Demand modeling at the formation level; Step 2.4.1: At the formation level, represent FDA as the quantity of resource requirements for A to restore its functional state, FDA = {PFDA, SFDA}, where PFDA represents the quantity of personnel resource requirements for A to restore its functional state, and SFDA represents the quantity of spare part resource requirements for A to restore its functional state, and: PFDA = {PFDA1, … PFDA i , … PFDA I} SFDA = {SFDA1, … SFDA i , … SFDA I} i = 1, 2,... I; Step 2.4.2: Represent MDA as the quantity of resource requirements for A to meet the task requirements, that is, the decision variable to be solved at this level, and MDA = {PMDA, SMDA}, where PMDA is the demand vector of the required personnel resources, and SMDA is the demand vector of the required spare part resources; Step 2.4.3: Calculate MDA through the following formula: MDA = λA * t * ∑ i γ i MDA i Among them, λA represents the failure probability of A, and γ i represents the weight of the i-th equipment.

5. The method for establishing a task-oriented hierarchical model of grouped support requirements according to claim 3, wherein In Step 3, task success probability modeling is constructed for unit components, subsystems, equipment, formations, and multi-stage tasks respectively, to calculate the success rates of unit components, subsystems, equipment, formations, and multi-stage tasks respectively. Specifically, it includes: 1) Resource and state relationship model at the unit component level where, length(MDA ijk -FDA ijk ≥0) is the number of elements in MDA ijk -FDA ijk that are greater than or equal to 0, and ndims(FDA ijk ) is the dimension of FDA ijk ; 2) Resource and state relationship model at the subsystem level σ ij = [σ ij1 ,…σ ijk ,…σ ijK *SA ij *[σ ij1 ,…σ ijk ,…σ ijK T ​ Among them, [σ ij1 , … σ ijk , … σ ijK is the state row vector of A ijk , [σ ij1 , … σ ijk , … σ ijK T is the state column vector of A ijk , SA ij is the adjacency matrix of the structure of A ij , that is, the association relationship of A ij in A ijk . If there is an association, it is represented by 1, and if there is no association, it is represented by 0;​ 3) Resource and state relationship model of a single piece of equipment σ i = [σ i1 ,…σ ij ,…σ iJ * SA i * [σ i1 ,…σ ij ,…σ iJ T ​ Among them, [σ i1 ,…σ ij ,…σ iJ is the state row vector of A ij , [σ i1 ,…σ ij ,…σ iJ T is the state column vector of A ij , SA i is the adjacency matrix of the structure of A i , that is, the association relationship of A i in A ij , it is 1 when there is a relationship and 0 when there is no relationship;​ 4) Resource and state relationship model of equipment formations σ = [σ1, … σ i , … σ I *SA*[σ1, … σ i , … σ I T ​ i = 1, 2,... I Among them, [σ1,…σ i ,…σ I is the state row vector of A i , and [σ1,…σ i ,…σ I T is the state column vector of A i . SA i is the adjacency matrix of the A structure, that is, the association relationship of A i in A. If there is a relationship, it is represented by 1, and if there is no relationship, it is represented by 0.​ 6. The method for establishing a task-oriented hierarchical model of grouped support requirements according to claim 5, wherein The single-equipment hierarchical support requirement prediction algorithm includes the following steps: Step 4.1: Obtain unit component A respectively ijk The failure probabilities ferrytask_λA ijk , attacktask_λA ijk , returntask_λA ijk during the ferry stage, attack stage, and return stage; ijk The resource vectors ferrytasj_FNum ijkl , attacktask_FNum ijkl , returntask_FNum ijkl required for the l-th single failure during the ferry stage, attack stage, and return stage; ij The adjacency matrices SA of each hierarchical system structure i , SA total task The overall mission success rate requirement TSR ferry task and the mission success rate requirements TSR attack task during the ferry stage, attack stage, and return stage return task ; Step 4.2: Initialize MDA ijk , set the current iteration number id and the total number of iterations num; Step 4.3: Calculate the unit component-level FDA for the navigation, attack, and return phases respectively ijk ; Step 4.4: Calculate the subsystem-level MDA ij and σ ij and the single-unit-level MDA i and σ i ; Step 4.5: Determine whether σ i is equal to 1. If it is, proceed to Step 4.6; otherwise, proceed to Step 4.12; Step 4.6: Enter the attack phase and recalculate the subsystem-level MDA ij and σ ij and the single-warhead-level MDA i and σ i ; Step 4.7: Determine whether σ i is equal to 1. If it is, proceed to Step 4.8; otherwise, proceed to Step 4.

12. Step 4.8: Enter the return flight phase and recalculate the subsystem-level MDA ij and σ ij and the single-unit-level MDA i and σ i ; Step 4.9: Determine whether σ i is equal to 1. If it is, proceed to Step 4.12 and increment the task success count count by 1; Step 4.10: Calculate the task success probability; Step 4.11: If the task success probability is less than the sum of the task success rate requirements for each different stage, return to Step 4.2 and adjust the MDA successively according to the resource types, increasing the multiple of the resource group required for each adjustment unit component by 1; otherwise, output the MDA ijk , increasing the multiple of the resource group required for each adjustment unit component by 1 each time; otherwise, output the MDA ijk and calculate the MDA i ; Step 4.12: Count the MDA that meets the requirements i , and obtain the resource combination and its probability, which is the demand prediction result.

7. The method for establishing a task-oriented hierarchical model of grouped support requirements according to claim 5, characterized in that, The formation hierarchical support requirement prediction algorithm includes the following steps: Step 5.1: Obtain unit component A separately ijk The failure probabilities ferrytask_λA ijk , attacktask_λA ijk , and returntask_λA ijk during the ferry stage, attack stage, and return stage of unit component A ijk The resource vectors ferrytask_FNum ijkl , attacktask_FNum ijkl , and returntask_FNum ijkl required for the l-th single failure during the ferry stage, attack stage, and return stage of each hierarchical system structure ij , SA i , SA, and the overall mission success rate requirement TSR total task as well as the mission success rate requirements TSR ferry task , TSR attack task , TSR return task ; Step 5.2: Calculate the unit component-level FDA for the navigation, attack, and return phases respectively ijkl ; Step 5.3: Initialize MDA ijk , set the current iteration number id and the total number of iterations num; Step 5.4: Enter the navigation stage, and calculate the subsystem-level MDA ij and σ ij and the single-unit-level MDA i and σ i ; Step 5.5: Calculate the formation resource requirements MDA and σ; Step 5.6: Determine whether σ i is equal to 1. If it is, proceed to Step 5.7; otherwise, proceed to Step 5.

13. Step 5.7: Enter the attack phase and recalculate the subsystem-level MDA ij and σ ij and the single-unit-level MDA i and σ i ; Step 5.8: Determine whether σ i is equal to 1. If it is, proceed to Step 5.9; otherwise, proceed to Step 5.

13. Step 5.9: Enter the return flight phase and recalculate the subsystem-level MDA ij and σ ij and the single-unit-level MDA i and σ i ; Step 5.10: Determine whether σ i is equal to 1. If it is, proceed to Step 5.13 and increment the task success count count by 1; Step 5.11: Calculate the task success probability; Step 5.12: If the task success probability is less than the sum of the task success rate requirements for each different stage, return to Step 5.2 and adjust the MDA successively according to the resource types, increasing the multiple of the resource group required for each unit component by 1; otherwise, output the MDA ijk , and calculate the MDA ijk and calculate the MDA; Step 5.13: Statistically analyze the MDA that meets the requirements, and obtain the resource combination and its probability, which is the demand prediction result.

Images