Technical support product system contribution rate prediction method based on BP neural network

Abstract

The invention provides a technical support product system contribution rate prediction method based on a BP neural network, and the method comprises the steps: obtaining a support task, serializing the current support task according to the execution process of the support task, and obtaining a support task sequence, the task atomic models required by all the process steps of the task sequence are retrieved from the task atomic model set in combination with the task atomic models required by all the process steps of the task sequence; matching a technical support product with corresponding technical capability information from a technical support product capability model library according to a technical specialty required by execution of the support task; and inputting the guarantee task, the task atomic model corresponding to the guarantee task, the attribute information of the technical support product, the service life information and the environment information into the trained technical support product system contribution rate prediction model, and outputting the contribution rate of the technical support product to the guarantee system. The artificial intelligence algorithm is introduced to predict the system contribution rate of a novel technology guarantee product, meanwhile, product life data is added, and the problem that adaptability is not high in a conventional prediction method is solved.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of equipment support system, and relates to a technical support product system contribution rate prediction method based on a BP neural network. BACKGROUND

[0002] The most important evaluation index for determining whether a new product needs to be developed is the contribution rate of the product in the entire product system. Technical support vehicles and other products are important means for related users to perform field detection and repair. At present, with the increasing updating of product systems, new technical support product systems also need to be updated to adapt to the product support needs in new field environments. At the same time, many already-assigned technical support products still play an important role in the front line due to the large number of assigned quantities, high personnel proficiency, and good support effectiveness. How to coordinate the relationship between the new and old systems of technical support products and determine whether a new product should be included in the research and development or procurement plan is the most important evaluation index, that is, the contribution degree in the entire system.

[0003] The existing technical solutions for evaluating the contribution rate of equipment support still have the following defects:

[0004] The evaluation and prediction method is fixed: there are many related technologies for predicting or evaluating the contribution rate at present, and the algorithms used include but are not limited to the analytic hierarchy process, fuzzy comprehensive evaluation method, entropy method, and module analysis method. Among them, the analytic hierarchy process is suitable for hierarchical index system, the fuzzy comprehensive analysis method is suitable for comprehensive evaluation of large systems, the module analysis is suitable for software modules, and the entropy method is suitable for evaluation and prediction of evaluation criteria with strong objectivity. It is not suitable for evaluating the contribution rate of single product to collaborative marshalling support effectiveness.

[0005] Less attention from support professionals: current joint operations increasingly exhibit characteristics such as intelligent decision-making, precise execution, three-dimensional deployment, multi-domain implementation, and cross-professional support. The elements, units, and systems are integrated and integrated, which has brought profound changes to the support mechanism of the product system and is mostly based on the output of joint operation effectiveness. However, whether the operation is completed well and whether the support is good are key factors. It is not yet a set of related professional methods to measure the contribution rate of new technical support products in the entire system and the system contribution rate in the entire joint support process, and to locate the short board problem in the entire technical support product marshalling, so as to achieve efficient output of support effectiveness.

[0006] Therefore, how to provide an equipment fault recommendation diagnosis method capable of comprehensively evaluating the collaborative marshalling support effectiveness is a problem that needs to be solved by those skilled in the art. SUMMARY

[0007] Therefore, the application provides a technical support product system contribution rate prediction method based on a BP neural network, introduces an artificial intelligence algorithm to predict the system contribution rate of a new type of technical support product, and additionally adds product life data to solve the problem of low adaptability in conventional prediction methods.

[0008] To achieve the above object, the application adopts the following technical scheme:

[0009] The application discloses a technical support product system contribution rate prediction method based on a BP neural network, comprising the following steps:

[0010] Model training step:

[0011] S1: According to the execution process of different support tasks, the support tasks are divided into a plurality of task atomic models, and a task atomic model set is constructed;

[0012] S2: According to the technical ability information of different technical support products, a technical support product ability model library is constructed, wherein the technical support product is equipment used for executing the support task;

[0013] S3: A technical support product system contribution rate prediction model is constructed based on a BP neural network;

[0014] S4: According to the technical professional required by the support task execution, a technical support product with corresponding technical ability information is matched; the support task, the attribute information, the life information and the environmental information of the task atomic model corresponding to the support task and the technical support product are taken as inputs, and the contribution rate of the technical support product to the support system is taken as output to construct a training sample set, and the technical support product system contribution rate prediction model is trained;

[0015] Model prediction step:

[0016] S5: A current support task is acquired, the current support task is serialized according to the execution process of the current support task, a current support task sequence is obtained, and the task atomic model required by each process step of the current support task sequence is called from the task atomic model set;

[0017] S6: According to the technical professional required by the support task execution, a technical support product with corresponding technical ability information is matched from the technical support product ability model library;

[0018] S7: The current support task, the attribute information, the life information and the environmental information of the task atomic model corresponding to the current support task and the technical support product are input into the trained technical support product system contribution rate prediction model, and the contribution rate of the technical support product to the support system is output.

[0019] Preferably, the attribute information of the technical support product comprises one or more of the following: product name, product number, product type, product function information.

[0020] Preferably, the environment information of the technical support product comprises environment data affecting the execution of the technical support product task, including one or more of the following: terrain, temperature, altitude, humidity, and climate.

[0021] Preferably, the technical support product capability model library is used to store attribute information of the technical support product, and a technical support product capability dictionary table associated with the corresponding technical support product, wherein the capability information in the capability dictionary table is the same as the content of the technical professional information required for matching support task execution, so as to match the technical support product with the corresponding technical capability information according to the technical professional required for support task execution.

[0022] Preferably, the execution flow steps of the current support task sequence correspond to a plurality of task atomic models, the technical professional required by each task atomic model is a plurality of technical support products, and a plurality of technical support products with corresponding technical capability information are matched.

[0023] Preferably, the method further comprises the following steps:

[0024] The current support task sequence corresponds to a plurality of different technical support product combinations, and the trained technical support product system contribution rate prediction model outputs a support system contribution rate list of different technical support product combinations;

[0025] The time consumption of each task atomic model combination is calculated, and the support time of the support task sequence under different task atomic model combinations is obtained through the critical path algorithm, and the optimal marshalling scheme is output through comparative analysis.

[0026] Preferably, the task atomic model set comprises: a general task atomic model, a rescue support task atomic model, a maintenance support task atomic model, a transportation support task atomic model, and an energy support task atomic model.

[0027] According to the above technical solution, compared with the prior art which relies on pre-defined reasoning path and realizes diagnosis through simple search matching, the beneficial effects of the present application include:

[0028] The present application aims at the influence of the single technical support product to be established on the contribution rate of the whole technical support product system, and further extends to the influence analysis of multiple technical support products on the system contribution rate output of the cooperative support organization, on the basis of the existing cooperative support organization system contribution rate prediction method, the product life influence factor of the cooperative support organization system contribution rate is added, and a new type of technical support product contribution rate prediction method based on BP neural network is summarized, so that the support organization support efficiency is effectively exerted, and is used for assisting the cooperative support organization decision output of the support team commander.

[0029] The present application adds the product life influence factor in view of the influence of different life states of different products on the overall effectiveness of support in different environments, further refines the system contribution rate prediction index of the technical support product in the support effect, and further more accurately outputs the related prediction results of the system contribution rate. BRIEF DESCRIPTION OF DRAWINGS

[0030] In order to more clearly illustrate the technical solutions of the embodiments of the present application or the prior art, the drawings needed to be used in the embodiment or the prior art description will be briefly introduced as follows. Obviously, the drawings in the following description are only embodiments of the present application, and other drawings can be obtained by the provided drawings without creative labor for those skilled in the art;

[0031] Figure 1 The technical support product system contribution rate prediction method based on BP neural network provided by the embodiment of the present application is shown in the flowchart;

[0032] Figure 2 The BP neural network algorithm flowchart provided by the embodiment of the present application is shown in the flowchart;

[0033] Figure 3 The environment model organization block diagram provided by the embodiment of the present application is shown in the block diagram;

[0034] Figure 4 The technical support product model organization block diagram provided by the embodiment of the present application is shown in the block diagram;

[0035] Figure 5 The rescue support task atomic model business flowchart provided by the embodiment of the present application is shown in the flowchart;

[0036] Figure 6 The maintenance support task atomic model business flowchart provided by the embodiment of the present application is shown in the flowchart;

[0037] Figure 7 The atomic task model set schematic diagram provided by the embodiment of the present application is shown in the schematic diagram;

[0038] Figure 8 The support task sequence schematic diagram provided by the embodiment of the present application is shown in the schematic diagram.

[0039] Figure 9 The equipment fault knowledge graph ontology model structure schematic diagram provided for the embodiment of the present application is shown in the figure.

[0040] Figure 10 The collaborative guarantee marshalling scheme generation flowchart provided for the embodiment of the present application is shown in the figure. DETAILED DESCRIPTION

[0041] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative work fall within the protection scope of the present application.

[0042] The embodiment of the present application provides a technology guarantee product system contribution rate prediction method based on BP neural network, constructs a technology guarantee product system contribution rate prediction model based on BP neural network intelligent algorithm with intelligent processing, compiles a training data sample set according to past expert experience values, research and development unit scientific research production data, use records and other data of using units, and continuously collects system contribution rate data of single technology guarantee product in real-time marshalling based on the technology guarantee product system contribution rate prediction model based on BP neural network intelligent algorithm, continuously trains the intelligent model, improves the accuracy of intelligent output of the model, and achieves the adaptive intelligent prediction ability of different change situations. After preliminary training, the technology guarantee product system contribution rate prediction model based on BP neural network intelligent algorithm is applied to the software system, a guarantee effectiveness system contribution rate prediction system based on multi-life state and oriented to multi-technology guarantee product collaborative guarantee marshalling is designed and developed, and the efficiency, benefit and effectiveness of the technology guarantee of the use department are effectively improved. Figure 1 As shown in the figure, the method comprises the following steps:

[0043] Model training step:

[0044] S1: According to the execution process of different guarantee tasks, the guarantee tasks are divided into a plurality of task atomic models, and a task atomic model set is constructed;

[0045] S2: According to the technical ability information of different technology guarantee products, a technology guarantee product ability model library is constructed, and the technology guarantee product is equipment used for executing guarantee tasks;

[0046] S3: A technology guarantee product system contribution rate prediction model is constructed based on BP neural network;

[0047] S4: According to the technical professional required by the guarantee task, match the technical support product with the corresponding technical ability information; take the guarantee task, the task atomic model corresponding to the guarantee task, the attribute information, the life information and the environment information of the technical support product as input, and the contribution rate of the technical support product to the guarantee system as output to construct a training sample set, and train a technical support product system contribution rate prediction model;

[0048] Model prediction step:

[0049] S5: Obtain the current guarantee task, serialize the current guarantee task according to the execution process of the current guarantee task, obtain the current guarantee task sequence, and call from the task atomic model set according to the task atomic model required by each process step of the current guarantee task sequence;

[0050] S6: According to the technical professional required by the guarantee task, match the technical support product with the corresponding technical ability information from the technical support product ability model library;

[0051] S7: Input the current guarantee task, the task atomic model corresponding to the current guarantee task, the attribute information, the life information and the environment information of the technical support product into the trained technical support product system contribution rate prediction model, and output the contribution rate of the technical support product to the guarantee system.

[0052] The embodiment is based on technical support products, takes the product life state as the breakthrough point, introduces the PB neural network intelligent algorithm, outputs the system contribution rate in the technical support product collaborative grouping, predicts the quantitative guarantee effectiveness, improves the technical support product grouping effectiveness in the joint guarantee action, and provides theoretical data support for product grouping of guarantee team decision makers.

[0053] It should be noted that the model adopts a BP neural network intelligent algorithm as a basic intelligent model. The BP neural network algorithm has a wide range of application fields and is a highly nonlinear algorithm with self-adaptation. It can complete the mapping task in a multi-dimensional space, has a high self-adaptation capability for data through the self-organizing structure of internal links, can learn knowledge from instances, and automatically establishes the control law and cognitive model of a complex system. The BP neural network algorithm has the advantages of strong comprehensive capability, low data requirement, timely learning, simple operation, time saving, and the like. Even if the essence of the algorithm is not understood, the function-rich function can be directly applied to realize the construction of the technical support product system contribution rate prediction model based on the BP neural network intelligent algorithm.

[0054] The BP neural network includes an input layer, a hidden layer and an output layer. The hidden layer can have multiple layers. The number of nodes in the input layer and the output layer is fixed (the number of variables of the input sample and the number of output labels, respectively), but the number of nodes in the hidden layer is not fixed.

[0055] The BP neural network mainly includes two processes, i.e. forward propagation of working signals and backward propagation of error signals, which are responsible for the calculation of the output in the neural network and the update of the weights and thresholds, respectively. The forward propagation of working signals is to obtain the actual output of the BP neural network by calculation, and the backward propagation of error signals is to correct the weights and thresholds from the back to the front layer by layer in order to make the actual output closer to the expected output.

[0056] Forward propagation of working signals: the input signals enter from the input layer, pass through the synapses into the hidden layer neurons, and after the operation of the transfer function, are transmitted to the output layer, and the output signals are calculated in the output layer and transmitted out. When the working signals are forward propagated, the weights and thresholds are fixed and unchanged, and the state of each layer in the neural network is only related to the net output, weights and thresholds of the previous layer. If the expected output is obtained in the output layer by forward propagation, the learning is ended and the current weights and thresholds are retained; if the expected output cannot be obtained in the output layer by forward propagation, the weights and thresholds are corrected in the backward propagation of error signals.

[0057] Backward propagation of error signals: if the expected output cannot be obtained after the forward propagation of working signals, the backward propagation is performed by calculating the error signals, the difference between the actual output and the expected output of the BP neural network is taken as the error signal, and the error signal is propagated from the output layer to the input layer. In this process, the weights and thresholds of each layer are modified as the error signal is propagated forward, and the process is propagated forward until the input layer. This process is to make the result of the neural network closer to the expected result.

[0058] After one forward propagation and backward propagation, if the error still cannot meet the requirements, the process continues until the error meets the accuracy or other set end conditions such as the number of iterations.

[0059] The algorithm is a multi-layer feedforward neural network trained according to the error backpropagation algorithm, and its main feature is that the signal is forward propagated, and the error is backward propagated.

[0060] Before the model prediction operation starts, the agent training of the technology support product system contribution rate prediction model based on the BP neural network intelligent algorithm is needed. Through the analysis of various types of information in actual use, combined with business standards, related support rules, support laws and expert knowledge base as carriers, relevant data in product use environment, support tasks and support execution atomic sequence are learned and analyzed, business rules are sorted out and summarized, and formatted training samples for training the input layer of the technology support product system contribution rate prediction model based on the BP neural network intelligent algorithm are made. The sample content includes training set, validation set and test set.

[0061] The technology support product system contribution rate prediction model based on the BP neural network intelligent algorithm is intelligently generated through machine learning. The model is evaluated and tested through verification sample data and test sample data. The model construction is completed when the project requirements are met. If the project requirements are not met, iterative training is repeated until the project requirements are met.

[0062] The generated training samples will be constructed and used to continuously train the technology support product system contribution rate prediction model based on the BP neural network intelligent algorithm until the model has the ability to independently analyze and output.

[0063] 1. Implementation method

[0064] Read the sample data CSV file and convert the data into the format required for intelligent algorithm training;

[0065] Set the relevant parameters for intelligent algorithm training and input sample data into the intelligent algorithm training method;

[0066] Adjust the sample data according to the training results and retrain the model to enable it to achieve the ability to independently analyze and output.

[0067] 2. Design of training set, validation set and test sample set

[0068] In machine learning, the training set, validation set, and test set are three key components of a dataset, used to train, evaluate, and test the performance of machine learning models. Based on the neuron input requirements of the technology support product system contribution rate prediction model based on the BP neural network intelligent algorithm, the model is designed to input environmental information, product information, and support mission information, and output the product's performance output ratio.

[0069] In practical applications, the proportions of training, validation, and test sets depend on the specific problem and dataset size. Generally speaking, the training set is larger, typically 60%-80% of the total dataset; the validation set is smaller, typically 10%-20%; and the test set is also smaller, typically 10%-20%. Based on this standard, we designed the training, validation, and test sets for this method, as shown below.

[0070] Table 1 Training sample set

[0071]

[0072]

[0073] Table 2 Validation sample set

[0074]

[0075] Table 3 Test sample set

[0076]

[0077] 3. Process logic

[0078] When the user performs the start training operation, the system reads the sample training data and inputs it to the neural network algorithm for training. If the training number is reached, a prediction model is generated. If the training number is not reached, the training continues. Finally, the system reads the sample test data for model testing. If the training target is reached, the training ends. If the training target is not reached, the sample can be corrected, and the training operation is performed again. The process logic is shown in Figure 2 .

[0079] In one embodiment, the input layer data of the technology support product system contribution rate prediction model based on the BP neural network intelligent algorithm is divided into environmental information, technology support product information, and support task information. The output data is the technology support product atomic support task system contribution rate prediction result based on the life profile.

[0080] In this embodiment, as shown in Figure 3 , the typical environment that can affect the execution of the technology support product task is selected as the environmental information model data. The terrain, temperature, altitude, humidity, and weather (wind, rain, fog, snow, etc.) are used to describe the actual environmental conditions. Different life states of the product have different adaptability to the environment.

[0081] In this embodiment, as shown in Figure 4 , as a support team commander, it is necessary to have the ability to master the support performance of the technology support product. The technology support product model is constructed and instantiated from the product basic attributes, support capability information, and product life state information.

[0082] In this embodiment, the large business model or data model is disassembled into a single function atomic task model, and then data analysis and processing are performed. Then, the complex model is generated by structuring multiple simple atomic task models. The support task model that can express complex requirements is generated by structuring the complex model.

[0083] The following takes the rescue support task and maintenance support task as an example to analyze and disassemble the atomic task model.

[0084] The rescue support task atomic task model disassembly example is shown in

[0085] Equipment battlefield rescue is a rescue unit, i.e. an operation implementation subject, which, under the restriction of a battlefield environment, according to an order of a superior, uses rescue equipment to make a trapped, overturned, wounded, and action-incapable equipment in distress safe. The equipment battlefield rescue mode usually has three modes of self-rescue, towing rescue, and evacuation, wherein the towing rescue can be further divided into light trapping towing rescue, medium trapping towing rescue, heavy trapping towing rescue, and super heavy trapping towing rescue, and the evacuation can be further divided into traction evacuation and transportation.

[0086] The equipment battlefield rescue business process of the rescue unit is as shown in Figure 5 In a combat process, a tank, an armored vehicle, a self-propelled gun, and other equipment support objects generate a trapped, overturned, and other dangerous situations. The rescue unit, according to a rescue order requirement, makes preparations for personnel, support equipment, and support equipment, and quickly moves to a rescue site. After arriving at the rescue site, it first judges whether the equipment in distress needs to be hoisted, and if it needs to be hoisted, it unfolds and implements hoisting. Then, according to the specific dangerous situation of the equipment, it judges which rescue mode to use. For a trapped equipment, the towing rescue mode is used to make it safe; for a damaged equipment that cannot be repaired on site, the traction mode is used to transport it to a repair point. When the towing rescue mode is used, according to the actual situation of the equipment in distress, the rescue personnel, tools, equipment, and other resources needed are calculated, and it is judged whether it has the corresponding ability, if it has the corresponding ability, it unfolds and organizes and implements, otherwise it applies for support to the superior, and after the support force arrives, it unfolds and organizes and implements. When the traction mode is used, its activity process is similar to the towing rescue, and is not described here. After the rescue unit completes the rescue task, it should be withdrawn and collected, and the situation should be reported.

[0087] Therefore, according to the rescue business process, the rescue business is abstracted, the part that improves support efficiency through cooperation is decomposed, and the steps that are relatively independent of the business are decomposed, so that the rescue business function can be decomposed into key action atoms of receiving an order, moving, unfolding, hoisting, rescuing, traction, withdrawing and collecting, and reporting. The atomic task model can be instantiated into different atomic tasks according to the specific task situation, for example, the rescue can be instantiated into 50-ton rescue, 40-ton rescue, and the like.

[0088] An example of decomposition of the maintenance support atomic task model is shown in

[0089] Equipment battlefield maintenance is a repair unit, i.e. an operation implementation subject, which, under the restriction of a battlefield environment, according to an order of a superior, uses support equipment to evaluate and quickly repair damaged parts of a battle-damaged or faulty equipment. According to different repair methods, it can be divided into emergency repair, spare part repair, patching repair, original part repair, and comprehensive repair, etc. In wartime, due to the urgency of the situation and the changing situation of the battlefield, the first three repair methods are usually used to shorten the repair time and improve the rapid response capability. The equipment battlefield repair process model is as shown in Figure 6 .

[0090] Thus, according to the repair service process, the repair service is abstracted, and the repair service function can be divided into: receiving instructions, maneuvering, battle damage assessment, deployment, repair, repair and repair, withdrawal, and reporting several key action models, wherein the repair and repair and repair are instantiated into different atomic tasks according to specific business needs.

[0091] The process of data requirement analysis and data entry according to the model structure is the pre-contents of the guarantee task model construction work; the guarantee task model is displayed and viewed in the form of using the operation business process based on the atomic task model constructed in advance, and the system analyzes and evaluates the data structure, data content, and theoretical feasibility of the guarantee task model data, verifies the accuracy of the data, and also verifies the feasibility of the guarantee task model whether it can completely express the product guarantee process. In this embodiment, as shown in Figure 7 The task atomic model set includes: general task atomic model, rescue guarantee task atomic model, maintenance guarantee task atomic model, transportation guarantee task atomic model, and energy guarantee task atomic model.

[0092] According to the atomic task model construction results, the guarantee task model is prepared according to the guarantee business operation step law, upstream and downstream business processes, flow content, processing efficiency, personnel characteristics, etc., and the divided atomic task model is specified to the technical guarantee product according to the business function, and the upstream and downstream relationships of the business information flow are determined according to the business steps, and the business model design and implementation are completed. The maintenance guarantee task model example is shown in Figure 8 .

[0093] For example: a certain guarantee group receives a repair task, and three equipment in a nearby place are repaired, and the traditional guarantee mode is three repair guarantee businesses, and the maintenance guarantee scheme developed according to the battle damage assessment is on-site repair, equipment application, and repair and repair, and the corresponding business steps are expressed in the form of task atoms as follows:

[0094] Business 1: receiving instructions -> maneuvering to the equipment warehouse -> loading rescue equipment and materials -> maneuvering to the equipment to be rescued -> deployment -> rescue and repair -> withdrawal -> reorganization;

[0095] Business 2: receiving instructions -> maneuvering to the equipment warehouse -> loading rescue equipment and materials -> maneuvering to the equipment to be rescued -> deployment -> repair -> withdrawal -> reorganization;

[0096] Business 2: receiving instructions -> maneuvering to the equipment warehouse -> loading rescue equipment and materials -> maneuvering to the equipment to be rescued -> deployment -> hoisting -> repair and repair -> withdrawal -> reorganization;

[0097] The corresponding relationship between the atomic task and the guarantee specialty is shown in the following table.

[0098] Table 4: Correspondence table between atomic tasks of support service 1 and support specialties

[0099]

[0100] Table 5: Correspondence table between atomic tasks of support service 2 and support specialties

[0101]

[0102] Table 6: Correspondence table between atomic tasks of support service 3 and support specialties

[0103]

[0104] All current technical support products are sorted out, a technical support product capability model library is constructed, and individual technical support product capabilities are instantiated, and a related capability dictionary library is constructed.

[0105] Table 7: Instance of technical support product capability dictionary table

[0106]

[0107]

[0108] According to the correlation relationship model established between support tasks <-> atomic tasks <-> support specialties <-> support products, we can systematically and sequentially process the support task demand list to calculate multiple support product combinations, thereby generating a comprehensive support formation required to complete the support task. Different technical support products can complete mutually intersecting support atomic tasks, so for the same support task, multiple different technical support product combinations can be output. The corresponding mode is shown in Figure 9

[0109] In one embodiment, the attribute information of the technical support product includes one or more of the following: product name, product number, product type, product function information.

[0110] In one embodiment, the environmental information of the technical support product includes environmental data affecting the execution of the technical support product task, including one or more of the following: terrain, temperature, altitude, humidity, and climate.

[0111] In one embodiment, the technical support product capability model library is used to store attribute information of the technical support product, and a capability dictionary table of the technical support product associated with the corresponding technical support product. The capability information in the capability dictionary table is the same as the content of the technical professional information required for matching support task execution, so as to match the technical support product with the corresponding technical capability information according to the technical professional required for support task execution.

[0112] ​​In one embodiment, the execution flow step of the current guarantee task sequence corresponds to several task atomic models, the technical specialty required by each task atomic model is several, and matches several technical guarantee products with corresponding technical ability information; for the current guarantee task, obtain a plurality of different technical guarantee product combinations.

[0113] In one embodiment, according to the characteristics of the diversity of guarantee methods and product guarantee capabilities, there are usually multiple guarantee schemes to complete a certain guarantee demand, and it is the main purpose of the method for the guarantee detachment commander to select a guarantee scheme with good timeliness, strong environmental adaptability, and low cost-effectiveness, and the key problem of determining the guarantee scheme is to determine the collaborative guarantee formation and guarantee business process. The multi-product collaborative guarantee formation guarantee efficiency system contribution rate prediction system focuses on solving the problem of how to complete the optimization of high-efficiency, high-coordination, and low-cost technical guarantee product collaborative formation through predicting the guarantee system contribution rate of a single technical guarantee product in the case of multiple life states, complex environments, and multiple guarantee business processes coexisting. It should be noted here that a high system contribution rate does not necessarily mean good guarantee effect, and conditions such as high overall timeliness, low total number of product deployments, and sufficient product residual strength should be prioritized. Therefore, the following steps are also included:

[0114] The current guarantee task sequence corresponds to a plurality of different technical guarantee product combinations, and the trained technical guarantee product system contribution rate prediction model outputs a list of guarantee system contribution rates of different technical guarantee product combinations;

[0115] The time consumption of each task atomic model combination is calculated, and the guarantee duration of the guarantee task sequence under different task atomic model combinations is obtained through the critical path algorithm. Through comparative analysis, the optimal formation scheme is output.

[0116] In specific execution, the execution flow is:

[0117] First, obtain the guarantee demand, obtain all feasible guarantee task sequences for this guarantee demand through the guarantee task sequence library, match technical guarantee product entities (including life information) and guarantee atomic tasks in the feasible guarantee task sequences in combination with the current available technical guarantee product entity information in the guarantee detachment, input the current environmental information as parameters into the technical guarantee product system contribution rate prediction model based on the BP neural network intelligent algorithm, perform single-product single-task atomic system contribution rate prediction, until all combination information is predicted, output the guarantee atomic task system contribution rate list, and finally combine the guarantee task sequence library to calculate the comprehensive guarantee effect of all feasible formation schemes, recommend a multi-product collaborative guarantee formation scheme under multiple life states in combination with the guarantee duration of the guarantee task sequence, and the formation scheme includes a guarantee entity list, a guarantee business sequence, and system contribution rate information of each atomic task. The specific process is as followsFigure 10 are shown.

[0118] The above describes in detail the technical support product system contribution rate prediction method based on the BP neural network provided by the present application. The principle and implementation manner of the present application are described by applying specific examples in the embodiment. The above embodiment description is only used to help understand the method of the present application and its core idea. Meanwhile, for the general technical personnel in the field, the specific implementation manner and application range will be changed according to the idea of the present application. In conclusion, the content of the specification should not be understood as a limitation of the present application.

[0119] The above description of the disclosed embodiments enables those skilled in the art to carry out or use the present application. Various modifications to the embodiments will be apparent to those skilled in the art, and the general principles defined in the present embodiment can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown in the present embodiment, but will conform to the widest scope consistent with the principles and novel features disclosed in the present embodiment.

Claims

1. A method for predicting the contribution rate of a technology support product system based on a BP neural network, characterized in that: The steps include: Model training steps: S1: According to the execution process of different guarantee tasks, the guarantee tasks are divided into several task atomic models, and a set of task atomic models is constructed; S2: Constructing a technical support product capability model library based on the technical capability information of different technical support products, wherein the technical support products are equipment used to perform the support mission; S3: Construct a technology support product system contribution rate prediction model based on BP neural network; S4: Match technical support products with corresponding technical capability information based on the technical expertise required to execute the support task; use the support task, the task atomic model corresponding to the support task, the attribute information, lifespan information, and environmental information of the technical support product as input, and the contribution rate of the technical support product to the support system as output to construct a training sample set, and train the technical support product system contribution rate prediction model; Model prediction steps: S5: Obtain the current assurance task, serialize the current assurance task according to the execution process of the current assurance task to obtain the current assurance task sequence, and retrieve the task atomic model required by each process step of the current assurance task sequence from the task atomic model set; S6: Matching technical support products with corresponding technical capability information from the technical support product capability model library according to the technical expertise required for the execution of the support task; S7: Input the current support task, the task atomic model corresponding to the current support task, the attribute information, life information and environmental information of the technical support product into the trained technical support product system contribution rate prediction model, and output the contribution rate of the technical support product to the support system.

2. The method for predicting the contribution rate of a technology support product system based on a BP neural network according to claim 1, characterized in that: The attribute information of the technical support product includes one or more of the following: product name, product number, product type, and product function information.

3. The method for predicting the contribution rate of a technology support product system based on a BP neural network according to claim 1, characterized in that: The environmental information of the technical support product includes: environmental data that affects the execution of the technical support product mission, including one or more of the following: terrain, temperature, altitude, humidity and climate.

4. The method for predicting the contribution rate of a technology support product system based on a BP neural network according to claim 1, characterized in that: The technical support product capability model library is used to store the attribute information of technical support products, as well as the capability dictionary table of technical support products associated with the corresponding technical support products. The capability information in the capability dictionary table is the same as the content of the technical professional information required for the execution of the matching support tasks, so as to match technical support products with corresponding technical capability information according to the technical expertise required for the execution of the support tasks.

5. The method for predicting the contribution rate of a technology support product system based on a BP neural network according to claim 1, characterized in that: The execution process steps of the current support task sequence correspond to several of the task atomic models. Each of the task atomic models requires several technical expertise and is matched with several technical support products with corresponding technical capability information. For the current support task, a variety of different technical support product combinations are obtained.

6. The method for predicting the contribution rate of a technology support product system based on a BP neural network according to claim 5, characterized in that: The following steps are also included: The current support task sequence corresponds to a variety of different technical support product combinations, and the trained technical support product system contribution rate prediction model outputs a list of contribution rates of different technical support product combinations to the support system; The time consumption of each task atomic model combination is calculated one by one, and the guarantee duration of the guarantee task sequence under different task atomic model combinations is obtained through the critical path algorithm. Through comparative analysis, the optimal grouping plan is output.

7. The method for predicting the contribution rate of a technology support product system based on a BP neural network according to claim 1, characterized in that: The mission atomic model set includes: a general mission atomic model, a rescue support mission atomic model, a maintenance support mission atomic model, a transportation support mission atomic model and an energy support mission atomic model.

Images