CN110930054A - Data-driven battle system key parameter rapid optimization method - Google Patents

Data-driven battle system key parameter rapid optimization method Download PDF

Info

Publication number
CN110930054A
CN110930054A CN201911222070.0A CN201911222070A CN110930054A CN 110930054 A CN110930054 A CN 110930054A CN 201911222070 A CN201911222070 A CN 201911222070A CN 110930054 A CN110930054 A CN 110930054A
Authority
CN
China
Prior art keywords
parameters
neural network
optimization
combat
key
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201911222070.0A
Other languages
Chinese (zh)
Inventor
宫琳
朱明仁
陈西
吴开放
陈伟
周金鹏
张明恩
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing Institute of Technology BIT
Original Assignee
Beijing Institute of Technology BIT
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Beijing Institute of Technology BIT filed Critical Beijing Institute of Technology BIT
Priority to CN201911222070.0A priority Critical patent/CN110930054A/en
Publication of CN110930054A publication Critical patent/CN110930054A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
    • G06Q10/063—Operations research, analysis or management
    • G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/08—Learning methods
    • G06N3/084—Backpropagation, e.g. using gradient descent

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Human Resources & Organizations (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Educational Administration (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Economics (AREA)
  • Development Economics (AREA)
  • General Physics & Mathematics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Computational Linguistics (AREA)
  • Quality & Reliability (AREA)
  • General Business, Economics & Management (AREA)
  • Operations Research (AREA)
  • Marketing (AREA)
  • Health & Medical Sciences (AREA)
  • Game Theory and Decision Science (AREA)
  • Artificial Intelligence (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Tourism & Hospitality (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • General Engineering & Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a data-driven critical parameter fast optimization method for a combat system, which comprises the steps of firstly, extracting combat parameters as critical parameters and extracting combat assessment parameters for the combat system with the parameters to be optimized; constructing a training sample containing key parameters and combat assessment parameters, and training a neural network; the intelligent optimization algorithm takes the trained neural network as a target function, optimizes key parameter values and outputs an optimization result; and the intelligent optimization algorithm executes multiple times of optimization, performs simulation evaluation on all optimizing results, and selects a key parameter value corresponding to the optimal simulation evaluation result as a final optimization result. The invention can improve the optimization efficiency.

Description

Data-driven battle system key parameter rapid optimization method
Technical Field
The invention relates to the technical field of parameter optimization, in particular to a data-driven method for quickly optimizing key parameters of a combat system.
Background
The design of the combat system refers to a process of organically managing various weapon equipment systems which are functionally interconnected and interacted with each other by taking the realization of a combat task as a target under the condition of meeting strategic guidance, combat command and guarantee, and forming a higher-level system through a system engineering method. And for a specific combat system design process, generating a system design scheme based on combat mission requirements, and evaluating the efficiency of the system design scheme based on system key parameters to measure the system design quality and performance.
The system performance evaluation method mainly comprises an analytic method and a simulation evaluation method. The analytical method mainly aims at an architecture model with simple index relationship, low coupling and understandable mathematical relationship, but is not applicable to a complex model with high index coupling and implicit relationship; the simulation evaluation method has the advantages of capability of simulating a wartime environment, consideration of multi-system cooperative combat, high model controllability, convenience in directly aiming at a certain factor and adjustment of relevant parameters, and the like, and is widely used in design evaluation of a complex combat system. However, the simulation method has the problem of long time consumption in the simulation process, and is not suitable for quickly evaluating a combat system. From the direction of evaluating the system efficiency, the method is mainly divided into three aspects: the method comprises the steps that firstly, in the aspect of the operational capacity, the evaluation process is oriented to each equipment, and the influence of the equipment on the operational efficiency of a system is directly evaluated; secondly, analyzing the system combat effectiveness from the perspective of the overall effectiveness and the system contribution rate; and thirdly, analyzing the influence of a single device on the integrity of the whole system from the overall view. Currently, a performance evaluation method is in a relatively mature and complete stage, but a scheme for optimizing and adjusting the existing system based on an evaluation result so as to improve the overall performance of the system is lacking. Therefore, even if the evaluation of the system is completed, how to improve and optimize the problems and short boards found based on the evaluation result is not enough, the current academic world also lacks a method for optimally designing key index parameters oriented to the maximization of the system efficiency based on the evaluation result and lacks a method theory for optimally designing the key parameters of the system.
Disclosure of Invention
In view of the above, the invention provides a data-driven method for quickly optimizing key parameters of a combat system, so as to improve optimization efficiency.
In order to solve the technical problem, the invention is realized as follows:
a data-driven method for quickly optimizing key parameters of a combat system comprises the following steps:
extracting operation parameters as key parameters and extracting fighting evaluation parameters aiming at an operation system of parameters to be optimized;
constructing a training sample containing key parameters and combat assessment parameters, and training a neural network;
step three, the intelligent optimization algorithm takes the trained neural network as a target function, optimizes key parameter values and outputs an optimization result;
and fourthly, executing multiple times of optimization by the intelligent optimization algorithm, performing simulation evaluation on all the optimization results, and selecting the key parameter value corresponding to the optimal simulation evaluation result as the final optimization result.
Optionally, in the step, at least 2 kinds of neural networks are trained, and the neural network with the best precision is selected as the target function.
Optionally, the selecting the neural network with the best precision as the target function is: and calculating the root mean square error between the output of the neural network and the sample, and selecting the neural network with the minimum root mean square error as an objective function.
Optionally, the neural network selects a single hidden layer BP neural network and a single hidden layer RBF neural network.
Optionally, selecting at least 2 intelligent optimization algorithms to optimize the key parameter values; and each intelligent optimization algorithm in the fourth step executes multiple times of optimization.
Optionally, in step three, the determination method of the end of the iteration is: and finishing the iteration after the obtained satisfactory battle evaluation parameter value or the iteration number reaches a set value.
Has the advantages that:
the invention trains the neural network by using the prior combat simulation data, takes the trained neural network as a target function, realizes the optimization design of key parameter values by an intelligent optimization algorithm, and can reduce the optimization time because a large amount of simulation calculation is not needed.
Drawings
FIG. 1 is a flow chart of a method for quickly optimizing key parameters of a battle system based on data driving according to the present invention.
Detailed Description
The invention is described in detail below by way of example with reference to the accompanying drawings.
The invention provides a data-driven method for quickly optimizing key parameters of a combat system, which comprises the following steps as shown in figure 1:
analyzing a combat system of parameters to be optimized, extracting combat parameters as key parameters of the combat system, and extracting combat evaluation parameters.
In this step, the combat system to be evaluated is analyzed, combat parameters having a large influence on the combat result are extracted as key parameters of the combat system, and combat evaluation parameters concerned by a combat party are extracted as the combat evaluation parameters of the combat system. The operational parameters such as operational preparation time, quantity of combat supplies, etc. are generally selected as key parameters. The combat assessment parameters such as combat effectiveness, missile interception rate, casualties of our party, etc. may be one or more combat assessment parameters, in the following example combat effectiveness is selected as the combat assessment parameter.
And step two, constructing a training sample containing the key parameters and the combat assessment parameters, and training the neural network. The second step comprises the following substeps:
substep 21: and extracting the past combat simulation data. In this embodiment, the combat assessment parameter obtained by the simulation assessment method is selected as a sample label y, and the combat parameter of the simulation assessment method is input as a sample x. Where x is the same as the key parameter selected in step one and y is the same as the battle assessment parameter selected in step one. x and y may each be a parameter or a vector of parameters. (x, y) constitutes the prototype of the training sample. In practice, the battle evaluation parameters obtained by other schemes can be selected, and are not limited to simulation evaluation methods.
Substep 22: the combat parameters selected in the substep 21 are preprocessed.
Preprocessing here refers to normalization. Aiming at different influence conditions of different operational parameters on efficiency (battle evaluation parameters), the operational parameters are divided into benefit type and cost type:
a) benefit type-increase of combat parameters has a promoting effect on the effect value;
b) cost type-increase of combat parameters has a restraining effect on the efficacy value.
The samples were normalized using a linear scale transform method:
Figure RE-GDA0002366447570000041
in the above formula, rijFor the battle parameters after normalization, xijFor the battle parameters before normalization, XiIs a data set of the ith type of combat parameter. i represents the ith operational parameter, and j represents the jth data in one operational parameter.
Substep 23: and training the neural network.
In this embodiment, at least two different types of neural networks are selected for training to fit the functional relationship between the operational parameters and the combat assessment parameters. And after the training is finished, calculating the fitting precision of the neural network, and selecting the neural network with the best precision as the target function of the step three.
When selecting the neural network, a single hidden layer BP neural network and a single hidden layer RBF neural network may be adopted, and the number of hidden layer units may be determined by using the following empirical formula:
Figure RE-GDA0002366447570000051
wherein hiddens is the number of hidden layer units, inputs is the number of input units, and outputs is the number of output units.
The training samples were divided into training sets and test sets by 10 to 1. Training a neural network by using training set data; and then inputting each trained neural network by adopting a test set to obtain an output value, calculating the root mean square error between the output value and the sample label y, and selecting the neural network with the minimum root mean square error as the neural network with the best precision.
And step three, the intelligent optimization algorithm takes the trained neural network as an objective function, optimizes the key parameter values and outputs an optimization result.
Preferably, at least 2 intelligent optimization algorithms are selected to optimize the key parameter values. For example, Particle Swarm Optimization (PSO) and Genetic Algorithm (GA) can be selected. According to the method, an intelligent optimization algorithm is not required to be improved, only the trained neural network selected in the second step is used as an objective function, and after each step of iteration, the combat parameters obtained in the iteration are substituted into the neural network to obtain combat assessment parameter values. Selecting a better combat parameter by using the combat assessment parameter value by using the operation of the intelligent optimization algorithm, and entering the next iteration; and after the obtained satisfactory battle evaluation parameter value or the iteration number reaches a set value, ending the iteration and outputting an optimization result.
And fourthly, executing multiple times of optimization by the intelligent optimization algorithm, performing simulation evaluation on all the optimization results, and selecting the optimal key parameter value according to the simulation evaluation results.
In the step three, each intelligent optimization algorithm used in the step three is executed for multiple times to obtain multiple optimization results. And inputting the optimizing result into the simulation model, obtaining a combat assessment parameter by using a traditional simulation assessment method, and selecting a combat parameter corresponding to the optimal combat assessment parameter to obtain a final optimizing result.
This flow ends by this point.
An example is given below. The example is based on battlefield confrontation of a blue party and a red party, wherein the blue party is an attacking party, the red party is a defending party, and a blue party corresponding battle system design case is generated. The specific evaluation contents are as follows:
the blue party develops an attack on some military base of the red party by utilizing an air-to-ground attack formation. And designing an air defense system by the Hongfang based on the existing weaponry system. The example selects key parameters including radar discovery probability I12Preparation time for battle I21Single missile launch interval I22Intercept slope I31Number of single target missile launches I32Number of channels for use of fire I34Missile launching interval I42And a target RCS value I43And selecting the combat effectiveness as a combat assessment parameter.
(1) Preprocessing key index data: and acquiring n groups of key parameters and efficiency data thereof, and performing normalization processing on the key parameters by adopting the benefit-type and cost-type classification processing mode.
TABLE 1
Figure RE-GDA0002366447570000061
(2) Neural network training
And (3) dividing the normalized key ability index data samples in the step (1) into a training set and a testing set according to the ratio of 10 to 1. Training a BP neural network by using training set data, wherein model parameters of the BP neural network are shown in a table 2:
TABLE 2
Figure RE-GDA0002366447570000062
Based on the trained neural network, the effect of the proxy model is tested by using the reserved test set, and the statistical result is shown in table 3:
TABLE 3
Figure RE-GDA0002366447570000071
(3) Optimization model solution
And (3) based on the BP neural network obtained by training in the step (2), optimizing and solving the key index parameters by utilizing a Particle Swarm Optimization (PSO). The PSO algorithm parameters were set as shown in table 4:
TABLE 4
Figure RE-GDA0002366447570000072
The key index parameters are optimized and solved for four times to obtain four groups of optimal design parameters, and the results of the four times are verified by respectively utilizing a simulation evaluation method and a neural network evaluation method, as shown in table 5:
TABLE 5
Figure RE-GDA0002366447570000073
(4) Analysis of results
And analyzing the PSO optimization solution result, wherein the third optimization parameter value has the best effect and is higher than the optimal design in the original sample data, and the technical effectiveness of the method is verified. On the other hand, the four times of optimization are carried out for 200 times of iterative optimization calculation, 10000 groups of key parameter design values are calculated by using the agent model, the total time consumption is 0.53h, the time consumption of a simulation evaluation system under the same operation condition is about 36.36h, and the operation efficiency of the method is far higher than that of a key parameter optimization method based on a simulation system.
Therefore, the method lays a foundation for efficiency evaluation result prediction and rapid optimization design of key parameter indexes of the battle system design scheme, and provides an idea direction for the optimization design of the key parameters of the battle system.
In summary, the above description is only a preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (6)

1. A data-driven method for quickly optimizing key parameters of a combat system is characterized by comprising the following steps:
extracting operation parameters as key parameters and extracting fighting evaluation parameters aiming at an operation system of parameters to be optimized;
constructing a training sample containing key parameters and combat assessment parameters, and training a neural network;
step three, the intelligent optimization algorithm takes the trained neural network as a target function, optimizes key parameter values and outputs an optimization result;
and fourthly, executing multiple times of optimization by the intelligent optimization algorithm, performing simulation evaluation on all the optimization results, and selecting the key parameter value corresponding to the optimal simulation evaluation result as the final optimization result.
2. The method of claim 1, wherein said step trains two or more neural networks, and selects the neural network with the best precision as the objective function.
3. The method of claim 2, wherein the selecting the neural network with the best accuracy as the objective function is: and calculating the root mean square error between the output of the neural network and the sample, and selecting the neural network with the minimum root mean square error as an objective function.
4. The method of claim 2, wherein the neural network selects a single hidden layer BP neural network and a single hidden layer RBF neural network.
5. The method of claim 1, wherein step three selects at least 2 intelligent optimization algorithms to optimize key parameter values; and each intelligent optimization algorithm in the fourth step executes multiple times of optimization.
6. The method of claim 1, wherein in step three, the end of the iteration is determined by: and finishing the iteration after the obtained satisfactory battle evaluation parameter value or the iteration number reaches a set value.
CN201911222070.0A 2019-12-03 2019-12-03 Data-driven battle system key parameter rapid optimization method Pending CN110930054A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911222070.0A CN110930054A (en) 2019-12-03 2019-12-03 Data-driven battle system key parameter rapid optimization method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911222070.0A CN110930054A (en) 2019-12-03 2019-12-03 Data-driven battle system key parameter rapid optimization method

Publications (1)

Publication Number Publication Date
CN110930054A true CN110930054A (en) 2020-03-27

Family

ID=69848522

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911222070.0A Pending CN110930054A (en) 2019-12-03 2019-12-03 Data-driven battle system key parameter rapid optimization method

Country Status (1)

Country Link
CN (1) CN110930054A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112926739A (en) * 2021-03-11 2021-06-08 北京计算机技术及应用研究所 Network countermeasure effectiveness evaluation method based on neural network model

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6658396B1 (en) * 1999-11-29 2003-12-02 Tang Sharon S Neural network drug dosage estimation
CN107886161A (en) * 2017-11-03 2018-04-06 南京航空航天大学 A kind of global sensitivity analysis method for improving Complex Information System efficiency
CN109614729A (en) * 2018-12-18 2019-04-12 上海机电工程研究所 A kind of equipment Efficacy assessment result rapid analysis method based on meta-model
CN110162807A (en) * 2018-02-12 2019-08-23 沈阳理工大学 Based on improvement ant group optimization BP neural network Ammunition Storage Reliability Evaluation method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US6658396B1 (en) * 1999-11-29 2003-12-02 Tang Sharon S Neural network drug dosage estimation
CN107886161A (en) * 2017-11-03 2018-04-06 南京航空航天大学 A kind of global sensitivity analysis method for improving Complex Information System efficiency
CN110162807A (en) * 2018-02-12 2019-08-23 沈阳理工大学 Based on improvement ant group optimization BP neural network Ammunition Storage Reliability Evaluation method
CN109614729A (en) * 2018-12-18 2019-04-12 上海机电工程研究所 A kind of equipment Efficacy assessment result rapid analysis method based on meta-model

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
孟晶 等: ""基于BP神经网络和遗传算法的扫雷参数优化"", 《水雷战与舰船防护》 *
钱晓超 等: ""基于效能评估的武器装备体系优化设计方法"", 《***仿真技术》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112926739A (en) * 2021-03-11 2021-06-08 北京计算机技术及应用研究所 Network countermeasure effectiveness evaluation method based on neural network model
CN112926739B (en) * 2021-03-11 2024-03-19 北京计算机技术及应用研究所 Network countermeasure effectiveness evaluation method based on neural network model

Similar Documents

Publication Publication Date Title
Przemieniecki Mathematical methods in defense analyses
CN108647414A (en) Operation plan adaptability analysis method based on emulation experiment and storage medium
CN108615122A (en) A kind of air-defense anti-missile system combat capability assessment method
CN107766668B (en) Complex simulation model verification method based on neural network
CN109614729A (en) A kind of equipment Efficacy assessment result rapid analysis method based on meta-model
CN112861257B (en) Aircraft fire control system precision sensitivity analysis method based on neural network
CN113627749B (en) Damage efficiency evaluation and fire planning method based on convolutional neural network
CN111080108A (en) Data-driven weapon equipment combat effectiveness evaluation index screening method and system
CN114358141A (en) Multi-agent reinforcement learning method oriented to multi-combat-unit cooperative decision
CN112417710B (en) Weapon equipment system contribution degree evaluation method based on combat ring
CN115222271A (en) Weapon target distribution method based on neural network
CN105739304B (en) A kind of more UCAV based on antithetical ideas improved adaptive GA-IAGA strike target distribution method online
CN109299491B (en) Meta-model modeling method based on dynamic influence graph strategy and using method
CN113177583B (en) Aerial target clustering method
Lappi Computational methods for tactical simulations
CN110930054A (en) Data-driven battle system key parameter rapid optimization method
Li et al. Multi-uav air combat weapon-target assignment based on genetic algorithm and deep learning
CN109325301A (en) Weapon Equipment System efficiency fast Optimization
CN116050515B (en) XGBoost-based parallel deduction multi-branch situation prediction method
Wu et al. Dynamic multitarget assignment based on deep reinforcement learning
CN115186378A (en) Real-time solution method for tactical control distance in air combat simulation environment
Zhang Research on Weighted Model of Air Defense Weapon-Target Assignment Based on AHP and Entropy Method
Lu et al. Research on User-oriented Product Development Mode of Systematic Weapon and Equipment
Li et al. Dynamic weapon target assignment based on deep q network
Li et al. Multi-aircraft cooperative target allocation in BVR air combat using cultural-genetic algorithm

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
WD01 Invention patent application deemed withdrawn after publication
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20200327