CN114841068A - Three-dimensional high-simulation war game deduction platform and method - Google Patents

Abstract

The invention discloses a three-dimensional high-simulation war game deduction platform and a method. The platform includes: the system comprises a planning module, a database module, a deduction module, a transmission module, a display module, an algorithm module and a group evolution training module; the method comprises the following steps: initializing parameters of a deduction platform; an operator automatically constructs a two-party algorithm for battle or uses an algorithm in a population learning algorithm library in the platform as the two-party algorithm for battle and operates the two-party algorithm, and situation information output by the platform is input into the algorithm to obtain a decision instruction; the situation information is displayed visually in real time; when the task ending condition is reached, ending the battle, repeating the steps and training according to the algorithms of the two parties of the battle; and when the algorithm is converged, the fighting visual information is observed to obtain the fighting thought of the intelligent agent algorithm for the reference of actual command. The invention provides a three-dimensional, high-precision, more detailed and high-battlefield reducibility military chess deduction platform and a method which are beneficial to quick editing.

Description

Three-dimensional high-simulation war game deduction platform and method

Technical Field

The invention relates to the field of a war game deduction system, in particular to a three-dimensional high-simulation war game deduction platform and a method.

Background

At present, a war and chess deduction system is an important national defense safety support tool, and the war and chess deduction system mainly has the functions of simulating multi-party participating sea, land and air and special combined combat action patterns and judging war situation changes caused by command decisions of deductions. The computer war game deduction system is mainly divided into military and commercial.

However, the simulation environments of the two deduction systems are basically consistent, the local battlefield environment is simulated based on the two-dimensional plane geographic images, the real-time display of the three-dimensional situation cannot be realized, the utilization of height information is insufficient, operation details such as the hitting positions of operation units and the like are ignored, the functions are single, the space operation simulation is extremely unfavorable, and the command issuing of operation commands is further influenced. In addition, the two parties of the battle mainly interact with each other by human, the battle model development mainly takes a two-dimensional battle space and a digital battle unit as the basis, the battle space lacks authenticity, the battle unit lacks related complex modeling, the integral visualization capability is poor, and the gap between the integral visualization capability and the actual battle process is large; the existing simulation platform has limited combat scenarios and is not easy to modify, and the process of setting up a newly-planned environment is tedious, so that the rapid simulation of a modern combined combat system on emergencies is not facilitated.

Therefore, the technical personnel in the field need to solve the problem of how to provide a three-dimensional, high-precision, more detailed and battlefield-restoring chess-deducing platform and method which are beneficial to fast editing.

Disclosure of Invention

In view of the above, the present invention provides a three-dimensional highly simulated war game deduction platform based on an illusion engine,

in order to achieve the purpose, the invention adopts the following technical scheme:

the method comprises the following steps: the system comprises a planning module, a database module, a deduction module, a transmission module, a display module, an algorithm module and a group evolution training module;

the imagination module edits the operation unit, weapon performance, operation force, operation environment and task based on the operation assumption to obtain initialization parameters. The imagination module is used for imagination of basic situation, operation attempt and operation development situation of both parties of the battle.

The database module comprises: the system comprises a three-dimensional example model, an environment library, a scenario database, a deduction record database, a game rule database and a population learning algorithm library;

the deduction module is driven by a fantasy four-physical engine and is used for triggering action response, collision response, weather response, communication response and adjudication response and counting countermeasure indexes according to the initialization parameters to obtain situation information;

namely, the illusion engine has the advantages of supporting high-precision three-dimensional model and physical rule operation and supporting construction of more complex overall environment, so that a complex combined combat system is built;

the display module is used for visually displaying the situation information;

the algorithm module is used for fighting training and comprises the steps of obtaining decision instructions through the situation information and inputting the decision instructions to the transmission module to enable the transmission module to make corresponding actions, so that excellent strategies after each wheel pair fight are obtained;

the transmission module transmits the situation information in the deduction module to the algorithm module, receives a decision instruction from the algorithm module, and transmits the situation information to the display module for display and output;

the population evolution training module uses a population learning algorithm to form a population training strategy by each pair of excellent fighting strategies, and the population training strategy is used as an embedded strategy for training of the algorithm module.

Preferably, the editing content of the initialized parameter task in the scenario module is to set the attack task to destroy all the forces of the defending party or destroy the defending target of the defending party.

Preferably, the information transmitted by the transmission module includes: the information of the combat unit, the visible light visual information, the infrared sensing information and the three-dimensional environment information.

Preferably, the statistics of the countermeasure index in the deduction module include: the fighting damage ratio, the resource consumption, the accidental injury ratio and the task completion progress in the battle, wherein the accidental injury ratio refers to the proportion of the randomly occurring neutral party damage amount in the battlefield to the total damage amount.

Preferably, the display module includes: main visual angle display, follow visual angle display, free visual angle display, small map display and countermeasure index display; the main visual angle display is a fixed global visual angle display; the main visual angle display and the free visual angle display are main windows; the following visual angle display, the small map display and the confrontation index display are small windows and are attached to the main window.

Preferably, the scenario module, the deduction module and the display module are realized through a client; the client comprises a Windows client and a linux client, and the actions of the combat units in the platform are controlled by a human player or an AI algorithm through the client.

Preferably, the system comprises a server which can be accessed by the client, and the server is placed in a Linux system mirror image and is packaged into a docker mirror image file; the server obtains data from a database module.

The invention also provides a three-dimensional high-simulation chess deduction method based on the illusion engine, which comprises the following steps:

s1: initializing parameters of a combat unit, weapon performance, combat force, combat environment and mission, and adopting default parameters if the parameters are not manually set;

s2: respectively constructing and operating an intelligent agent algorithm or a manual rule setting algorithm of both parties of the battle, inputting situation information output by the deduction platform into the algorithm to obtain a decision instruction, and inputting the decision instruction into the platform, wherein the decision instruction changes the situation information;

s3: the situation information is displayed visually in real time;

s4: ending the battle when the task ending condition is reached, and repeating S2;

s5: and outputting the strategy of the intelligent agent algorithm when the algorithms of the two parties of the battle converge.

Preferably, in step S2, the algorithm for both parties to the battle is constructed by using one host alone or by using two hosts separately.

Preferably, the method further comprises the following steps: s6: and calling a population learning algorithm of the platform according to the converged fighting algorithm to train so as to obtain a population training strategy.

Compared with the prior art, the invention discloses a three-dimensional, high-precision, more detailed and battlefield-restoring chess-deducing platform and method beneficial to quick editing, which have the advantages that the editing performance of the operation environment and the operation unit model improves the planning and re-editing efficiency of the platform and supports multi-level and large-scale modern combined operation deduction. The fighting thought of the algorithm can be obtained by observing a certain office for reference of actual command. The platform obviously improves the visual experience of human observers in the battle process, effectively reduces damage judgment errors in the battle process through collision detection, and improves the accuracy of the confrontation result.

Drawings

In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.

Fig. 1 is a schematic structural diagram provided by the present invention.

Detailed Description

The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

The embodiment of the invention discloses a three-dimensional high-simulation war game deduction platform, which comprises: the system comprises a planning module, a database module, a deduction module, a transmission module, a display module, an algorithm module and a group evolution training module;

in one embodiment, the scenario module compiles the operational units, weapon performance, operational capacity, operational environment and mission based on the operational scenario to obtain initialization parameters. The thought module is used for thinking about basic situations, fighting attempts and fighting development situations of two parties in the fighting. In the planning module, the combat unit can be edited, for example, airplanes of different types, radars of different types and missiles of different ranges can be selected from the database; the weapon performance can be edited, for example, the unit speed of operation, the number of guided missiles carried, the missile range and the like can be modified; editing the fighting capacity, for example, setting different quantities of various different types of fighting units of both parties; editing the landform, such as adding lakes, islands, sand dunes and other landforms in the existing landform; editing the weather, such as setting the battlefield environment to be sunny, rainy, sandstorm, snowy and the like.

The database module includes: the system comprises a three-dimensional example model, an environment library, a scenario database, a deduction record database, a game rule database and a population learning algorithm library; the data in the database are all obtained by collecting the public data on the network. The three-dimensional example model comprises models of various terrain resources such as oceans, sand dunes and islands, models of various fighting units such as bombs 6K, SU-33, guard ships, missile models, models of various buildings such as command posts, radars and the like; the environment library is obtained from a fantasy mall and comprises scenes such as sunny days, rainy days, snowy days and the like; the scenario database stores the scenarios constructed each time, and if the scenarios are not needed to be used again, the scenarios can be deleted from the scenario database; the deduction record database stores the result and process of each deduction, or the deduction with pause in the middle, and the deduction record database can be loaded for deduction again; different battlefield point rules, win or lose judgment rules and the like are stored in the game rule database, and the rules can be defined by self and added into the game rule database; the population learning algorithm library is embedded with a population evolution algorithm in the platform, can be trained through the population evolution algorithm under a certain rule to obtain a decision strategy, and stores a population obtained after training into the population learning algorithm library for being used as the embedded strategy for the algorithm module to train.

The deduction module is driven by a phantom four-physical engine and is used for triggering action response, collision response, weather response, communication response and judgment response and counting countermeasure indexes according to the initialization parameters to obtain situation information; the action response trigger means that the deduction platform receives external instruction input and then triggers corresponding operation units in an operation scene to execute action instructions, specifically take-off, landing, regional patrol, attack, interference, radar switching and the like; the collision response triggering is the collision detection of a unit entity in a simulated battlefield, when the three-dimensional models are overlapped, a collision action is triggered, namely a barrier is formed, and the unit entity with the collision is damaged to different degrees by combining a platform algorithm; the weather response trigger is the influence on the battle environment under different weather conditions, such as the reduction of visual detection distance caused by foggy weather, the damage of plane caused by extreme severe weather thunderstorm, and the like; the communication response triggering refers to the condition for acquiring the information of the own party or the opposite party, the two parties can only acquire the information of a combat unit, a formation, a ground facility, a missile and the like in the detection range of the two parties, and the information in the interference radius range cannot be detected when an jammer drives an interference instruction; the adjudication response trigger refers to the adjudication of whether the whole station is finished or not, whether battle unit ground facilities are damaged or not by reaching the adjudication condition, and we initialize the adjudication rule as: the airplane and the ground are protected against air in the operation unit, the radar is damaged when being hit by the missile once, the moving capability of an engine in a protective ship is lost when being hit, the air defense capability of a weapon is lost when being hit, the damage and sinking are realized after being hit by three accumulated missiles, and the runway of a command post is damaged after being hit by two accumulated missiles; the task completion degree is the progress of the completion of the tasks of the two parties.

The display module includes: main visual angle display, follow visual angle display, free visual angle display, small map display and countermeasure index display; the device is used for visually displaying the situation information. And the display module is used for displaying output by the client.

The algorithm module is used for fighting training and comprises the steps of obtaining decision instructions through the situation information and inputting the decision instructions to the transmission module to enable the transmission module to make corresponding actions, so that excellent strategies after each wheel pair fight are obtained; the algorithm module is a battle training algorithm designed based on the platform, can be designed by self, and can also use the existing algorithm. At present, algorithms for decision problems are all intelligent algorithms based on reinforcement learning, representative algorithms comprise a DQN algorithm, a DDPG algorithm, an Actor-criticic algorithm, a VDN algorithm, a QPLEX algorithm and a QTRAN algorithm, and an algorithm module decides fighting instructions through situation information output by a simulation platform and outputs the fighting instructions to the simulation platform to enable the simulation platform to make corresponding actions. The formation of a mature algorithm requires training to converge, at which point the decision is made as the optimal solution under the algorithm. In fact, we finally observe the AI fighting idea, i.e. the decision output idea after the algorithm convergence.

The transmission module transmits the situation information in the deduction module to the algorithm module, receives a decision instruction from the algorithm module, and transmits the situation information to the display module for display and output;

the population evolution training module uses a population learning algorithm to form a population training strategy by each pair of excellent fighting strategies, and the population training strategy is used as an embedded strategy for the algorithm module to train. The population evolution training module uses a population learning algorithm to carry out the fight between the algorithm module and the self. The algorithm idea is as follows: each wheel pair is updated on the basis of a plurality of previous strong versions after fighting, the versions before updating can be stored and also participate in the later fighting, so that the problem that excellent strategies are restrained mutually is avoided, each wheel optimal strategy forms a population, and finally a plurality of population strategies are obtained and stored in a population learning algorithm library. The stored population training strategy can be used as an algorithm for competing with an external algorithm so as to be used for training the external algorithm.

In another embodiment, the editing content of the initialized parameter task in the scenario module is to set the attacking task to destroy all the forces of the defenders or destroy defending targets of the defenders.

In one embodiment, the information transmitted by the transmission module includes: the information of the combat unit, the visible light visual information, the infrared sensing information and the three-dimensional environment information. The information of the combat units is information of the position, the course, the speed, the residual oil quantity, the residual ammunition quantity, the damage condition, the formation information and the like of the combat units; the visible light senses visible light visual information, including weather information, visual perception information and the like, and is mainly used for the display module; the infrared sensing is sensing information in an infrared mode, and extra heat information can be acquired for accurate positioning; the three-dimensional environment information mainly comprises terrain information and ground facility information.

In one embodiment, the statistics of the countermeasure metrics in the deduction module include: the fighting damage ratio, the resource consumption, the accidental injury ratio and the task completion progress in the battle, wherein the accidental injury ratio refers to the proportion of the randomly occurring neutral party damage amount in the battlefield to the total damage amount. Resource consumption, namely missile consumption quantity, oil consumption and the like, is higher than the proportion of cubic damage to the damage quantity, namely civil ships, civil aviation and the like randomly appearing in a battlefield, namely, more realistic factors are considered, and the deduction accuracy is improved.

In another embodiment, a display module includes: displaying a main visual angle, a follow-up visual angle, a free visual angle, a small map and a confrontation index; the main visual angle display is a fixed global visual angle display; the main visual angle display and the free visual angle display are used as main windows; the following view angle display, the small map display and the confrontation index display are small windows and are attached to the main window. Namely, the main visual angle display is the global visual angle display, a fixed visual angle is set for observing the global combat condition, and the fixed visual angle is used as a main window; the following visual angle display is small window display, which is attached to the main window, an observer can select a certain unit entity in the simulation process, the visual angle of a camera attached to the unit entity can be displayed in the following visual angle window, and at least a plurality of following visual angles of three unit entities are simultaneously viewed in the platform; the free visual angle display means that when the fixed main visual angle is not needed, the free visual angle can be adopted, and after the function is started, an observer can realize the movement of the visual angle by controlling the keyboard and the mouse; displaying a small map, namely displaying two-dimensional information of the battlefield, wherein the two-dimensional information is the overlooking visual angle of the battlefield, the fighting units of different parties are marked by red and blue respectively, and the window is also used as a small window to be attached to the main window; the countermeasure index display is attached to the upper part of the main window as a small window, mainly performs the function similar to a score board, and displays the information of the armed force condition, the task completion degree, the fighting loss ratio, the resource consumption and the like of the two parties of the battle in real time.

In a specific embodiment, the scenario module, the deduction module and the display module are realized by a client; the client comprises a Windows client and a linux client, and the actions of the combat units in the platform are controlled by a human player or an AI algorithm through the client.

In a specific embodiment, the system comprises a server which can be accessed by a client, wherein the server is placed in a Linux system mirror image and is packaged into a docker mirror image file; the server obtains data from the database module.

The algorithm module is designed by an external operator, the platform inputs situation information to the algorithm, and the algorithm outputs decision instructions to control the actions executed by each unit in the platform. The population evolution training module is a population learning algorithm, is a training algorithm embedded in the platform, can be trained to a more excellent strategy by utilizing the existing strategy, is constructed and embedded in the platform, and can be directly called. Namely, the existing strategy algorithm can be evolved into a more excellent strategy algorithm population through the population training algorithm.

A three-dimensional high-simulation war game deduction method comprises the following steps:

s1: initializing parameters of a combat unit, weapon performance, combat force, combat environment and mission, and adopting default parameters if the parameters are not manually set;

s2: respectively constructing and operating an intelligent agent algorithm or a manual rule setting algorithm of both parties of the battle, inputting situation information output by the deduction platform into the algorithm to obtain a decision instruction, and inputting the decision instruction into the platform, wherein the decision instruction changes the situation information;

s3: the situation information is displayed visually in real time;

s4: ending the battle when the task ending condition is reached, and repeating S2;

s5: and outputting the strategy of the intelligent agent algorithm when the algorithms of the two parties of the battle converge.

In one embodiment, in step S2, the algorithm for both parties to battle is constructed using one host alone or two hosts separately.

In another embodiment, the method further comprises the steps of:

s6: and calling a population learning algorithm of the platform according to the converged fighting algorithm to train so as to obtain a population training strategy.

The embodiments in the present description are described in a progressive manner, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description.

The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (10)

1. The utility model provides a three-dimensional high simulation war chess deduction platform which characterized in that includes: the system comprises a planning module, a database module, a deduction module, a transmission module, a display module, an algorithm module and a group evolution training module;

the planning module edits the combat unit, weapon performance, combat force, combat environment and task based on the combat assumption to obtain initialization parameters;

the database module comprises: the system comprises a three-dimensional example model, an environment library, a scenario database, a deduction record database, a game rule database and a population learning algorithm library;

the deduction module is driven by a phantom four-physical engine and is used for triggering action response, collision response, weather response, communication response and judgment response and counting countermeasure indexes according to initialization parameters to obtain situation information;

the display module is used for visually displaying the situation information;

the algorithm module is used for fighting training and comprises the steps of obtaining decision instructions through the situation information and inputting the decision instructions to the transmission module to enable the transmission module to make corresponding actions, so that excellent strategies after each wheel pair fight are obtained;

the transmission module transmits the situation information in the deduction module to the algorithm module, receives a decision instruction from the algorithm module, and transmits the situation information to the display module for display and output;

the population evolution training module uses a population learning algorithm to form a population training strategy by each pair of excellent fighting strategies, and the population training strategy is used as an embedded strategy for training of the algorithm module.

2. The three-dimensional highly-simulated military chess deduction platform according to claim 1, wherein the contents edited by the initialization parameter task in the scenario module are to set an aggressor task to destroy all military forces of a defender or destroy a defender target of the defender.

3. The three-dimensional highly simulated chess-playing platform according to claim 1, wherein the information transmitted by said transmission module comprises: the information of the combat unit, the visible light visual information, the infrared sensing information and the three-dimensional environment information.

4. The three-dimensional highly simulated chess deduction platform according to claim 1, wherein the statistics of the confrontational factors in the deduction module include: the fighting damage ratio, the resource consumption, the accidental injury ratio and the task completion progress in the battle, wherein the accidental injury ratio refers to the proportion of the randomly occurring neutral party damage amount in the battlefield to the total damage amount.

5. The three-dimensional highly simulated chess-playing platform according to claim 1, characterized in that said display module comprises: main visual angle display, follow visual angle display, free visual angle display, small map display and countermeasure index display; the main visual angle display is a fixed global visual angle display; the main visual angle display and the free visual angle display are main windows; the following visual angle display, the small map display and the confrontation index display are small windows and are attached to the main window.

6. The three-dimensional highly-simulated chess deduction platform according to claim 1, wherein the planning module, the deduction module and the display module are realized through a client; the client comprises a Windows client and a linux client, and the actions of the combat units in the platform are controlled by a human player or an AI algorithm through the client.

7. The three-dimensional highly-simulated chess deduction platform according to claim 6, characterized by comprising a server which can be accessed by said client, said server being placed in a Linux system mirror and packing the whole into a docker mirror file; the server obtains data from a database module.

8. A three-dimensional highly simulated chess deduction method according to any one of claims 1-7, characterized by comprising the steps of:

s1: initializing parameters of a combat unit, weapon performance, combat force, combat environment and mission, and adopting default parameters if the parameters are not manually set;

s2: respectively constructing and operating an intelligent agent algorithm or a manual rule setting algorithm of both parties of the battle, inputting situation information output by the deduction platform into the algorithm to obtain a decision instruction, and inputting the decision instruction into the platform decision instruction to change the situation information;

s3: the situation information is displayed visually in real time;

s4: ending the battle when the task ending condition is reached, and repeating S2;

s5: and outputting the strategy of the intelligent agent algorithm when the algorithms of the two parties of the battle converge.

9. The three-dimensional highly simulated chess deduction method according to claim 8, wherein in step S2, one host computer is used alone to construct both algorithms for battle or two host computers are used to construct both algorithms for battle respectively.

10. The three-dimensional highly simulated chess deduction method according to claim 8, characterized by further comprising the steps of:

s6: and calling a population learning algorithm of the platform according to the converged fighting algorithm to train so as to obtain a population training strategy.

Images