CN110489340A - A kind of map balance test method, device, equipment and storage medium - Google Patents
A kind of map balance test method, device, equipment and storage medium Download PDFInfo
- Publication number
- CN110489340A CN110489340A CN201910689920.1A CN201910689920A CN110489340A CN 110489340 A CN110489340 A CN 110489340A CN 201910689920 A CN201910689920 A CN 201910689920A CN 110489340 A CN110489340 A CN 110489340A
- Authority
- CN
- China
- Prior art keywords
- artificial intelligence
- map
- intelligence body
- game
- training
- 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.)
- Granted
Links
Classifications
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63F—CARD, BOARD, OR ROULETTE GAMES; INDOOR GAMES USING SMALL MOVING PLAYING BODIES; VIDEO GAMES; GAMES NOT OTHERWISE PROVIDED FOR
- A63F13/00—Video games, i.e. games using an electronically generated display having two or more dimensions
- A63F13/50—Controlling the output signals based on the game progress
- A63F13/52—Controlling the output signals based on the game progress involving aspects of the displayed game scene
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63F—CARD, BOARD, OR ROULETTE GAMES; INDOOR GAMES USING SMALL MOVING PLAYING BODIES; VIDEO GAMES; GAMES NOT OTHERWISE PROVIDED FOR
- A63F13/00—Video games, i.e. games using an electronically generated display having two or more dimensions
- A63F13/80—Special adaptations for executing a specific game genre or game mode
- A63F13/837—Shooting of targets
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/36—Preventing errors by testing or debugging software
- G06F11/3668—Software testing
- G06F11/3672—Test management
- G06F11/3688—Test management for test execution, e.g. scheduling of test suites
-
- 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
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
Abstract
The present invention relates to game technical field, specifically a kind of map balance test method, device, equipment and storage medium, which comprises obtain the target game map of balance to be tested;Artificial intelligence body set is obtained, the artificial intelligence body in the artificial intelligence body set is directed toward different decision models, and the different decision model carries out machine learning training based on different neural network models and/or different excitation functions and determines;Game is carried out using the target game map using the artificial intelligence body simulation player in the artificial intelligence body set;Obtain the game data that the artificial intelligence body carries out game using the target game map;The income thermodynamic chart of the target game map is generated according to the game data;The income thermodynamic chart is analyzed, determines the balance of the target game map.Map balance test method of the invention can shorten the exploitation proving period of map.
Description
Technical field
The present invention relates to game technical field, in particular to a kind of map balance test method, device, equipment and
Storage medium.
Background technique
With the rapid development of network technology, online game is increasingly liked that especially the first person is penetrated by people
Class game (First-Person Shooter Game, FPS) is hit, due to that can swim with the subjective visual angle of player to be shot at
It plays and is liked deeply by user.Players manipulate the virtual portrait in screen no longer as other game to carry out game, and
It is experience game bring visual impact on the spot in person, this just greatly strengthens the initiative and the sense of reality of game.
In such game, the balance of map is to evaluate an important indicator of a game quality, if trip
Play map the income that balance is bad namely player can obtain in the region of some in map it is excessively high or too low,
It then will lead to player's more or less selection map area, cause entire game unbalance, to reduce the entertaining of game
Property.Therefore, it during development of games, needs to test the balance of map.
In the prior art, usually by the way that map is opened into game, game player is then collected in this trip
The income (killing number) that game is played in play map generates income heating power after collecting tens of thousands of incomes to ten tens of thousands of game of playing a game
Figure, game plan personnel can modify map structure according to the income thermodynamic chart of generation.It is longer to office data needs due to collecting
Time, game plan personnel modify map structure after also need the balance to modified map to test, lead
Cause the exploitation proving period of map longer, and map, without verifying, directly online risk is larger.
Summary of the invention
In view of the above problems in the prior art, the purpose of the present invention is to provide a kind of map balance test sides
Method, device, equipment and storage medium, can shorten the exploitation proving period of map, and reduce map without
Verifying, the directly online game data bring risk for obtaining true player.
To solve the above-mentioned problems, the present invention provides a kind of map balance test method, comprising:
Obtain the target game map of balance to be tested;
Artificial intelligence body set is obtained, the artificial intelligence body in the artificial intelligence body set is directed toward different decision models
Type, the different decision model carry out machine learning instruction based on different neural network models and/or different excitation functions
Practice and determines;
The target game map is used using the artificial intelligence body simulation player in the artificial intelligence body set
Carry out game;
Obtain the game data that the artificial intelligence body carries out game using the target game map;
The income thermodynamic chart of the target game map is generated according to the game data;
The income thermodynamic chart is analyzed, determines the balance of the target game map.
Another aspect of the present invention provides a kind of map balance system safety testing device, comprising:
First obtains module, for obtaining the target game map of balance to be tested;
Second obtains module, the artificial intelligence body for obtaining artificial intelligence body set, in the artificial intelligence body set
It is directed toward different decision models, the different decision model is based on different neural network models and/or different excitation letters
Number carries out machine learning training and determines;
Processing module, for using described in artificial intelligence body simulation player's use in the artificial intelligence body set
Target game map carries out game;
Third obtains module, uses the game of target game map progress game for obtaining the artificial intelligence body
Data;
Generation module, for generating the income thermodynamic chart of the target game map according to the game data;
Analysis module determines the balance of the target game map for analyzing the income thermodynamic chart.
Another aspect of the present invention provides a kind of electronic equipment, and the equipment includes processor and memory, the memory
In be stored at least one instruction, at least a Duan Chengxu, code set or instruction set, at least one instruction, described at least one
Duan Chengxu, the code set or instruction set are loaded by the processor and are executed to realize such as above-mentioned method.
Another aspect of the present invention provides a kind of computer readable storage medium, which is characterized in that deposits in the storage medium
Contain at least one instruction, at least a Duan Chengxu, code set or instruction set, at least one instruction, an at least Duan Cheng
Sequence, the code set or instruction set are loaded by processor and are executed to realize such as above-mentioned method.
Due to above-mentioned technical proposal, the invention has the following advantages:
Map balance test method, device, equipment and storage medium of the invention, by using deeply
Algorithm and course learning mode are practised using the multiple artificial intelligence bodies for being directed toward different decision models of target game map training, and
Game is carried out using the target game map using trained artificial intelligence body simulation player, obtains the trip of more innings of game
It plays data, and generates according to the game data thermodynamic chart of the target game map, and then with determining the target game
The balance of figure can shorten the exploitation proving period of map, and auxiliary project staff has found the defect of map in advance,
And map is reduced without verifying, the directly online game data bring risk for obtaining true player.
Detailed description of the invention
It, below will be to required in embodiment or description of the prior art in order to illustrate more clearly of technical solution of the present invention
The attached drawing used is briefly described.It should be evident that drawings in the following description are only some embodiments of the invention, it is right
For those of ordinary skill in the art, without creative efforts, it can also be obtained according to these attached drawings
Its attached drawing.
Fig. 1 is the flow chart of play map balance test method provided by one embodiment of the present invention;
Fig. 2 is the flow chart for the map balance test method that another embodiment of the present invention provides;
Fig. 3 is the flow chart for the map balance test method that another embodiment of the present invention provides;
Fig. 4 is the schematic diagram of neural network model provided by one embodiment of the present invention;
Fig. 5 is the schematic diagram of deeply learning training framework provided by one embodiment of the present invention;
Fig. 6 is the schematic diagram for the income thermodynamic chart that one embodiment of the invention is related to;
Fig. 7 is the structural schematic diagram of map balance system safety testing device provided by one embodiment of the present invention;
Fig. 8 is the structural schematic diagram for the map balance system safety testing device that another embodiment of the present invention provides;
Fig. 9 is the structural schematic diagram of server provided by one embodiment of the present invention.
Specific embodiment
Following will be combined with the drawings in the embodiments of the present invention, and technical solution in the embodiment of the present invention carries out clear, complete
Site preparation description, it is clear that described embodiments are only a part of the embodiments of the present invention, instead of all the embodiments.It is based on
Embodiment in the present invention, those of ordinary skill in the art without making creative work it is obtained it is all its
His embodiment, shall fall within the protection scope of the present invention.
It should be noted that description and claims of this specification and term " first " in above-mentioned attached drawing, "
Two " etc. be to be used to distinguish similar objects, without being used to describe a particular order or precedence order.It should be understood that using in this way
Data be interchangeable under appropriate circumstances, so as to the embodiment of the present invention described herein can in addition to illustrating herein or
Sequence other than those of description is implemented.In addition, term " includes " and " having " and their any deformation, it is intended that cover
Covering non-exclusive includes to be not necessarily limited to for example, containing the process, method of a series of steps or units, device, product or equipment
Step or unit those of is clearly listed, but may include be not clearly listed or for these process, methods, product
Or other step or units that equipment is intrinsic.
In order to which objects, technical solutions and advantages disclosed by the embodiments of the present invention are more clearly understood, below in conjunction with attached drawing
And embodiment, the embodiment of the present invention is further elaborated.It should be appreciated that specific embodiment described herein is only used
To explain the embodiment of the present invention, it is not intended to limit the present invention embodiment.Firstly, the embodiment of the present invention solves following concepts
It releases:
Artificial intelligence: Artificial Intelligence (AI), artificial intelligence are to utilize digital computer or number
Machine simulation, extension and the intelligence for extending people of computer control, perception environment are obtained knowledge and are most preferably tied using Knowledge Acquirement
Theory, method, technology and the application system of fruit.In other words, artificial intelligence is a complex art of computer science, it is looked forward to
Figure understands the essence of intelligence, and produces a kind of new intelligence machine that can be made a response in such a way that human intelligence is similar.People
Work intelligently namely studies the design principle and implementation method of various intelligence machines, and machine is made to have perception, reasoning and decision
Function.
Deeply study: Deep Reinforcement Learning (DRL), deeply learn deep learning
Sensing capability and the decision-making capability of intensified learning combine, be the artificial intelligence approach of closer mankind thought mode a kind of.
Neural network: artificial neural network (Artificial Neural Networks, ANN), is by numerous nerves
The adjustable connection weight of member is formed by connecting, and has MPP, distributed information storage, good self-organizing self study
The features such as ability.
Course learning: Curriculum Training, course learning is similar to the Learning in School of the mankind, by a branch of instruction in school
Content the courses of different grade of difficulty is divided into according to complexity, improve grade of difficulty step by step in the training process and instructed
Practice.
Figure of description 1 is please referred to, it illustrates map balance provided by one embodiment of the present invention test sides
The process of method, the map balance test method can be applied to map balance provided in an embodiment of the present invention and survey
Trial assembly is set, and map balance system safety testing device is configured in electronic equipment, the electronic equipment can be terminal or
Person's server.Wherein, terminal can be smart phone, desktop computer, tablet computer, laptop etc. with various operations system
The hardware device of system.Server may include an independently operated server perhaps distributed server or by multiple clothes
The server cluster of business device composition.
It should be noted that present description provides the method operating procedures as described in embodiment or flow chart, but it is based on
Routine may include more or less operating procedure without creative labor.The step of enumerating in embodiment sequence is only
Only one of numerous step execution sequence mode does not represent and unique executes sequence.System or product in practice is held
When row, can be executed according to embodiment or method shown in the drawings sequence or it is parallel execute (such as parallel processor or
The environment of multiple threads).As shown in Figure 1, the method may include following steps:
S110: the target game map of balance to be tested is obtained.
In the embodiment of the present invention, the target game map can be the new game of map project staff design
Figure needs to test the being balanced property of map before opening into the map in game.
S120: obtaining artificial intelligence body set, and the artificial intelligence body in the artificial intelligence body set is directed toward different determine
Plan model, the different decision model carry out engineering based on different neural network models and/or different excitation functions
Training is practised to determine.
In the embodiment of the present invention, the decision model of the artificial intelligence body includes at least one full articulamentum.It is described artificial
Artificial intelligence body in intelligent body set is directed toward different decision models, so that artificial intelligence body different in game process can
It, being capable of more preferable simulation real gaming player progress game to there is different decisions.
In practical applications, the artificial intelligence body in the artificial intelligence body set can be by constructing different nerve nets
Network model and/or different excitation functions are determined by identical machine learning training.
In a possible embodiment, the method can also include using the artificial intelligence of target game map training
It can body.As shown in Fig. 2, described may comprise steps of using the artificial intelligent body of target game map training:
S210: the training mission of the artificial intelligence body is divided into multiple grade of difficulty.
In the embodiment of the present invention, the operation and map due to game player are all more complicated, can use course
The artificial intelligent body of mode training of habit, is divided into multiple grade of difficulty, each difficulty for the training mission of the artificial intelligence body
Different training map and training objective is arranged in grade, improves grade of difficulty step by step in the training process and is trained.
S220: being that the training mission of each grade of difficulty matches corresponding trained map from the target game map.
In the embodiment of the present invention, since the training objective of the training mission of each grade of difficulty is different, instruction to be used is needed
It is also different to practice map, simple map can be used for simple training mission and be trained, with training difficulty
It is promoted, training map can also become more complicated.
In one example, it by taking FPS game as an example, attacked since the artificial intelligence body needs to learn shooting, hide enemy
A series of tactical operations such as map are explored in the attack of people.In training process, it can allow it first can basic operation from simple cartography
(such as upper bullet, attack etc.), have these basic operations again by the artificial intelligence body be put into more complicated map into
Row training association complex operations (such as exploring map).Specifically, training mission can be divided into 5 grade of difficulty, instructed
Grade of difficulty is improved during practicing step by step to be trained:
First order difficulty: shooting attacks no motion of enemy and is limited in the artificial intelligence body in first order difficulty
Some region of map is trained, and enemy is not moved will not attack in map.The artificial intelligence body
Need to learn to find in training that no motion of enemy and shooting attack no motion of enemy.
Second level difficulty: the artificial intelligence body is limited in ground in the difficulty of the second level by the enemy of shooting attack movement
Some region of figure is trained, and makes enemy's random movement in map.The artificial intelligence body needs in training middle school
The enemy of random movement can be attacked.
Third level difficulty: the enemy that shooting attack is moved and can be struck back, in third level difficulty, by the artificial intelligence body
Some region for being limited in map is trained, and make enemy in map not only can random movement, but also find the people
Enemy can make attack after work intelligent body.The artificial intelligence body needs to learn how that attack enemy's is same in training
When, hide the attack of enemy.
Fourth stage difficulty: small-scale simple map attack in fourth stage difficulty, expands the activity of the artificial intelligence body
Region allows the artificial intelligence body to learn how to explore in map, find enemy using fairly simple map.
Level V difficulty: the attack of large-scale complex map using complicated map, allows the people in level V difficulty
Complicated map is explored by work intelligent body association.
S230: being successively directed to the training mission of each grade of difficulty, is based on deeply learning algorithm, uses the training
Map is trained the artificial intelligence body.
In the embodiment of the present invention, the training mission of lower difficulty can be first carried out, the training mission of lower difficulty has been trained
Cheng Houzai carries out more highly difficult training mission, until all training missions are completed.
In a possible embodiment, as shown in figure 3, the training mission for being successively directed to each grade of difficulty, base
In deeply learning algorithm, being trained using the trained map to the artificial intelligence body be may comprise steps of:
S231: the neural network model and excitation function of the artificial intelligence body are constructed.
In the embodiment of the present invention, the neural network model and excitation function can be constructed according to the actual situation, building is not
Same excitation function may result in the neural network model study to different strategies.In practical applications, can pass through
Different neural network models and/or different excitation functions are constructed to train the artificial intelligence for being directed toward different decision models
Body.
S232: it is obtained according to the neural network model and the excitation function and utilizes the artificial intelligence body simulation
The training sample data that player is trained using the trained map.
It is described to be obtained according to the neural network model and the excitation function using described artificial in the embodiment of the present invention
Intelligent body simulation player may include: using the training sample data that the trained map is trained
During a game running process, the current status data of the artificial intelligence body is obtained;
The current status data is inputted in the neural network model of the artificial intelligence body, generates the current state
The corresponding decision data of data;
The decision movement of the artificial intelligence body is determined according to the decision data, and is controlled the artificial intelligence body and executed
The decision movement;
Determine that the artificial intelligence body executes the reward data after the decision acts according to the excitation function;
It obtains the artificial intelligence body and executes the succeeding state data after the decision movement;
By the current status data, the decision data, the reward data and the succeeding state data organization at
Training sample data.
In one example, by taking FPS game as an example, as shown in the table, the current status data of the artificial intelligence body can
To include the personal attribute of the artificial intelligence body, shield the information such as object and map datum together, the personal attribute may include
Blood volume, ammunition number kill several and visual angle etc., and described with screen object may include enemy position etc., and the map datum can wrap
It includes map and describes file, teammate location coordinate and oneself position coordinates etc..
The input of the neural network model is the current status data, may include the blood of the artificial intelligence body
Amount, ammunition number kill several and visual angle, enemy position and map and describe file, teammate location coordinate and oneself position coordinates.Its
In, the blood volume can be 8 or be by the numerical value blood volume that perhaps percentage is indicated such as artificial intelligence body
75%, the map describes file can be indicated by the color of pixel, such as map is 229 × 229 map, institute
It states map and describes file one and share 229 × 229 × 3 values.As shown in figure 4, it illustrates offer of the embodiment of the present invention
The schematic diagram of neural network model it is multiple not need design since the input of the neural network model is one group of vector
Miscellaneous convolutional neural networks, it is only necessary to design full articulamentum.
The current status data is inputted in the neural network model, decision data, the decision number can be exported
According to the action data that may include the artificial intelligence body, as shown in the table, the action data may include the artificial intelligence
The movement of energy body, attack, visual angle adjustment, the switching information such as weapon and hopped data.
After getting the decision data of output, the decision of the artificial intelligence body can be determined according to the decision data
Movement, and control the artificial intelligence body and execute the decision movement.It, can after the artificial intelligence body executes the decision movement
To obtain succeeding state data, the particular content of the succeeding state data is identical as the current status data, herein not
It repeats again.
After the artificial intelligence body executes the decision movement, the artificial intelligence can also be determined according to the excitation function
Energy body executes the reward data after the decision movement.Specifically, the excitation function can be set to consider the artificial intelligence
The ammunition number of energy body kills number, blood volume, time-to-live and the number killed, can indicate are as follows:
F (x)=w1×x1+w2×x2+w3×x3+w4×x4+w5×x5
Wherein, x1,x2,x3,x4,x5It respectively represents ammunition number, kill number, blood volume, time-to-live and the number killed,
w1,w2,w3,w4,w5It respectively represents ammunition number, kill the power of number, blood volume, the number killed and time-to-live in excitation function
Weight, can obtain different excitation functions by adjusting the weight.
For example, setting w1,w2,w3,w4,w5For (- 0.01,1,0,0,0), this group of parameter is indicated if the artificial intelligence body
Shooting consumption ammunition can obtain -0.01 excitation, if the artificial intelligence body shoots an enemy to death and can obtain 1
Excitation, and the blood volume of the artificial intelligence body, time-to-live and to be killed the excitation of acquisition be all 0, this indicates described artificial intelligence
Energy body is killed or oligemia under attack will not all pay for.Therefore the artificial intelligence body can more incline in training
To in attack enemy because very high excitation can be obtained by killing enemy.If w is arranged1,w2,w3,w4,w5For (- 0.01,0.1,0,
0.4, -1), this group of parameter indicates if artificial intelligence body shooting consumption ammunition can obtain -0.01 excitation, if institute
0.1 excitation can be obtained by stating artificial intelligence body and shooting an enemy to death, while if artificial intelligence body survival when
Between longer can obtain higher excitation (because of w4For 0.4), can be obtained if the artificial intelligence body is killed -1 swash
It encourages.Therefore the artificial intelligence body is more likely to protect oneself in training, hides the attack of enemy first, because being hit by enemy
Very high negative energize can be obtained after killing.
Therefore, the training sample data may include the current blood volume of artificial intelligence body, current ammunition number, currently kill
Several and current visual angle, current enemy position and current map describe file, current teammate location coordinate and oneself current position
Coordinate, movement, attack, visual angle adjustment, switching weapon and the hopped data of the artificial intelligence body, the artificial intelligence body execution
The excitation and the artificial intelligence body obtained after decision movement executes the subsequent blood volume after the decision movement, subsequent
Ammunition number, subsequent several and subsequent visual angle, subsequent enemy position and the subsequent map of killing describe file, subsequent teammate location coordinate
With the data such as oneself subsequent position coordinates.
S233: it is based on deeply learning algorithm, using the training sample data to the nerve of the artificial intelligence body
Network model is trained, and adjusts the parameter of the neural network model, obtains the decision model of the artificial intelligence body.
In the embodiment of the present invention, suitable deeply learning algorithm, including depth Q net can be selected according to actual needs
Network (Deep Q Network, DQN) algorithm, depth deterministic policy gradient (Deep Deterministic Policy
Gradient, DDPG) algorithm and asynchronous advantage performer-reviewer (Asynchronous AdvantageActor-Critic,
A3C) algorithm etc., wherein the DQN algorithm, DDPG algorithm and the A3C algorithm are prior art content, and the present invention is implemented
Example is no longer repeated herein.
It in one example, can be using deeply learning training framework as shown in Figure 5 to the artificial intelligence body
Neural network model be trained.As shown in figure 5, the trained framework may include game server 510, parameter server
520 and at least one host 530 for being communicated to connect respectively with the game server 510 and the parameter server 520.Institute
Parameter server 520 is stated for storing the parameter of multiple neural network models, operation has multiple artificial on each host 530
The training process of intelligent body, artificial intelligence body on each host 530 use identical neural network model and identical
Excitation function is trained, then can use A3C algorithm to improve the training speed of the artificial intelligence body;Different hosts 530
On artificial intelligence body be trained using different network models and/or different excitation functions, then can train simultaneously more
A artificial intelligence body for being directed toward different decision models.
S130: the target game is used using the artificial intelligence body simulation player in the artificial intelligence body set
Map carries out game.
In the embodiment of the present invention, the artificial intelligence body simulation of different number can be used according to the type difference of game
Player carries out game.For example, can only use single artificial intelligent body simulation player in solitaire game carries out game, more people
The member of game war team can be the game player of artificial intelligence body simulation in battle game.The artificial intelligence body is in game
Current status data can be obtained in real time in the process, the current status data is inputted to the decision model of the artificial intelligence body
In, the corresponding decision data of the current status data is generated, determining for the artificial intelligence body is determined according to the decision data
It instigates to make, and executes the decision movement, game is carried out with simulation player.Due to the people in the artificial intelligence body set
Work intelligent body is directed toward different decision models, therefore the artificial intelligence body can take different decisions in the state of same
Movement, simulation are more in line with actual conditions.
In practical applications, the artificial intelligence body can using and game server it is directly interactive by the way of obtain it is current
Status data does not need rendering game client, can significantly speed up artificial intelligence body simulation player and carry out game
Process.Furthermore it is also possible to the concurrency for promoting game in such a way that multiple artificial intelligence bodies carry out game parallel, thus
It realizes and is completed within a short period of time to tens of thousands of or even hundreds of thousands part game data needed for a map balance test
Demand.
S140: the game data that the artificial intelligence body carries out game using the target game map is obtained.
In the embodiment of the present invention, it can use the artificial intelligence body in the artificial intelligence body set and simulate more innings of game,
And obtain the game data of more innings of game.
S150: the income thermodynamic chart of the target game map is generated according to the game data.
In a possible embodiment, the income heat that the target game map is generated according to the game data
Try hard to may include:
Determine that the artificial intelligence body obtains the location information of income according to the game data;
The location information is mapped in default map grid corresponding with the target game map;
Count the quantity for the corresponding income of location information that each grid includes;
Mapping each grid is color corresponding with the quantity, generates income thermodynamic chart.
In one example, by taking FPS game as an example, the location information that the artificial intelligence body obtains income includes the people
The location information for the enemy that the location information and the artificial intelligence body that work intelligent body is killed kill.It is attached to please refer to specification
Fig. 6 illustrates the income thermodynamic chart that one embodiment of the invention is related to, as shown in fig. 6, more partially bright region indicates
The income that actual player can obtain in the region is higher.
S160: analyzing the income thermodynamic chart, determines the balance of the target game map.
In the embodiment of the present invention, map designer can determine the map by the income thermodynamic chart
Balance can modify, if the map does not meet the expection of map designer to modified
Map can continue to use map balance test method as shown in Figure 1 and be tested, until the game
Until figure reaches the expection of map designer.
In conclusion map balance test method of the invention, by using deeply learning algorithm and class
Journey mode of learning uses training using the multiple artificial intelligence bodies for being directed toward different decision models of target game map training
Artificial intelligence body simulation player carries out game using the target game map, obtains the game data of more innings of game, and
The thermodynamic chart of the target game map is generated according to the game data, and then determines the balance of the target game map
Property, the exploitation proving period of map can be shortened, auxiliary project staff has found the defect of map in advance, and reduces
Map is without verifying, the directly online game data bring risk for obtaining true player.
Figure of description 7 is please referred to, it illustrates map balance provided by one embodiment of the present invention tests to fill
700 structural schematic diagram is set, described device 700 may include:
First obtains module 710, for obtaining the target game map of balance to be tested;
Second obtains module 720, the artificial intelligence for obtaining artificial intelligence body set, in the artificial intelligence body set
Body is directed toward different decision models, and the different decision model is based on different neural network models and/or different excitations
Function carries out machine learning training and determines;
Processing module 730, for being used using the artificial intelligence body simulation player in the artificial intelligence body set
The target game map carries out game;
Third obtains module 740, carries out game using the target game map for obtaining the artificial intelligence body
Game data;
Generation module 750, for generating the income thermodynamic chart of the target game map according to the game data;
Analysis module 760 determines the balance of the target game map for analyzing the income thermodynamic chart
Property.
In a possible embodiment, described device 700 can also include training module 770, as shown in figure 8, described
Training module 770 may include:
Division unit 771, for the training mission of the artificial intelligence body to be divided into multiple grade of difficulty;
Matching unit 772, the training mission matching for from the target game map being each grade of difficulty correspond to
Training map;
Training unit 773, for successively for the training mission of each grade of difficulty, being based on deeply learning algorithm,
The artificial intelligence body is trained using the trained map.
In a possible embodiment, the training unit 773 may include:
Subelement is constructed, for constructing the neural network model and excitation function of the artificial intelligence body;
Subelement is obtained, utilizes the artificial intelligence for obtaining according to the neural network model and the excitation function
The training sample data that body simulation player is trained using the trained map;
Training subelement, for being based on deeply learning algorithm, using the training sample data to the artificial intelligence
The neural network model of energy body is trained, and adjusts the parameter of the neural network model, obtains determining for the artificial intelligence body
Plan model.
In a possible embodiment, the generation module 750 may include:
Determination unit, for determining that the artificial intelligence body obtains the location information of income according to the game data;
Map unit, for the location information to be mapped to default map grid corresponding with the target game map
In;
Statistic unit, for counting the quantity for the corresponding income of location information that each grid includes;
Generation unit is color corresponding with the quantity for mapping each grid, generates income thermodynamic chart.
It should be noted that device provided by the above embodiment, when realizing its function, only with above-mentioned each functional module
It divides and carries out for example, can according to need in practical application and be completed by different functional modules above-mentioned function distribution,
The internal structure of equipment is divided into different functional modules, to complete all or part of the functions described above.
The embodiment of the invention provides a kind of electronic equipment, the electronic equipment includes processor and memory, described to deposit
Be stored at least one instruction, at least a Duan Chengxu, code set or instruction set in reservoir, at least one instruction, it is described extremely
A few Duan Chengxu, the code set or instruction set are loaded by the processor and are executed to realize that above method embodiment such as provides
Map balance test method.
Memory can be used for storing software program and module, and processor is stored in the software program of memory by operation
And module, it is tested thereby executing various function application and map balance.Memory can mainly include storage program
Area and storage data area, wherein storing program area can application program needed for storage program area, function etc.;Storage data area
It can store and created data etc. are used according to the equipment.In addition, memory may include high-speed random access memory,
It can also include nonvolatile memory, a for example, at least disk memory, flush memory device or other volatile solid-states are deposited
Memory device.Correspondingly, memory can also include Memory Controller, to provide access of the processor to memory.
Embodiment of the method provided by the embodiment of the present invention can be filled in terminal, server or similar operation
Middle execution is set, i.e., above-mentioned electronic equipment may include terminal, server or similar arithmetic unit.To operate in clothes
For on business device, Fig. 9 is a kind of the hard of the server provided in an embodiment of the present invention for running map balance test method
Part structural block diagram.As shown in figure 9, the server 900 can generate bigger difference because configuration or performance are different, may include
One or more central processing units (Central Processing Units, CPU) 910 (processor 910 may include but
Be not limited to the processing unit of Micro-processor MCV or programmable logic device FPGA etc.), memory 930 for storing data, one
(such as one or more mass memories of storage medium 920 of a or more than one storage application program 923 or data 922
Equipment).Wherein, memory 930 and storage medium 920 can be of short duration storage or persistent storage.It is stored in storage medium 920
Program may include one or more modules, and each module may include to the series of instructions operation in server.More
Further, central processing unit 910 can be set to communicate with storage medium 920, execute storage medium on server 900
Series of instructions operation in 920.Server 900 can also include one or more power supplys 960, one or more
Wired or wireless network interface 950, one or more input/output interfaces 940, and/or, one or more operations
System 921, such as Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM, FreeBSDTM etc..
Input/output interface 940 can be used for that data are received or sent via a network.Above-mentioned network is specifically real
Example may include the wireless network that the communication providers of server 900 provide.In an example, input/output interface 940 includes
One network adapter (Network Interface Controller, NIC), can pass through base station and other network equipment phases
Even so as to be communicated with internet.In an example, input/output interface 940 can be radio frequency (Radio
Frequency, RF) module, it is used to wirelessly be communicated with internet.
It will appreciated by the skilled person that structure shown in Fig. 9 is only to illustrate, not to above-mentioned electronic device
Structure cause to limit.For example, server 900 may also include than shown in Fig. 9 more perhaps less component or have with
Different configuration shown in Fig. 9.
The embodiments of the present invention also provide a kind of computer readable storage medium, the storage medium may be disposed at service
To save for realizing a kind of relevant at least one instruction of map balance test method, an at least Duan Cheng among device
Sequence, code set or instruction set, at least one instruction, an at least Duan Chengxu, the code set or instruction set are by handling
Device loads and executes the map balance test method to realize above method embodiment offer.
Optionally, in the present embodiment, above-mentioned storage medium can include but is not limited to: USB flash disk, read-only memory (ROM,
Read-Only Memory), random access memory (RAM, RandomAccessMemory), mobile hard disk, magnetic disk or light
The various media that can store program code such as disk.
It should be understood that embodiments of the present invention sequencing is for illustration only, do not represent the advantages or disadvantages of the embodiments.
And above-mentioned this specification specific embodiment is described.Other embodiments are within the scope of the appended claims.One
In a little situations, the movement recorded in detail in the claims or step can be executed according to the sequence being different from embodiment and
Still desired result may be implemented.In addition, process depicted in the drawing not necessarily requires the particular order shown or company
Continuous sequence is just able to achieve desired result.In some embodiments, multitasking and parallel processing it is also possible or
It may be advantageous.
All the embodiments in this specification are described in a progressive manner, same and similar portion between each embodiment
Dividing may refer to each other, and each embodiment focuses on the differences from other embodiments.Especially for device,
For equipment and server example, since it is substantially similar to the method embodiment, so being described relatively simple, related place
Illustrate referring to the part of embodiment of the method.
Those of ordinary skill in the art will appreciate that realizing that all or part of the steps of above-described embodiment can pass through hardware
It completes, relevant hardware can also be instructed to complete by program, the program can store in a kind of computer-readable
In storage medium, storage medium mentioned above can be read-only memory, disk or CD etc..
The foregoing is merely presently preferred embodiments of the present invention, is not intended to limit the invention, it is all in spirit of the invention and
Within principle, any modification, equivalent replacement, improvement and so on be should all be included in the protection scope of the present invention.
Claims (10)
1. a kind of map balance test method characterized by comprising
Obtain the target game map of balance to be tested;
Artificial intelligence body set is obtained, the artificial intelligence body in the artificial intelligence body set is directed toward different decision models, institute
It states different decision models and machine learning training is carried out really based on different neural network models and/or different excitation functions
It is fixed;
It is carried out using the artificial intelligence body simulation player in the artificial intelligence body set using the target game map
Game;
Obtain the game data that the artificial intelligence body carries out game using the target game map;
The income thermodynamic chart of the target game map is generated according to the game data;
The income thermodynamic chart is analyzed, determines the balance of the target game map.
2. the method according to claim 1, wherein further including using the artificial intelligence of target game map training
Energy body, it is described to include: using the artificial intelligent body of target game map training
The training mission of the artificial intelligence body is divided into multiple grade of difficulty;
It is that the training mission of each grade of difficulty matches corresponding trained map from the target game map;
It is successively directed to the training mission of each grade of difficulty, deeply learning algorithm is based on, using the trained map to institute
Artificial intelligence body is stated to be trained.
3. according to the method described in claim 2, it is characterized in that, it is described successively be directed to each grade of difficulty training mission,
Based on deeply learning algorithm, the artificial intelligence body is trained using the trained map includes:
Construct the neural network model and excitation function of the artificial intelligence body;
It is obtained according to the neural network model and the excitation function and is used using the artificial intelligence body simulation player
The training sample data that the trained map is trained;
Based on deeply learning algorithm, using the training sample data to the neural network model of the artificial intelligence body into
Row training, adjusts the parameter of the neural network model, obtains the decision model of the artificial intelligence body.
4. according to the method described in claim 3, it is characterized in that, described according to the neural network model and the excitation letter
Number obtains the training sample data packet being trained using the artificial intelligence body simulation player using the trained map
It includes:
During a game running process, the current status data of the artificial intelligence body is obtained;
The current status data is inputted in the neural network model of the artificial intelligence body, generates the current status data
Corresponding decision data;
The decision movement of the artificial intelligence body is determined according to the decision data, and is controlled described in the artificial intelligence body execution
Decision movement;
Determine that the artificial intelligence body executes the reward data after the decision acts according to the excitation function;
It obtains the artificial intelligence body and executes the succeeding state data after the decision movement;
By the current status data, the decision data, the reward data and the succeeding state data organization at training
Sample data.
5. method according to claim 1 or 2, which is characterized in that the decision model of the artificial intelligence body includes at least
One full articulamentum.
6. method according to claim 1 or 2, which is characterized in that described to generate the target according to the game data
The income thermodynamic chart of map includes:
Determine that the artificial intelligence body obtains the location information of income according to the game data;
The location information is mapped in default map grid corresponding with the target game map;
Count the quantity for the corresponding income of location information that each grid includes;
Mapping each grid is color corresponding with the quantity, generates income thermodynamic chart.
7. a kind of map balances system safety testing device characterized by comprising
First obtains module, for obtaining the target game map of balance to be tested;
Second obtains module, and for obtaining artificial intelligence body set, the artificial intelligence body in the artificial intelligence body set is directed toward
Different decision models, the different decision model based on different neural network models and/or different excitation functions into
Row machine learning training determines;
Processing module, for using the target using the artificial intelligence body simulation player in the artificial intelligence body set
Map carries out game;
Third obtains module, uses the game number of target game map progress game for obtaining the artificial intelligence body
According to;
Generation module, for generating the income thermodynamic chart of the target game map according to the game data;
Analysis module determines the balance of the target game map for analyzing the income thermodynamic chart.
8. device according to claim 7, which is characterized in that further include training module, the training module includes:
Division unit, for the training mission of the artificial intelligence body to be divided into multiple grade of difficulty;
Matching unit, for being the corresponding training ground of training mission matching of each grade of difficulty from the target game map
Figure;
Training unit is based on deeply learning algorithm, using described for being successively directed to the training mission of each grade of difficulty
Training map is trained the artificial intelligence body.
9. a kind of electronic equipment, which is characterized in that the equipment includes processor and memory, be stored in the memory to
Few an instruction, at least a Duan Chengxu, code set or instruction set, it is at least one instruction, an at least Duan Chengxu, described
Code set or instruction set are loaded by the processor and are executed to realize method as claimed in any one of claims 1 to 6.
10. a kind of computer readable storage medium, which is characterized in that be stored at least one instruction, extremely in the storage medium
A few Duan Chengxu, code set or instruction set, at least one instruction, an at least Duan Chengxu, the code set or instruction
Collection is loaded by processor and is executed to realize method as claimed in any one of claims 1 to 6.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910689920.1A CN110489340B (en) | 2019-07-29 | 2019-07-29 | Game map balance testing method, device, equipment and storage medium |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910689920.1A CN110489340B (en) | 2019-07-29 | 2019-07-29 | Game map balance testing method, device, equipment and storage medium |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110489340A true CN110489340A (en) | 2019-11-22 |
CN110489340B CN110489340B (en) | 2021-06-15 |
Family
ID=68548590
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910689920.1A Active CN110489340B (en) | 2019-07-29 | 2019-07-29 | Game map balance testing method, device, equipment and storage medium |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110489340B (en) |
Cited By (8)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110898433A (en) * | 2019-11-28 | 2020-03-24 | 腾讯科技(深圳)有限公司 | Virtual object control method and device, electronic equipment and storage medium |
CN111111204A (en) * | 2020-04-01 | 2020-05-08 | 腾讯科技(深圳)有限公司 | Interactive model training method and device, computer equipment and storage medium |
CN111388995A (en) * | 2020-03-02 | 2020-07-10 | 腾讯科技(深圳)有限公司 | Game artificial intelligence information processing method, system, device and storage medium |
CN111443806A (en) * | 2020-03-26 | 2020-07-24 | 腾讯科技(深圳)有限公司 | Interactive task control method and device, electronic equipment and storage medium |
CN112221151A (en) * | 2020-10-23 | 2021-01-15 | 腾讯科技(深圳)有限公司 | Map generation method and device, computer equipment and storage medium |
CN112263835A (en) * | 2020-11-16 | 2021-01-26 | 网易(杭州)网络有限公司 | Game balance testing method and device, electronic equipment and storage medium |
CN112370790A (en) * | 2020-12-01 | 2021-02-19 | 腾讯科技(深圳)有限公司 | Game map drawing method and device, electronic equipment and storage medium |
CN112783781A (en) * | 2021-01-28 | 2021-05-11 | 网易(杭州)网络有限公司 | Game testing method and device, electronic equipment and storage medium |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20160343211A1 (en) * | 2015-05-18 | 2016-11-24 | Gameco, Inc. | Video game gaming system |
CN106503787A (en) * | 2016-10-26 | 2017-03-15 | 腾讯科技(深圳)有限公司 | A kind of method for obtaining game data and electronic equipment |
CN106621336A (en) * | 2015-10-30 | 2017-05-10 | 网易(杭州)网络有限公司 | FPS (first-person shooting) game scene map design visual verification method and system |
US20170157518A1 (en) * | 2015-12-04 | 2017-06-08 | Sony Interactive Entertainment America Llc | Balancing multiple team based games |
CN107982920A (en) * | 2017-11-28 | 2018-05-04 | 上海波克城市网络科技股份有限公司 | Become more meticulous adjusting of difficulty method based on the game of outpost of the tax office class |
CN109815155A (en) * | 2019-02-26 | 2019-05-28 | 网易(杭州)网络有限公司 | A kind of method and device of game test, electronic equipment, storage medium |
CN110052031A (en) * | 2019-04-11 | 2019-07-26 | 网易(杭州)网络有限公司 | The imitation method, apparatus and readable storage medium storing program for executing of player |
-
2019
- 2019-07-29 CN CN201910689920.1A patent/CN110489340B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20160343211A1 (en) * | 2015-05-18 | 2016-11-24 | Gameco, Inc. | Video game gaming system |
CN106621336A (en) * | 2015-10-30 | 2017-05-10 | 网易(杭州)网络有限公司 | FPS (first-person shooting) game scene map design visual verification method and system |
US20170157518A1 (en) * | 2015-12-04 | 2017-06-08 | Sony Interactive Entertainment America Llc | Balancing multiple team based games |
CN106503787A (en) * | 2016-10-26 | 2017-03-15 | 腾讯科技(深圳)有限公司 | A kind of method for obtaining game data and electronic equipment |
CN107982920A (en) * | 2017-11-28 | 2018-05-04 | 上海波克城市网络科技股份有限公司 | Become more meticulous adjusting of difficulty method based on the game of outpost of the tax office class |
CN109815155A (en) * | 2019-02-26 | 2019-05-28 | 网易(杭州)网络有限公司 | A kind of method and device of game test, electronic equipment, storage medium |
CN110052031A (en) * | 2019-04-11 | 2019-07-26 | 网易(杭州)网络有限公司 | The imitation method, apparatus and readable storage medium storing program for executing of player |
Non-Patent Citations (1)
Title |
---|
郝爱峰: "具有平衡性的历史策略游戏系统初探", 《科技情报开发与经济》 * |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110898433A (en) * | 2019-11-28 | 2020-03-24 | 腾讯科技(深圳)有限公司 | Virtual object control method and device, electronic equipment and storage medium |
CN110898433B (en) * | 2019-11-28 | 2021-09-03 | 腾讯科技(深圳)有限公司 | Virtual object control method and device, electronic equipment and storage medium |
CN111388995A (en) * | 2020-03-02 | 2020-07-10 | 腾讯科技(深圳)有限公司 | Game artificial intelligence information processing method, system, device and storage medium |
CN111388995B (en) * | 2020-03-02 | 2022-07-26 | 腾讯科技(深圳)有限公司 | Game artificial intelligence information processing method, system, device and storage medium |
CN111443806A (en) * | 2020-03-26 | 2020-07-24 | 腾讯科技(深圳)有限公司 | Interactive task control method and device, electronic equipment and storage medium |
CN111443806B (en) * | 2020-03-26 | 2023-08-11 | 腾讯科技(深圳)有限公司 | Interactive task control method and device, electronic equipment and storage medium |
CN111111204A (en) * | 2020-04-01 | 2020-05-08 | 腾讯科技(深圳)有限公司 | Interactive model training method and device, computer equipment and storage medium |
CN111111204B (en) * | 2020-04-01 | 2020-07-03 | 腾讯科技(深圳)有限公司 | Interactive model training method and device, computer equipment and storage medium |
CN112221151A (en) * | 2020-10-23 | 2021-01-15 | 腾讯科技(深圳)有限公司 | Map generation method and device, computer equipment and storage medium |
CN112263835A (en) * | 2020-11-16 | 2021-01-26 | 网易(杭州)网络有限公司 | Game balance testing method and device, electronic equipment and storage medium |
CN112370790A (en) * | 2020-12-01 | 2021-02-19 | 腾讯科技(深圳)有限公司 | Game map drawing method and device, electronic equipment and storage medium |
CN112783781A (en) * | 2021-01-28 | 2021-05-11 | 网易(杭州)网络有限公司 | Game testing method and device, electronic equipment and storage medium |
Also Published As
Publication number | Publication date |
---|---|
CN110489340B (en) | 2021-06-15 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110489340A (en) | A kind of map balance test method, device, equipment and storage medium | |
Vinyals et al. | Starcraft ii: A new challenge for reinforcement learning | |
CN110119773B (en) | Global situation assessment method, system and device of strategic gaming system | |
CN109091868B (en) | Method, apparatus, computer equipment and the storage medium that battle behavior determines | |
Togelius et al. | Controllable procedural map generation via multiobjective evolution | |
CN108211362A (en) | A kind of non-player role fight policy learning method based on depth Q learning networks | |
CN108553903A (en) | Control robot player's method and device | |
Uriarte et al. | Game-tree search over high-level game states in RTS games | |
Stanescu et al. | Hierarchical adversarial search applied to real-time strategy games | |
Mahlmann et al. | Towards procedural strategy game generation: Evolving complementary unit types | |
Stephenson et al. | Procedural generation of complex stable structures for angry birds levels | |
Barriga et al. | Improving RTS game AI by supervised policy learning, tactical search, and deep reinforcement learning | |
CN109847366A (en) | Data for games treating method and apparatus | |
Mahlmann et al. | Spicing up map generation | |
Uriarte et al. | Combat models for RTS games | |
CN109541960A (en) | A kind of system and method for the confrontation of aircraft digital battlefield | |
Adil et al. | State-of-the-art and open challenges in RTS game-AI and Starcraft | |
Perez-Liebana et al. | Generating diverse and competitive play-styles for strategy games | |
Oh et al. | Playing real-time strategy games by imitating human players’ micromanagement skills based on spatial analysis | |
Fernández-Ares et al. | Adaptive bots for real-time strategy games via map characterization | |
Dahlskog et al. | Mapping the game landscape: Locating genres using functional classification | |
Dockhorn et al. | Game state and action abstracting monte carlo tree search for general strategy game-playing | |
Guerrero-Romero et al. | Map-elites to generate a team of agents that elicits diverse automated gameplay | |
Barriga et al. | Building placement optimization in real-time strategy games | |
Mora et al. | Evolving the cooperative behaviour in Unreal™ bots |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant |