CN106055339A - Method for determining card playing strategy of computer player in two-against-one game - Google Patents
Method for determining card playing strategy of computer player in two-against-one game Download PDFInfo
- Publication number
- CN106055339A CN106055339A CN201610406925.5A CN201610406925A CN106055339A CN 106055339 A CN106055339 A CN 106055339A CN 201610406925 A CN201610406925 A CN 201610406925A CN 106055339 A CN106055339 A CN 106055339A
- Authority
- CN
- China
- Prior art keywords
- card
- player
- neutral net
- game
- training data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/448—Execution paradigms, e.g. implementations of programming paradigms
- G06F9/4498—Finite state machines
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/008—Artificial life, i.e. computing arrangements simulating life based on physical entities controlled by simulated intelligence so as to replicate intelligent life forms, e.g. based on robots replicating pets or humans in their appearance or behaviour
-
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield 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
- G06N3/06—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons
- G06N3/061—Physical realisation, i.e. hardware implementation of neural networks, neurons or parts of neurons using biological neurons, e.g. biological neurons connected to an integrated circuit
Abstract
The invention provides a method for determining a card playing strategy of a computer player in a two-against-one game. The method comprises: creating a BP neural network; training the BP neural network by using training data; and when it is the computer player's turn to play a card in a current round of the two-against-one game, invoking the BP neural network to calculate card round parameters in the two-against-one game so as to determine the card playing strategy of the computer player. By implementation of the invention, the card playing strategy of the computer player can adapt to an actual card round situation, then the card playing intelligent level and simulation degree of the computer player can be improved, and the game experience of a real player who plays a game with the computer player is enhanced.
Description
Technical field
The present invention relates to artificial intelligence field, particularly relate to a kind of two and make a call to a game determining, computer player is played a card strategy
Method.
Background technology
Two make a call to the board intermediate item that a game is a kind of physical culture intelligent sports that three players the participate in, " bucket being namely commonly called as
Landlord " cards game.The two basic laws of the game plaing a game are by crying score, make a player become banker, remaining two
Player becomes defender and banker and resists, and judges gambling party victory or defeat and according to row board board with a certain player hands that taken the lead in
Score settled accounts by type.
Along with the development of computer technology, prior art provides online or unit electronization two and makes a call to a game journey
Sequence, further, in unit training mode or compound match rule, occurs in that utilize artificial intelligence technology to realize two play a game
Computer player, this computer player can simulate the row board of true player according to two rules plaing a game, in order to substitutes true
Real player carries out two together with other true players and plays a game.Computer player of the prior art generally uses finite state
Machine realizes, such as Chinese patent " implementation method of intelligent algorithm for computer player in Doudizhu game " (application number
Technical scheme disclosed in 201010246269.X), this technical scheme is exactly that a kind of typical use finite state machine combines something lost
Propagation algorithm realizes the method for two logic structure model making a call to a game Computer player.
The defect of prior art is, due to the characteristic of of finite state machine itself, builds meter even with genetic algorithm
The row board logic of calculation machine player, is also easy to the incomplete problem of algorithm occur, thus causes the level of intelligence of computer player
Low and style of playing a card is single, it is impossible to correctly identify ping-pong situation complicated in gambling party, the defender served as with true player
During cooperation, artificial intelligence's degree of verisimilitude is low.Therefore the true player played is carried out with the computer player using finite state machine to realize
It is difficult to obtain satisfied game experiencing.
Summary of the invention
In order to overcome drawbacks described above of the prior art, the invention provides a kind of two and make a call in a game and determine that computer is played
Family play a card strategy method, the method includes:
Create BP neutral net；
Training data is used to train described BP neutral net；
An innings two currently carried out make a call to one game in take turns to computer player play a card time, call described BP neutral net pair
The described two gambling party parameters made a call in a game calculate, to determine the strategy of playing a card of described computer player.
According to another aspect of the present invention, before using training data to train described BP neutral net, the method is also
Including: pre-stored two makes a call to the gambling party information of true player in a history gambling party played；Generate described according to described gambling party information
Training data.
According to a further aspect of the invention, described in the method, training data includes:
For describing the gaming identity of described true player, hands quantity, hands board type, active/passive play a card state, quilt
Dynamic board type faced by when playing a card, the gaming identity of the player that plays a card faced by when passively playing a card, the board type got, get
Board quantity, the overall situation remain hands quantity, board type of playing a card, cross the arbitrary or variable of a combination thereof and value thereof in board operation.
According to a further aspect of the invention, the method creates BP neutral net to include: be respectively created described BP neural
The input layer of network, hidden layer and output layer；Initialize described input layer and described hidden layer weights and side-play amount, and initially
Change described hidden layer and the weights of described output layer and side-play amount.
According to a further aspect of the invention, the method use training data train described BP neutral net to include: institute
Stating BP neutral net uses back-propagation algorithm to calculate described training data, makes the actual output of described BP neutral net with described
The error of the desired output in training data is less than predetermined threshold；And/or described BP neutral net uses back-propagation algorithm meter
Calculate described training data, make the correction number of times of the actual output of described BP neutral net reach pre-determined number.
According to a further aspect of the invention, the method use training data train the step of described BP neutral net to make
Realize by Distributed Calculation.
According to a further aspect of the invention, gambling party parameter described in the method includes: is used for describing described computer and plays
The gaming identity of family, hands quantity, hands board type, active/passive play a card state, board type faced by when passively playing a card, passively go out
The variable of the gaming identity of the player that plays a card faced by during board and value thereof；With make a call to one for describing described an innings two currently carried out
The board type that in game, all players have got, the board quantity got, the variable of overall situation residue hands quantity and value thereof.
According to a further aspect of the invention, strategy of playing a card described in the method at least includes for marking described calculating
The row board information of machine player's playing mode, described row board information includes play a card board type or mistake for describing described computer player
The variable of board operation and value thereof.
Provided by the present invention two make a call to one game in determine computer player play a card strategy method, utilize trained
BP neutral net calculates gambling party parameter in the first two plays a game and plays a card strategy determining computer player so that computer
The strategy of playing a card of player is adapted to the gambling party situation of reality, and then promotes play a card intelligence degree and the plan of computer player
True degree, enhances and carries out the game experiencing of the true player played with computer player.
Accompanying drawing explanation
By the detailed description that non-limiting example is made made with reference to the following drawings of reading, other of the present invention
Feature, purpose and advantage will become more apparent upon:
Fig. 1 be two according to the present invention make a call to one game in determine computer player play a card strategy method one be embodied as
The flow chart of mode；
Fig. 2 be two according to the present invention make a call to one game in determine computer player play a card strategy method one optional specifically
The flow chart of embodiment；
Fig. 3 is the detailed process schematic diagram of step S100 shown in Fig. 1 or Fig. 2；
In accompanying drawing, same or analogous reference represents same or analogous step.
Detailed description of the invention
For a better understanding and interpretation of the present invention, below in conjunction with accompanying drawing, the present invention is described in further detail.
The invention provides a kind of two make a call to one game in determine computer player play a card strategy method, refer to Fig. 1, figure
1 is that two according to the present invention make a call to and determine in a game that computer player is played a card the flow chart of method of strategy, the method include with
Lower step:
Step S100, creates BP neutral net；
Step S200, uses training data to train described BP neutral net；
Step S300, an innings two currently carried out make a call to one game in take turns to computer player play a card time, call described BP
The gambling party parameter that neutral net is made a call to described two in a game calculates, to determine the strategy of playing a card of described computer player.
Term " BP neutral net " refers to back propagation (Back Propagation) neutral net, people in the art
Member is appreciated that this BP neutral net is a kind of trainable Multi-layered Feedforward Networks, such as, use error backpropagation algorithm to enter
Row training, the BP neutral net after training can learn and store the mapping relations of substantial amounts of input-output pattern, and not
With the function of the mapping relations pre-set as finite state machine for describing input-output pattern, therefore BP nerve net
Network can be used for realizing artificial intelligence, especially sets up artificial intelligence's reaction for external environment.Especially, in the present invention should
BP neutral net is for determining two strategies of playing a card making a call to a game Computer player.
Specifically, refer to the detailed process schematic diagram that Fig. 3, Fig. 3 are steps S100 shown in Fig. 1 or Fig. 2, in step
The step creating BP neutral net in S100 includes:
Step S101, is respectively created input layer, hidden layer and the output layer of described BP neutral net；
Step S102, initializes described input layer and described hidden layer weights and side-play amount, and initializes described implicit
Layer and the weights of described output layer and side-play amount.
It will be understood by those skilled in the art that the described input layer in BP neutral net for receiving the information of input, and
Pass to described hidden layer.Described hidden layer is for processing and calculate the information of described input to generate object information, and will tie
Really information passes to described output layer, and described output layer is for outwardly transmitting described object information.Need according to information processing
Asking, described hidden layer can be designed as single layer structure or multiple structure.Correspondingly, it will be understood by those skilled in the art that BP to be made
Neutral net has normal learning capacity, needs to perform such as the operation in step S102, initializes described input layer with described
Hidden layer weights and side-play amount, and initialize described hidden layer and the weights of described output layer and side-play amount, namely regulation institute
Stating the initial value of weights and side-play amount, these weights and side-play amount can implement technical experience or the experiment knot of personnel according to the present invention
Really determine, it is also possible to selecting according to routine and determine, this is not limited by the present invention.
In view of the level of intelligence in order to promote computer player, need to use training data to train institute in step s 200
State BP neutral net.Owing to the BP neutral net of the present invention is for determining two plans of playing a card making a call to the computer player in a game
Slightly, the most described training data can generate according to the two gambling party information making a call to true player in the history gambling parties played.
Preferably, refer to Fig. 2, Fig. 2 be two according to the present invention make a call to one game in determine computer player play a card strategy method one
The flow chart of individual optional detailed description of the invention, compared with the technical scheme shown in Fig. 1, the technical scheme shown in Fig. 2 is in step
Can also comprise the steps: before S200
Step S400, pre-stored two makes a call to the gambling party information of true player in a history gambling party played；
Step S500, generates described training data according to described gambling party information.
Specifically, owing to described training data is to generate according to the playing a card action of true player, therefore use described
After BP neutral net is trained by training data, this BP neutral net can trend towards making the true object for appreciation of computer player simulation
The strategy of playing a card of family, to reach to promote the purpose of computer player level of intelligence.In order to record described true object for appreciation all sidedly as far as possible
Family's row board strategy in history gambling party, the most described training data includes the game body for describing described true player
Part, hands quantity, hands board type, active/passive play a card state, board type faced by when passively playing a card, faced by when passively playing a card
Play a card the gaming identity of player, the board type got, the board quantity got, the overall situation residue hands quantity, play a card board type, mistake
Arbitrary or the variable of a combination thereof and value thereof in board operation..The variable X of described gaming identity is such as described15Represent, wherein
X15=1 represents banker, X15=2 represent banker player whose turn comes next defender, X15=3 represent family defender on banker.Those skilled in the art
It is appreciated that in implementing the present invention, it may, variable and value thereof for describing above-mentioned gambling party relevant information can bases
The training demand of described BP neutral net freely selects, it is only necessary to meet the gambling party corresponding to specific value of described variable
Relevant information is uniquely.
Specifically, step S200 use training data train described BP neutral net to include: described BP neutral net makes
Calculate described training data with back-propagation algorithm, make actual output and the phase in described training data of described BP neutral net
Hope that the error of output is less than predetermined threshold；And/or described BP neutral net uses back-propagation algorithm to calculate described training data,
The correction number of times making the actual output of described BP neutral net reaches pre-determined number.Wherein said BP neutral net uses and reversely passes
Broadcast the algorithm described training data of calculating to refer to: described training data inputs described BP neutral net according to described weights with inclined
Shifting amount calculates, and records actual output, compares the error of this actual output and the desired output in described training data, root
According to weights and side-play amount described in described error update, the most again calculate described instruction according to the described weights after updating and side-play amount
Practice data and obtain actual output, the most described actual output and the error of described desired output.Circulation execution is described reversely
Propagation algorithm until described error reaches pre-determined number less than the correction number of times of predetermined threshold and/or described actual output, according to
The principle of back-propagation algorithm, described error can progressively reduce.The algorithm updating described weights and side-play amount can use under gradient
Fall algorithm.It should be noted that described desired output is typically to play a card in the face of identical according to player true in described training data
The information of playing a card of situation and generate.After described BP neutral net uses described training data to be trained, its actual output
Tend to and close to described desired output, namely play a card strategy trend the strategy of playing a card close to true player of computer player.
The gambling party information bigger in view of may including the order of magnitude in described training data, preferably step S200 is permissible
Use Distributed Calculation to realize, namely complete the training process to described BP neutral net on use multi-section processor.Use
The advantage of method execution step S200 of Distributed Calculation is can be with the computational efficiency of BP neutral net described in training for promotion.
One innings two make a call to Import computer player in a game, and perform step S300, i.e. make a call to one currently carrying out two
Game takes turns to computer player when playing a card, call the gambling party parameter that described BP neutral net makes a call to described two in a game and carry out
Calculate, to determine the strategy of playing a card of described computer player.The most described gambling party parameter includes: be used for describing described computer
The gaming identity of player, hands quantity, hands board type, active/passive are played a card state, board type faced by when passively playing a card, passive
The variable of the gaming identity of the player that plays a card faced by when playing a card and value thereof；With for describing described an innings currently carried out two dozens
The board type that in one game, all players have got, the board quantity got, the variable of overall situation residue hands quantity and value thereof,
Namely described gambling party parameter for reaction when the first two make a call to one play described in computer player play a card before faced by gambling party shape
Condition.Described strategy of playing a card at least includes the row board information for marking described computer player playing mode, and described row board is believed
Breath includes the board type of playing a card for describing described computer player or the variable crossing board operation and value thereof.Those skilled in the art
Should be understood that the above-mentioned board type of playing a card for describing described computer player or the variable of board operation excessively and value thereof are according to described BP
Neutral net calculates the actual output of gained according to described gambling party parameter and determines.
Although it should be noted that describe the operation of the inventive method in the accompanying drawings with particular order, but, this is not
Require or hint must perform these operations according to this particular order, could be real or have to carry out the most shown operation
Existing desired result.On the contrary, the step described in flow chart can change execution sequence.Additionally or alternatively, it is convenient to omit
Multiple steps are merged into a step and are performed by some step, and/or a step is decomposed into the execution of multiple step.
The present invention provide two make a call to one game in determine computer player play a card strategy method can use able to programme patrolling
Volume device realizes, it is also possible to be embodied as computer software, such as, can be a kind of calculating according to embodiments of the invention
Machine program product, runs this program product and makes computer perform for the method demonstrated.Described computer program includes
Computer-readable recording medium, this medium comprises computer program logic or code section, for realizing each of said method
Individual step.Described computer-readable recording medium can be mounted built-in medium in a computer or can be from calculating owner
The removable medium (such as hot-plugging technology storage device) of body dismounting.Described built-in medium includes but not limited to rewritable non-
Volatile memory, such as RAM, ROM, flash memory and hard disk.Described removable medium includes but not limited to: optical storage matchmaker
Body (such as CD-ROM and DVD), magneto-optic storage media (such as MO), magnetic recording medium (such as tape or portable hard drive), have
The media (such as storage card) of built-in rewritable nonvolatile memory and there are the media (such as ROM box) of built-in ROM.
It will be appreciated by those skilled in the art that any computer system with properly programmed device all will be able to carry out bag
All steps of the method for the present invention being contained in program product.Although the most detailed description of the invention all sides described in this specification
Overweight software program, but the alternate embodiment realizing the method for present invention offer as firmware and hardware equally will in the present invention
Ask within the scope of protection.
It is obvious to a person skilled in the art that the invention is not restricted to the details of above-mentioned one exemplary embodiment, Er Qie
In the case of the spirit or essential attributes of the present invention, it is possible to realize the present invention in other specific forms.Therefore, no matter
From the point of view of which point, all should regard embodiment as exemplary, and be nonrestrictive, the scope of the present invention is by appended power
Profit requires rather than described above limits, it is intended that all by fall in the implication of equivalency and scope of claim
Change is included in the present invention.Should not be considered as limiting involved claim by any reference in claim.This
Outward, it is clear that " including ", a word was not excluded for miscellaneous part, unit or step, and odd number is not excluded for plural number.
Provided by the present invention two make a call to one game in determine computer player play a card strategy method, utilize trained
BP neutral net calculates gambling party parameter in the first two plays a game and plays a card strategy determining computer player so that computer
The strategy of playing a card of player is adapted to the gambling party situation of reality, and then promotes play a card intelligence degree and the plan of computer player
True degree, enhances and carries out the game experiencing of the true player played with computer player.
Above disclosed be only the present invention some preferred embodiments, certainly can not with this limit the present invention it
Interest field, the equivalent variations therefore made according to the claims in the present invention, still belong to the scope that the present invention is contained.
Claims (8)
1. two make a call to one game in determine computer player play a card strategy a method, the method includes:
Create BP neutral net；
Training data is used to train described BP neutral net；
An innings two currently carried out make a call to one game in take turns to computer player play a card time, call described BP neutral net to described
The two gambling party parameters made a call in a game calculate, to determine the strategy of playing a card of described computer player.
Method the most according to claim 1, before using training data to train described BP neutral net, the method is also wrapped
Include:
Pre-stored two makes a call to the gambling party information of true player in a history gambling party played；
Described training data is generated according to described gambling party information.
Method the most according to claim 2, wherein, described training data includes:
For describing the gaming identity of described true player, hands quantity, hands board type, active/passive play a card state, passively go out
Board type faced by during board, the gaming identity of the player that plays a card faced by when passively playing a card, the board type got, the board number got
Amount, the overall situation remain hands quantity, board type of playing a card, cross the arbitrary or variable of a combination thereof and value thereof in board operation.
Method the most according to claim 1, wherein, creates BP neutral net and includes:
It is respectively created input layer, hidden layer and the output layer of described BP neutral net；
Initialize described input layer and described hidden layer weights and side-play amount, and initialize described hidden layer and described output layer
Weights and side-play amount.
Method the most according to claim 1, wherein, uses training data to train described BP neutral net to include:
Described BP neutral net uses back-propagation algorithm to calculate described training data, makes the actual output of described BP neutral net
It is less than predetermined threshold with the error of the desired output in described training data；And/or
Described BP neutral net uses back-propagation algorithm to calculate described training data, makes the actual output of described BP neutral net
Correction number of times reach pre-determined number.
6. according to the method described in any one of claim 1 to 5, wherein:
Training data is used to train the step of described BP neutral net to use Distributed Calculation to realize.
Method the most according to claim 1, wherein, described gambling party parameter includes:
For describing the gaming identity of described computer player, hands quantity, hands board type, active/passive play a card state, passive
Board type faced by when playing a card, the variable of gaming identity of the player that plays a card faced by when passively playing a card and value thereof；With
The board type that in a game, all players have got, the board number got is made a call to for describing described an innings two currently carried out
Amount, the variable of overall situation residue hands quantity and value thereof.
Method the most according to claim 1, wherein:
Described strategy of playing a card at least includes the row board information for marking described computer player playing mode, and described row board is believed
Breath includes the board type of playing a card for describing described computer player or the variable crossing board operation and value thereof.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610406925.5A CN106055339A (en) | 2016-06-08 | 2016-06-08 | Method for determining card playing strategy of computer player in two-against-one game |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610406925.5A CN106055339A (en) | 2016-06-08 | 2016-06-08 | Method for determining card playing strategy of computer player in two-against-one game |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN106055339A true CN106055339A (en) | 2016-10-26 |
Family
ID=57170734
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610406925.5A Pending CN106055339A (en) | 2016-06-08 | 2016-06-08 | Method for determining card playing strategy of computer player in two-against-one game |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN106055339A (en) |
Cited By (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108874377A (en) * | 2018-06-21 | 2018-11-23 | 深圳市腾讯网络信息技术有限公司 | A kind of data processing method, device and storage medium |
| CN108970124A (en) * | 2018-07-05 | 2018-12-11 | 腾讯科技（深圳）有限公司 | It forms a team matching process, device, system and readable medium |
| CN109033309A (en) * | 2018-07-17 | 2018-12-18 | 腾讯科技（深圳）有限公司 | Data processing method, device, equipment and storage medium |
| CN109091868A (en) * | 2018-08-14 | 2018-12-28 | 腾讯科技（深圳）有限公司 | Method, apparatus, computer equipment and the storage medium that battle behavior determines |
| CN109508789A (en) * | 2018-06-01 | 2019-03-22 | 北京信息科技大学 | Predict method, storage medium, processor and the equipment of hands |
| CN110175619A (en) * | 2018-11-27 | 2019-08-27 | 深圳市腾讯信息技术有限公司 | Method, equipment and storage medium are determined based on the board group of playing a card of machine learning model |
| CN110368690A (en) * | 2019-07-31 | 2019-10-25 | 腾讯科技（深圳）有限公司 | Gaming decision model training method, tactics of the game generation method and device |
| CN110852436A (en) * | 2019-10-18 | 2020-02-28 | 桂林力港网络科技股份有限公司 | Data processing method, device and storage medium for electronic poker game |
| CN111185010A (en) * | 2019-12-25 | 2020-05-22 | 北京理工大学 | System and method for constructing landlord card-playing program by using pulse neural network |
| CN111330282A (en) * | 2020-02-20 | 2020-06-26 | 北京金山云网络技术有限公司 | Method and device for determining card-playing candidate items |
| CN111729316A (en) * | 2020-06-15 | 2020-10-02 | 北京智明星通科技股份有限公司 | Card-out recommendation method and system in card battle game and game terminal |
| CN112446424A (en) * | 2020-11-16 | 2021-03-05 | 桂林力港网络科技股份有限公司 | Word card game data processing method, system and storage medium |
| CN113159313A (en) * | 2021-03-02 | 2021-07-23 | 北京达佳互联信息技术有限公司 | Data processing method and device of game model, electronic equipment and storage medium |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1845131A (en) * | 2005-04-06 | 2006-10-11 | 20Q网络公司 | Artificial neural network guessing method and game |
| KR20090092007A (en) * | 2008-02-26 | 2009-08-31 | 한국전자통신연구원 | Method and apparatus for studying user control pattern of game character in game |
| US20130281195A1 (en) * | 2006-10-24 | 2013-10-24 | Brain Games, L.C. | System and method for conducting a game including a computer-controlled player |
-
2016
- 2016-06-08 CN CN201610406925.5A patent/CN106055339A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1845131A (en) * | 2005-04-06 | 2006-10-11 | 20Q网络公司 | Artificial neural network guessing method and game |
| US20130281195A1 (en) * | 2006-10-24 | 2013-10-24 | Brain Games, L.C. | System and method for conducting a game including a computer-controlled player |
| KR20090092007A (en) * | 2008-02-26 | 2009-08-31 | 한국전자통신연구원 | Method and apparatus for studying user control pattern of game character in game |
Non-Patent Citations (5)
| Title |
|---|
| HHDDZ: "心血来潮，用神经网络写了一个单机版斗地主", 《HTTPS://BBS.CSDN.NET/TOPICS/350176582》 * |
| 张仁津等: "基于人工神经网络游戏程序的研究和设计", 《广西师范大学学报(自然科学版）》 * |
| 戴红等编著: "《数据挖掘导论》", 31 January 2015, 清华大学出版社 * |
| 王永琦编著: "《MATLAB与音视频技术》", 30 November 2013, 清华大学出版社 * |
| 谢剑斌等编著: "《视觉机器学习20讲》", 30 June 2015, 清华大学出版社 * |
Cited By (19)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109508789A (en) * | 2018-06-01 | 2019-03-22 | 北京信息科技大学 | Predict method, storage medium, processor and the equipment of hands |
| CN109508789B (en) * | 2018-06-01 | 2022-03-15 | 北京信息科技大学 | Method, storage medium, processor and apparatus for predicting hand |
| CN108874377A (en) * | 2018-06-21 | 2018-11-23 | 深圳市腾讯网络信息技术有限公司 | A kind of data processing method, device and storage medium |
| CN108970124A (en) * | 2018-07-05 | 2018-12-11 | 腾讯科技（深圳）有限公司 | It forms a team matching process, device, system and readable medium |
| CN109033309A (en) * | 2018-07-17 | 2018-12-18 | 腾讯科技（深圳）有限公司 | Data processing method, device, equipment and storage medium |
| CN109033309B (en) * | 2018-07-17 | 2023-04-07 | 腾讯科技（深圳）有限公司 | Data processing method, device, equipment and storage medium |
| CN109091868A (en) * | 2018-08-14 | 2018-12-28 | 腾讯科技（深圳）有限公司 | Method, apparatus, computer equipment and the storage medium that battle behavior determines |
| CN110175619A (en) * | 2018-11-27 | 2019-08-27 | 深圳市腾讯信息技术有限公司 | Method, equipment and storage medium are determined based on the board group of playing a card of machine learning model |
| CN110175619B (en) * | 2018-11-27 | 2023-02-03 | 深圳市腾讯信息技术有限公司 | Method, device and storage medium for determining card-playing set based on machine learning model |
| CN110368690B (en) * | 2019-07-31 | 2021-10-22 | 腾讯科技（深圳）有限公司 | Game decision model training method, game strategy generation method and device |
| CN110368690A (en) * | 2019-07-31 | 2019-10-25 | 腾讯科技（深圳）有限公司 | Gaming decision model training method, tactics of the game generation method and device |
| CN110852436A (en) * | 2019-10-18 | 2020-02-28 | 桂林力港网络科技股份有限公司 | Data processing method, device and storage medium for electronic poker game |
| CN111185010A (en) * | 2019-12-25 | 2020-05-22 | 北京理工大学 | System and method for constructing landlord card-playing program by using pulse neural network |
| CN111330282A (en) * | 2020-02-20 | 2020-06-26 | 北京金山云网络技术有限公司 | Method and device for determining card-playing candidate items |
| CN111729316A (en) * | 2020-06-15 | 2020-10-02 | 北京智明星通科技股份有限公司 | Card-out recommendation method and system in card battle game and game terminal |
| CN112446424A (en) * | 2020-11-16 | 2021-03-05 | 桂林力港网络科技股份有限公司 | Word card game data processing method, system and storage medium |
| CN112446424B (en) * | 2020-11-16 | 2024-02-27 | 桂林力港网络科技股份有限公司 | Word card game data processing method, system and storage medium |
| CN113159313A (en) * | 2021-03-02 | 2021-07-23 | 北京达佳互联信息技术有限公司 | Data processing method and device of game model, electronic equipment and storage medium |
| CN113159313B (en) * | 2021-03-02 | 2022-09-09 | 北京达佳互联信息技术有限公司 | Data processing method and device of game model, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN106055339A (en) | Method for determining card playing strategy of computer player in two-against-one game | |
| Rosyid et al. | Deploying learning materials to game content for serious education game development: A case study | |
| US7636701B2 (en) | Query controlled behavior models as components of intelligent agents | |
| WO2022247791A1 (en) | Chess self-learning method and apparatus based on machine learning | |
| Lopes et al. | Authoring adaptive game world generation | |
| Bontchev | Modern trends in the automatic generation of content for video games | |
| Smith | Procedural content generation: An overview | |
| Al-Khateeb et al. | Introducing individual and social learning into evolutionary checkers | |
| Zook et al. | Temporal game challenge tailoring | |
| Baykal et al. | Reinforcement learning in card game environments using monte carlo methods and artificial neural networks | |
| US11797274B2 (en) | Interoperable composite data units for use in distributed computing execution environments | |
| Langi et al. | Generation quiz with genetic algorithm based on bloom's taxonomy classification in serious game based virtual environments | |
| Mozgovoy et al. | Behavior capture: Building believable and effective AI agents for video games | |
| Végh et al. | Simulations of Solving a Single-Player Memory Card Game with Several Implementations of a Human-Like Thinking Computer Algorithm | |
| Sriram | Automated playtesting of platformer games using reinforcement learning | |
| Drabik | An analysis of the connection between ai development and the creation of games and advanced computer systems | |
| Upadhyay | Analysis of q-learning based game playing agents for abstract board games with increasing state-space complexity | |
| Akinyemi et al. | Critical analysis of decision making experience with a machine learning approach in playing Ayo game | |
| Gong | Dota 2 Hero Selection Analysis | |
| Yen et al. | Monte Carlo Tree Search and Machine Learning Techniques on Block Go Programs | |
| Gunawan et al. | Development of a Scrum Learning Model Using Gamification with the Battleship Board Game | |
| Phon-Amnuaisuk | Learning to play tetris from examples | |
| Beaudoin | Serious lessons from the commercial games industry | |
| Scott | Complementary companion behavior in video games | |
| CHAN | THE UNIVERSITY OF HONG KONG FACULTY OF ENGINEERING DEPARTMENT OF COMPUTER SCIENCE FINAL YEAR PROJECT |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20161026 |
|
| RJ01 | Rejection of invention patent application after publication |