Game territory construction method, computer equipment and storage medium

Abstract

The embodiment of the invention discloses a game territory construction method, computer equipment and a storage medium. The method comprises the steps that environment image information, building feature information and camp state information in a game environment map are determined; determining a path planning end point according to the environment image information, the building feature information and the camp state information; performing path planning based on the path planning terminal point and the own current territory to obtain a target building point; and building a game territory based on the target building point. According to the method, the game territory construction is divided into two processes of determining the path planning end point and determining the target building point for hierarchical decision making, so that the decision complexity is reduced, a feasible game territory construction strategy is quickly generated when a high-complexity scene is dealt with, game response delay caused by long-time calculation is avoided, and the user experience is improved. And smoother and timely game experience is provided for players. Therefore, the problem that game decision-making difficulty is high in a high-complexity scene is solved.

Description

Technical Field

[0001] The present application relates to the field of machine learning, and in particular to a method for building a game territory, a computer device, and a storage medium. Background Art

[0002] In strategy games, players must choose building locations on massive maps to expand their territory, compete for resources, and control their opponents. While current reinforcement learning-based game AI methods can learn basic strategies through trial and error, they must process massive amounts of observational information in highly complex scenarios, significantly increasing the difficulty of decision-making and making optimal decisions quickly, which in turn impacts player performance and development. Summary of the Invention

[0003] This application provides a game territory construction method, computer equipment and storage medium, which can solve the problem of difficult game decision-making in highly complex scenarios.

[0004] In a first aspect, the present application provides a method for building a game territory, the method comprising:

[0005] Determine the environmental image information, building feature information and camp status information in the game environment map;

[0006] Determining a path planning endpoint based on the environmental image information, the building feature information, and the camp status information;

[0007] Performing path planning based on the path planning endpoint and the current territory of the party to obtain a target building point;

[0008] The game territory is constructed based on the target building point.

[0009] In a second aspect, the present application further provides a computer device, the computer device comprising a memory and a processor;

[0010] The memory is used to store computer programs;

[0011] The processor is used to execute the computer program and implement the above-mentioned game territory construction method when executing the computer program.

[0012] In a third aspect, the present application further provides a computer-readable storage medium storing a computer program, which, when executed by a processor, enables the processor to implement the game territory construction method as described above.

[0013] This application discloses a method, computer device, and storage medium for building game territory. The method determines environmental image information, building feature information, and faction status information from a game environment map; determines a path planning endpoint based on the environmental image information, building feature information, and faction status information; performs path planning based on the path planning endpoint and the player's current territory to obtain a target building point; and then constructs the game territory based on the target building point. This solution reduces decision-making complexity by breaking down game territory construction into two steps: determining the path planning endpoint and determining the target building point, and performing hierarchical decision-making. This approach addresses the difficulty of game decision-making in highly complex scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0015] Figure 1 This is a flowchart of a method for building a game territory provided by an embodiment of the present application;

[0016] Figure 2 is a schematic diagram of a game environment map provided in an embodiment of the present application;

[0017] Figure 3 This is a schematic diagram of the process of implementing the game territory construction provided by the embodiment of the present application;

[0018] Figure 4 It is a structural diagram of the computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0019] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0020] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, combined, or partially merged, so the actual execution order may vary depending on the actual situation.

[0021] It should be understood that the terms used in this specification are only for the purpose of describing specific embodiments and are not intended to limit the present application. As used in this specification and the appended claims, the singular forms "a", "an", and "the" are intended to include the plural forms unless the context clearly indicates otherwise.

[0022] It will also be understood that the term "and / or" used in this specification and the appended claims refers to and includes any and all possible combinations of one or more of the associated listed items.

[0023] In large-scale strategy game scenarios, the map size is usually more than 1000*1000 units. Players can expand their territory by building buildings on the map, and can also interact with buildings connected to their own territory, such as capturing neutral buildings, occupying resource points, attacking enemy buildings, etc. At the same time, players can also hinder the expansion of enemy territory by expanding their own territory, thereby preventing the enemy from capturing important targets.

[0024] Existing reinforcement learning methods, such as the widely used Proximal Policy Optimization (PPO) algorithm, are a policy gradient-based reinforcement learning algorithm. Its core idea is to continuously adjust policy parameters so that the actions generated by the policy can achieve higher cumulative rewards. This algorithm has achieved remarkable results in the field of game AI. However, in highly complex scenarios, the algorithm needs to process a massive amount of observational information when making decisions. This includes resource distribution, including the reserves and regeneration rates of various resources at different locations on the map; combat power comparison information, including the attack power, defense power, health points, and other attributes of different friendly and enemy units and buildings; and the restraint relationships between different units. This information is not only massive in quantity but also interconnected and dynamically changing, making decision-making significantly more difficult.

[0025] In view of this, the present invention aims to propose a game territory construction method that can solve the problem of difficult game decision-making in highly complex scenarios.

[0026] The following describes some embodiments of the present application in detail with reference to the accompanying drawings. In the absence of conflict, the following embodiments and features therein may be combined with each other.

[0027] See also Figure 1 , Figure 1 : is a flowchart of a method for building a game territory provided by an embodiment of the present application. The method for building a game territory may include steps S101 to S104, as follows:

[0028] Step S101: Determine environmental image information, building feature information, and camp status information in a game environment map.

[0029] Among them, the game environment map can be flexibly set according to the actual application scenario, and the specific content is not limited here. Figure 2 As shown, the game environment map uses a green background to simulate natural terrain, with brown defensive buildings scattered across it to defend against enemy attacks. Various colored dots are also scattered across the map, representing different resource mines, such as timber, quarries, and farmland. This map setting allows players to intuitively understand the layout of the map and the location of various elements, allowing them to better construct their territory.

[0030] In some embodiments, determining environmental image information, building feature information, and camp status information in a game environment map includes: intercepting initial environmental images of different scales with the overall centroid position of multiple buildings in the game environment map as the center, performing dimensionality reduction processing on the initial environmental images, and obtaining environmental image information; obtaining multiple candidate buildings in the game environment map, and determining building feature information based on the relative coordinates between each candidate building and the overall centroid position; and performing statistics on the status of one's own camp in the game environment map to obtain camp status information.

[0031] Exemplarily, based on the coordinate data of multiple buildings in the game environment map, the overall center of mass position of the multiple buildings is calculated. Taking the overall center of mass position as the center, three initial environment images of different scales of 500*500 units, 300*300 units, and 100*100 units are respectively intercepted, and each initial environment image is subjected to dimensionality reduction processing. For example, the maximum pooling operation can be used to reduce the dimensionality of the above three initial environment images of different scales to obtain standardized 40*40 unit environment image information, and the representation compression of key terrain elements is achieved by retaining the local maximum features.

[0032] It's important to note that the overall center of mass can be determined by averaging the coordinate data of multiple buildings. For example, consider the coordinate data for multiple buildings. These coordinate points are stored in an array, with each coordinate point consisting of a horizontal and vertical coordinate. Next, using a numerical computation library such as NumPy, this array is processed to calculate the average of all coordinate points on the horizontal and vertical axes. Specifically, the horizontal coordinates are averaged to obtain the horizontal coordinate of the center of mass, and the vertical coordinates are averaged to obtain the vertical coordinate of the center of mass, thereby accurately determining the overall center of mass.

[0033] For example, the 12 candidate buildings closest to the overall centroid position in the game environment map are counted, and the candidate buildings include 5 neutral buildings that have not been occupied by any camp, 3 buildings belonging to the enemy camp, and 4 resource mines rich in various resources. The relative coordinates between each candidate building and the overall centroid position are recorded to obtain the building feature information.

[0034] For example, the status of one's own camp includes the amount of one's own warehouse resources, combat capability, etc. Through detailed statistics of these statuses, the camp status information is obtained, and the camp status information can represent one's own combat capability and construction capability.

[0035] Step S102: Determine the destination of the path planning based on the environmental image information, the building feature information, and the camp status information.

[0036] The destination of the path planning is one of the unoccupied neutral buildings, enemy buildings, and resource mines.

[0037] In some embodiments, determining the path planning endpoint based on environmental image information, building feature information, and camp status information may include: inputting the environmental image information, building feature information, and camp status information into a trained target prediction model to predict the path planning endpoint, obtaining at least one candidate path planning endpoint and a confidence level corresponding to each candidate path planning endpoint; determining the path planning endpoint based on each candidate path planning endpoint and the confidence level corresponding to each candidate path planning endpoint.

[0038] In order to improve the accuracy of determining the path planning endpoint, after obtaining environmental image information, building feature information and camp status information, the path planning endpoint can be predicted through the trained target prediction model to obtain at least one candidate path planning endpoint and the confidence corresponding to each candidate path planning endpoint. Among them, the trained target prediction model can be flexibly set according to actual needs, and the specific content is not limited here. After obtaining at least one candidate path planning endpoint and the confidence corresponding to each candidate path planning endpoint, the target prediction model calculates the confidence corresponding to each candidate path planning endpoint by using the softmax function to obtain a probability distribution, and selects the candidate path planning endpoint with the highest probability as the path planning endpoint to be built according to the probability distribution result.

[0039] In some embodiments, before inputting environmental image information, building feature information, and camp status information into a preset target prediction model to predict the path planning endpoint and obtaining at least one candidate path planning endpoint and the confidence level corresponding to each candidate path planning endpoint, it also includes: randomly extracting interaction data from a data cache pool; and training the target prediction model to be trained according to the interaction data until convergence, until a trained target prediction model is obtained.

[0040] In order to improve the reliability of model training, interaction data can be randomly extracted from the data cache pool. The interaction data can include multiple data samples. Each data sample can be the data of a game interaction with the opponent. The game data covers all information in the game environment map, such as terrain, personnel distribution, building distribution, and the win or loss of the game.

[0041] In some embodiments, before randomly extracting interaction data from the data cache pool, it also includes: extracting an opponent from the opponent strategy pool according to a preset probability; interacting with the opponent in the game environment map to obtain interaction data; and storing the interaction data in the data cache pool.

[0042] The preset probability is not a fixed value and will be dynamically adjusted based on the opponent's strength. Specifically, for weak opponents, their extraction probability is reduced. Since it is difficult to obtain effective strategic feedback when playing against opponents with a large gap in strength, it is an ineffective battle and will waste game resources and time. By reducing the extraction probability of weak opponents, the game process can be made more targeted and resources can be concentrated on more valuable battles. For strong opponents, their extraction probability is increased. Fighting against strong opponents can force the game model to encounter more complex and challenging strategic scenarios, thereby accelerating the optimization process of the model strategy. For opponents with equal strength, the normal extraction probability is maintained. This setting can maintain a stable and challenging game environment, allowing the game model to continuously learn and improve in competition and balance, thereby ensuring the continuous improvement of the quality of interactive data.

[0043] In some embodiments, drawing an opponent from an opponent strategy pool according to a preset probability includes: obtaining the ratio of one's own score when playing against the opponent to the opponent's own score; obtaining the sum of the ratios of one's own score when playing against all opponents in the opponent strategy pool to the opponent's own score; and obtaining the probability of drawing the opponent based on the ratio of the ratio of one's own score when playing against the opponent to the opponent's own score and the sum of the ratios of one's own score to the opponent's own score when playing against all opponents.

[0044] It is expressed as a mathematical formula Among them, P a is the probability of extracting opponent α, W α is the ratio of one's own score to the opponent's own score when playing against the opponent α, W i is the ratio of the own score to the score of the i-th opponent when playing against the i-th opponent. The value of i ranges from 0 to n, and n is the total number of opponents in the opponent strategy pool. For example, the probability P of extracting opponent α is calculated. a =0.7, then the opponent α has a 70% probability of being selected to participate in the interaction of this game.

[0045] For example, a running device can run multiple game games. For example, if the running device is an 8-core device and each CPU core can run 3 game games, then the running device supports 24 game games running simultaneously. These simultaneously running games will continuously generate interaction data, which is stored in the data buffer pool. The target prediction model to be trained randomly extracts the interaction data from it for training.

[0046] Step S103: Path planning is performed based on the path planning endpoint and the current territory of the party to obtain the target building point.

[0047] For example, in the game environment map, the current territory of one's side contains 5 buildings, each building is defined as a square area and has 4 vertices, that is, the 5 buildings generate a total of 20 vertices. From these 20 vertices, the 10 vertices closest to the path planning end point are selected as candidate path planning starting points. The 10 candidate path planning starting points are normalized by the softmax function to generate a probability distribution. According to the probability distribution, one candidate path planning starting point is randomly selected as the path planning starting point. Combined with the path planning end point, path planning is performed to finally obtain the target building point.

[0048] In some embodiments, performing path planning based on the path planning endpoint and one's own current territory to obtain a target building point may include: performing path planning using an Astar algorithm based on the path planning endpoint and the current territory to generate a target path; delineating a circular area of ​​a preset radius with the intersection of the target path and a preset construction range boundary as the center to obtain a target area; determining points within the target area as candidate building points; and determining the target building point in the target area based on the sphere of influence bordering degree, resource coverage, new territory expansion degree, and enemy threat degree corresponding to at least one candidate building point in the target area.

[0049] The preset radius can be flexibly set according to the actual application scenario, and the specific content is not limited here, ensuring that the area can fully cover the possible distribution range of candidate building points, providing a reasonable spatial basis for the subsequent accurate screening and evaluation of candidate building points.

[0050] In some embodiments, determining the target building point in the target area based on the influence sphere bordering degree, resource coverage degree, new territory expansion degree, and enemy threat degree corresponding to at least one candidate building point in the target area may include: performing weight calculation based on the influence sphere bordering degree, resource coverage degree, new territory expansion degree, and enemy threat degree corresponding to each candidate building point to obtain a comprehensive score corresponding to each candidate building point; and determining the target building point in the target area based on the comprehensive score corresponding to each candidate building point.

[0051] In order to improve the accuracy of evaluating candidate building points, each candidate building point is weighted based on four aspects: sphere of influence bordering, resource coverage, new territory expansion, and enemy threat. The comprehensive score corresponding to each candidate building point is obtained. Based on the comprehensive score corresponding to each candidate building point, the candidate building point with the highest comprehensive score is selected and determined as the target building point in the target area.

[0052] Step S104: construct the game territory based on the target building point.

[0053] After obtaining the target building site, you can construct your game territory there. If the target building site is in an unoccupied neutral building area, you can build fortifications such as city walls and arrow towers to resist possible enemy attacks. If the target building site is in an enemy building area, you can build barracks such as infantry and cavalry battalions for combat. If the target building site is in a resource mining area, you can build resource gathering buildings to quickly accumulate resources and lay a solid foundation for subsequent development.

[0054] See also Figure 3 , Figure 3 The embodiment of the present application also provides a schematic diagram of the process of implementing the construction of a game territory, such as Figure 3 As shown, the running device synchronously starts multiple game processes to interact with opponents in the opponent strategy pool, solves task allocation and resource scheduling by replacing the flow shop scheduling problem, and collects interaction data and stores it in the data cache pool; randomly extracts interaction data from the data cache pool to train the target prediction model until the model converges, and obtains the trained target prediction model. The trained target prediction model inputs the environmental image information, building feature information and camp status information in the game environment map, outputs the selected path planning end point, performs path planning based on the path planning end point and the current territory of the party, obtains the target building point, and constructs the game territory based on the target building point.

[0055] The game territory construction method provided in embodiments of the present application determines environmental image information, building feature information, and faction status information from a game environment map; determines a path planning endpoint based on this information; performs path planning based on the path planning endpoint and the player's current territory to obtain a target building point; and then constructs the game territory based on the target building point. This method reduces decision-making complexity by breaking down game territory construction into two steps: determining the path planning endpoint and determining the target building point, and performing hierarchical decision-making. This addresses the difficulty of game decision-making in highly complex scenarios.

[0056] See also Figure 4 , Figure 4 This is a schematic block diagram of the structure of a computer device provided in an embodiment of the present application.

[0057] See also Figure 4 The computer device includes a processor, a memory, and a network interface connected via a system bus, wherein the memory may include a storage medium and an internal memory. The storage medium may be a non-volatile storage medium or a volatile storage medium.

[0058] The processor is used to provide computing and control capabilities and support the operation of the entire computer equipment.

[0059] The internal memory provides an environment for the operation of the computer program in the storage medium. When the computer program is executed by the processor, the processor can execute any one of the methods for generating video description text.

[0060] The network interface is used for network communication, such as sending assigned tasks, etc.

[0061] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0062] It should be understood that the processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc.

[0063] In some embodiments, the processor is configured to execute a computer program stored in the memory to implement the following steps:

[0064] Determine the environmental image information, building feature information and camp status information in the game environment map; determine the path planning end point based on the environmental image information, building feature information and camp status information; perform path planning based on the path planning end point and the current territory of the party to obtain the target building point; and construct the game territory based on the target building point.

[0065] In some embodiments, when determining the environmental image information, building feature information, and camp status information in the game environment map, the processor is configured to implement:

[0066] Taking the overall centroid position of multiple buildings in the game environment map as the center, initial environment images of different scales are intercepted, and the initial environment images are subjected to dimensionality reduction processing to obtain environment image information; multiple candidate buildings in the game environment map are obtained, and the building feature information is determined based on the relative coordinates between each candidate building and the overall centroid position; the status of one's own camp in the game environment map is counted to obtain camp status information.

[0067] In some embodiments, when determining the path planning endpoint based on the environmental image information, the building feature information, and the camp status information, the processor is configured to implement:

[0068] Environmental image information, building feature information, and camp status information are input into the trained target prediction model to predict the path planning endpoint, and at least one candidate path planning endpoint and the confidence level corresponding to each candidate path planning endpoint are obtained; the path planning endpoint is determined based on each candidate path planning endpoint and the confidence level corresponding to each candidate path planning endpoint.

[0069] In some embodiments, before inputting environmental image information, building feature information, and camp status information into a preset target prediction model to predict a path planning endpoint and obtaining at least one candidate path planning endpoint and a confidence score corresponding to each candidate path planning endpoint, the processor is further configured to:

[0070] Interaction data is randomly extracted from the data cache pool; based on the interaction data, the target prediction model to be trained is trained until convergence, until the trained target prediction model is obtained.

[0071] In some embodiments, before randomly extracting the interaction data from the data cache pool, the processor is further configured to implement:

[0072] An opponent is extracted from an opponent strategy pool according to a preset probability; interaction is performed with the opponent in a game environment map to obtain interaction data; and the interaction data is stored in the data cache pool.

[0073] In some embodiments, when extracting an opponent from an opponent strategy pool according to a preset probability, the processor is configured to implement:

[0074] Obtain the ratio of one's own score to the opponent's own score when playing against the opponent; obtain the sum of the ratios of one's own score to the opponent's own score when playing against all opponents in the opponent strategy pool; obtain the probability of drawing an opponent based on the ratio of the ratio of one's own score to the opponent's own score when playing against the opponent and the sum of the ratios of one's own score to the opponent's own score when playing against all opponents.

[0075] In some embodiments, when performing path planning based on the path planning endpoint and the current territory of the own party to obtain the target building point, the processor is configured to implement:

[0076] The Astar algorithm is used to plan the path based on the planned endpoint and the current territory to generate a target path. A circular area with a preset radius is delineated with the intersection of the target path and the preset construction range boundary as the center to obtain the target area. Points within the target area are identified as candidate building points. The target building point in the target area is determined based on the sphere of influence bordering degree, resource coverage, new territory expansion degree, and enemy threat level corresponding to at least one candidate building point in the target area.

[0077] In some embodiments, when determining a target building point in a target area based on the influence sphere bordering degree, resource coverage degree, new territory expansion degree, and enemy threat degree corresponding to at least one candidate building point in the target area, the processor is configured to:

[0078] A weighted calculation is performed based on the degree of influence bordering, resource coverage, new territory expansion, and enemy threat corresponding to each candidate building point to obtain a comprehensive score for each candidate building point; based on the comprehensive score corresponding to each candidate building point, the target building point in the target area is determined.

[0079] In the above embodiments, the description of each embodiment has its own focus. For the part that is not described in detail in a certain embodiment, please refer to the detailed description of the game territory construction method above, which will not be repeated here.

[0080] In this embodiment of the application, a computer device determines environmental image information, building feature information, and faction status information from a game environment map; determines a path planning endpoint based on this information; performs path planning based on the path planning endpoint and the team's current territory to obtain a target building point; and then constructs the game territory based on the target building point. This solution reduces decision-making complexity by breaking down the construction of the game territory into two steps: determining the path planning endpoint and determining the target building point, and performing hierarchical decision-making. This solves the problem of difficult game decision-making in highly complex scenarios.

[0081] The embodiments of the present application also provide a computer-readable storage medium that stores a computer program. The computer program includes program instructions, and a processor executes the program instructions to implement any of the game territory construction methods provided in the embodiments of the present application. For example, when the computer program is loaded by the processor, the following steps may be executed:

[0082] Determine the environmental image information, building feature information and camp status information in the game environment map; determine the path planning end point based on the environmental image information, building feature information and camp status information; perform path planning based on the path planning end point and the current territory of the party to obtain the target building point; and construct the game territory based on the target building point.

[0083] The specific implementation of the above operations can be found in the previous embodiments and will not be repeated here.

[0084] The computer-readable storage medium may be an internal storage unit of the computer device described in the aforementioned embodiment, such as a hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a SmartMedia Card (SMC), a Secure Digital Card (SD Card), a flash memory card, etc. equipped on the computer device.

[0085] Furthermore, the computer-readable storage medium may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function, etc.; the data storage area may store data created according to the use of the blockchain node, etc.

[0086] The blockchain referred to in this application is a new application model for computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Blockchain is essentially a decentralized database, a series of data blocks generated using cryptographic methods. Each data block contains information about a batch of network transactions, which is used to verify the validity of this information (to prevent counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, the platform product service layer, and the application service layer.

[0087] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present application, and such modifications or substitutions should be included in the scope of protection of the present application. Therefore, the scope of protection of the present application should be based on the scope of protection of the claims.

Claims

1. A method for building a game territory, characterized in that: include: Determine the environmental image information, building feature information and camp status information in the game environment map; Determining a path planning endpoint based on the environmental image information, the building feature information, and the camp status information; Performing path planning based on the path planning endpoint and the current territory of the party to obtain a target building point; The game territory is constructed based on the target building point.

2. The game territory construction method according to claim 1, characterized in that: The determining of the environment image information, building feature information, and camp status information in the game environment map includes: Taking the overall centroid position of multiple buildings in the game environment map as the center, intercepting initial environment images of different scales, and performing dimensionality reduction processing on the initial environment images to obtain the environment image information; Acquire multiple candidate buildings in the game environment map, and determine the building feature information based on the relative coordinates between each candidate building and the overall centroid position; The status of one's own camp in the game environment map is counted to obtain the camp status information.

3. The game territory construction method according to claim 1, characterized in that: The determining of a path planning endpoint according to the environmental image information, the building feature information, and the camp status information includes: Inputting the environmental image information, the building feature information, and the camp status information into the trained target prediction model to predict the path planning endpoint, and obtaining at least one candidate path planning endpoint and a confidence score corresponding to each candidate path planning endpoint; The path planning endpoint is determined according to each candidate path planning endpoint and the confidence level corresponding to each candidate path planning endpoint.

4. The game territory construction method according to claim 3, characterized in that: Before inputting the environmental image information, the building feature information, and the camp status information into the trained target prediction model to predict the path planning endpoint and obtaining at least one candidate path planning endpoint and the confidence level corresponding to each candidate path planning endpoint, the method further includes: Randomly extract interaction data from the data cache pool; The target prediction model to be trained is trained according to the interaction data until convergence, until the trained target prediction model is obtained.

5. The game territory construction method according to claim 4, characterized in that: Before randomly extracting the interaction data from the data cache pool, the method further includes: Select opponents from the opponent strategy pool according to the preset probability; interacting with the opponent in the game environment map to obtain interaction data; The interaction data is stored in the data cache pool.

6. The method for building a game territory according to claim 5, characterized in that: The step of extracting an opponent from the opponent strategy pool according to a preset probability includes: Get the ratio of your own score to your opponent's score when playing against him; Get the sum of the ratios of your own score to the opponent's own score in all opponents' strategy pools; The probability of drawing the opponent is obtained according to the ratio of the ratio of the own score to the opponent's own score when playing against the opponent and the sum of the ratios of the own score to the opponent's own score when playing against all opponents.

7. The game territory construction method according to claim 1, characterized in that: The path planning is performed based on the path planning endpoint and the current territory of the party to obtain the target building point, including: Performing path planning using an Astar algorithm based on the path planning endpoint and the current territory to generate a target path; Delineating a circular area with a preset radius with the intersection of the target path and the preset construction range boundary as the center to obtain the target area; Determine a point located within the target area as a candidate building point; A target building point in the target area is determined based on the sphere of influence bordering degree, resource coverage degree, new territory expansion degree and enemy threat degree corresponding to at least one candidate building point in the target area.

8. The game territory construction method according to claim 7, characterized in that: Determining a target building point in the target area according to the sphere of influence bordering degree, resource coverage degree, new territory expansion degree, and enemy threat degree corresponding to at least one candidate building point in the target area includes: A weighted calculation is performed based on the sphere of influence bordering degree, resource coverage, new territory expansion degree, and enemy threat degree corresponding to each candidate building point to obtain a comprehensive score corresponding to each candidate building point; A target building point in the target area is determined according to the comprehensive score corresponding to each candidate building point.

9. A computer device, characterized in that: The computer device includes a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein when the computer program is executed by the processor, the steps of the game territory construction method according to any one of claims 1 to 8 are implemented.

10. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the steps of the game territory construction method according to any one of claims 1 to 8 are implemented.

Images