Test and evaluation method and apparatus for virtual character, and server and storage medium

Abstract

A test and evaluation method for a virtual character, the method comprising: performing a battle capability test on a plurality of non-player virtual characters, so as to obtain initial evaluation results for the non-player virtual characters; on the basis of the initial evaluation results for the non-player virtual characters and preset evaluation results for player virtual characters, performing a battle capability test on the non-player virtual characters and the player virtual characters, so as to obtain a first battle result; and on the basis of the first battle result, updating the initial evaluation results and the preset evaluation results, so as to obtain target evaluation results for the non-player virtual characters and target evaluation results for the player virtual characters. A battle capability test performed on non-player virtual characters and a battle capability test performed on the non-player virtual characters and player virtual characters are performed by means of offline evaluation and online evaluation, such that skill levels of the non-player virtual characters and the player virtual characters can be accurately and objectively evaluated.

Description

Test evaluation method and device for virtual role, server and storage medium

[0001] Cross-reference to Related Applications

[0002] The present application claims priority to the Chinese patent application No. 202410989423.4, filed on July 23, 2024, and entitled "Test evaluation method and device for virtual role, server and storage medium", the content of which is incorporated herein by reference in its entirety. TECHNICAL FIELD

[0003] The present disclosure relates to the technical field of games, in particular to a test evaluation method and device for virtual role, a server and a storage medium. BACKGROUND

[0004] The skill level of players in online games is diverse. It is not a suitable strategy to deploy the same difficulty robot (non-player virtual role) for all players. In order to obtain better robot deployment effect, it is necessary to match the difficulty according to the level of the players, and it is necessary to clearly know the skill level of the robot. The evaluation of the skill level of the robot has become a research hotspot.

[0005] In the related art, a large number of manual tests are used to evaluate the ability level of the robot through the sense of the test personnel. However, there is a problem that the evaluation result is relatively subjective and inaccurate. SUMMARY

[0006] According to an aspect of the present disclosure, a test evaluation method for a virtual role is provided. The method includes: performing a battle ability test on a plurality of non-player virtual roles to obtain an initial evaluation result of the non-player virtual roles; performing a battle ability test on the non-player virtual roles and a player virtual role according to the initial evaluation result of the non-player virtual roles and a preset evaluation result of the player virtual role to obtain a first battle result; and updating the initial evaluation result and the preset evaluation result according to the first battle result to obtain a target evaluation result of the non-player virtual roles and a target evaluation result of the player virtual role.

[0007] According to one aspect of the present disclosure, a virtual role test evaluation device is provided, which comprises: a test module configured to perform a battle capability test on a plurality of non-player virtual roles to obtain initial evaluation results of the non-player virtual roles; and perform a battle capability test on the non-player virtual roles and a player virtual role according to the initial evaluation results of the non-player virtual roles and preset evaluation results of the player virtual role to obtain a first battle result; and an update module configured to perform an update on the initial evaluation results and the preset evaluation results according to the first battle result to obtain target evaluation results of the non-player virtual roles and target evaluation results of the player virtual role.

[0008] According to one aspect of the present disclosure, a server is provided, which comprises: a memory and a processor, the memory storing a computer program executable by the processor, and the processor implementing the virtual role test evaluation method of any one of the above when executing the computer program.

[0009] According to one aspect of the present disclosure, a computer readable storage medium is provided, which stores a computer program, and the computer program is read and executed to implement the virtual role test evaluation method of any one of the above.

[0010] The present disclosure has the following beneficial effects: the present disclosure provides a virtual role test evaluation method, which comprises: performing a battle capability test on a plurality of non-player virtual roles to obtain initial evaluation results of the non-player virtual roles; performing a battle capability test on the non-player virtual roles and a player virtual role according to the initial evaluation results of the non-player virtual roles and preset evaluation results of the player virtual role to obtain a first battle result; and performing an update on the initial evaluation results and the preset evaluation results according to the first battle result to obtain target evaluation results of the non-player virtual roles and target evaluation results of the player virtual role. The battle capability test is first performed on a plurality of non-player virtual roles to realize offline evaluation of the skill levels of the non-player virtual roles; then the battle capability test is performed on the non-player virtual roles and a player virtual role based on the initial evaluation results obtained through offline evaluation and the preset evaluation results to realize online skill level evaluation, and the battle capability test is automatically performed in combination of offline evaluation and online evaluation, which can realize accurate and objective evaluation of the skill levels of the non-player virtual roles and the player virtual role, and reduces unnecessary waste of human resources. BRIEF DESCRIPTION OF DRAWINGS

[0011] FIG. 1 is a flow diagram of a virtual role test evaluation method according to an embodiment of the present disclosure;

[0012] FIG. 2 is a flow diagram of a virtual role test evaluation method according to another embodiment of the present disclosure;

[0013] Fig. 3 is a flow diagram of a virtual role testing evaluation method according to an embodiment of the present disclosure;

[0014] Fig. 4 is a flow diagram of a virtual role testing evaluation method according to an embodiment of the present disclosure;

[0015] Fig. 5 is a flow diagram of a virtual role testing evaluation method according to an embodiment of the present disclosure;

[0016] Fig. 6 is a flow diagram of a virtual role testing evaluation method according to an embodiment of the present disclosure;

[0017] Fig. 7 is a structural diagram of a virtual role testing evaluation device according to an embodiment of the present disclosure;

[0018] Fig. 8 is a structural diagram of a server according to an embodiment of the present disclosure. DETAILED DESCRIPTION

[0019] In order to make the objectives, technical solutions and advantages of the embodiments of the present disclosure clearer, the technical solutions in the embodiments of the present disclosure will be described clearly and completely below with reference to the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are only part of the embodiments of the present disclosure, rather than all the embodiments of the present disclosure.

[0020] Therefore, the following detailed description of the embodiments of the present disclosure provided in the drawings is not intended to limit the scope of the claimed present disclosure, but only represents selected embodiments of the present disclosure. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present disclosure without creative labor are within the scope of protection of the present disclosure.

[0021] In the description of the present disclosure, it should be noted that if the terms "upper", "lower", etc. indicate the orientation or positional relationship shown in the drawings, or the orientation or positional relationship commonly used when the product is used, and are only used to facilitate the description of the present disclosure and simplify the description, and do not indicate or imply that the device or element referred to must have a particular orientation, be constructed and operated in a particular orientation, and therefore cannot be understood as a limitation on the present disclosure.

[0022] In addition, the terms "first", "second", and the like in the description and claims of the present disclosure and the above drawings are used to distinguish similar objects, and do not necessarily indicate a particular order or sequence. It should be understood that the data thus used can be interchanged, where appropriate, so that the embodiments of the present disclosure described herein can be carried out in other sequences than those illustrated or described herein. Furthermore, the terms "comprise" and "have", and any variations thereof, are intended to cover non-exclusive inclusion, for example, a process, method, system, product, or apparatus that comprises a list of steps or units is not necessarily limited to those steps or units that are clearly listed, but can include other steps or units that are not clearly listed or inherent to such processes, methods, products, or apparatuses.

[0023] It should be noted that the features in the embodiments of the present disclosure can be combined with each other without conflict.

[0024] The following first explains the terms involved in the embodiments of the present disclosure.

[0025] Deep reinforcement learning: a branch of artificial intelligence that combines deep learning and reinforcement learning techniques. Deep learning uses multi-layer neural networks to process and capture patterns in data; reinforcement learning is a method of learning optimal strategies through the interaction of AI with the environment. Deep reinforcement learning estimates the value or strategy of AI interacting with the environment through deep neural networks, so as to effectively train AI that can cope with complex environments.

[0026] MySQL: an open source relational database management system in the field of software development.

[0027] Hive: a data warehouse infrastructure built on Apache Hadoop (an open source distributed storage and processing framework). It provides a SQL-like query language for querying and managing large data sets stored in Hadoop (distributed system infrastructure) clusters.

[0028] Redis database: an open source in-memory data structure storage system.

[0029] Statistical learning method: a series of algorithms and techniques based on statistical theory, which can be used to extract patterns or knowledge from data.

[0030] QuickSkill: a statistical learning method that can quickly learn the ability score of a novice.

[0031] Elo: A method used to assess the relative skill level of individuals participating in a competition, such as chess players, athletes, or other competitors. Originally designed for chess, it has been widely adopted in various sports and games.

[0032] Kafka: An open-source distributed streaming platform that provides a high-throughput, low-latency messaging system for building real-time data pipelines and stream processing applications.

[0033] Sample: A single instance or data point in a dataset, which is the smallest unit of data required for training artificial intelligence models.

[0034] Data-driven: A decision-making and problem-solving approach that emphasizes using relatively objective data analysis to guide the decision-making process, rather than relying solely on human intuition and subjective experience.

[0035] Skill rating: A numerical system used in online competitive games to express the skill level of players.

[0036] Competitive robot: A participant in online competitive games controlled by artificial intelligence technology.

[0037] The embodiments of the present disclosure provide a virtual role test evaluation method, applied to a server. The virtual role test evaluation method provided by the embodiments of the present disclosure is explained and described below.

[0038] FIG. 1 is a flowchart of a virtual role test evaluation method provided by the embodiments of the present disclosure. As shown in FIG. 1, the method can include the following steps S101-S103:

[0039] S101, test the fighting ability of a plurality of non-player virtual roles, and obtain an initial evaluation result of the non-player virtual roles.

[0040] The non-player virtual role can also be referred to as an NPC (non-player character) or a robot or a competitive robot. The non-player virtual role is a competitive robot trained by deep reinforcement learning technology, and the skill levels of different non-player virtual roles in the plurality of non-player virtual roles are different. The plurality of non-player virtual roles can belong to a non-player virtual role pool, and the non-player virtual role pool includes a plurality of candidate non-player virtual roles, and the plurality of non-player virtual roles are part of the plurality of candidate non-player virtual roles.

[0041] In some embodiments, the multiple non-player virtual characters are controlled to fight each other to test the fighting ability of the multiple non-player virtual characters, a fighting result between the non-player virtual characters is obtained, and an initial evaluation result of each non-player virtual character is determined based on the fighting result between the non-player virtual characters.

[0042] In addition, the process of testing the fighting ability of the multiple non-player virtual characters can be performed offline, and the process can also be referred to as a process of offline self-playing evaluation.

[0043] S102, according to the initial evaluation result of the non-player virtual character and the preset evaluation result of the player virtual character, the fighting ability of the non-player virtual character and the player virtual character is tested, and a first fighting result is obtained.

[0044] The player virtual character is controlled by a player, and the non-player virtual character is not controlled by the player. In addition to not being controlled by the player, the appearance and skill basic attributes of the non-player virtual character and the player virtual character are similar.

[0045] In actual application, the preset evaluation result of the player virtual character is a preset evaluation result. When the player virtual character participates in a competition for the first time, the preset evaluation result of each player virtual character is the same, which is a default preset evaluation result. With multiple fighting ability tests on the player virtual character, the preset evaluation result of the player virtual character will be different.

[0046] It should be noted that the process of testing the fighting ability of the non-player virtual character and the player virtual character can be referred to as an online evaluation process. Controlling the non-player virtual character and the player virtual character to fight can realize the testing of the fighting ability of the non-player virtual character and the player virtual character.

[0047] S103, according to the first fighting result, the initial evaluation result and the preset evaluation result are updated to obtain a target evaluation result of the non-player virtual character and a target evaluation result of the player virtual character.

[0048] In the embodiments of the present disclosure, after each fighting between the non-player virtual character and the player virtual character ends, the initial evaluation result of the non-player virtual character and the preset evaluation result of the player virtual character can be updated simultaneously according to the first fighting result, or the initial evaluation result of the non-player virtual character and the preset evaluation result of the player virtual character can be updated sequentially. The embodiments of the present disclosure do not specifically limit this.

[0049] It is worth mentioning that the target evaluation result of the non-player virtual role is based on the evaluation results obtained through two evaluations, i.e., offline self-play evaluation and online alignment evaluation, and the target evaluation result of the non-player virtual role is more accurate. The offline self-play evaluation is a battle between non-player virtual roles, and the online alignment evaluation is a mixed battle between non-player virtual roles and player virtual roles, so that the system deviation between the ability scores of the human and the robot can be corrected.

[0050] In addition, based on the initial evaluation result of the non-player virtual role and the preset evaluation result of the player virtual role, the battle ability test is performed on the non-player virtual role and the player virtual role, and the allocation of the non-player virtual role and the player virtual role participating in the battle is also more reasonable, which makes the first battle result obtained more accurate, so that the target evaluation result obtained by updating is also more accurate.

[0051] In summary, the embodiment of the disclosure provides a virtual role test evaluation method, which comprises: performing a battle ability test on a plurality of non-player virtual roles to obtain an initial evaluation result of the non-player virtual roles; performing a battle ability test on the non-player virtual roles and player virtual roles based on the initial evaluation result of the non-player virtual roles and a preset evaluation result of the player virtual roles to obtain a first battle result; and updating the initial evaluation result and the preset evaluation result based on the first battle result to obtain a target evaluation result of the non-player virtual roles and a target evaluation result of the player virtual roles. The battle ability test is first performed on a plurality of non-player virtual roles to realize offline evaluation of the skill level of the non-player virtual roles; and then the initial evaluation result obtained through offline evaluation and the preset evaluation result are used to perform a battle ability test on the non-player virtual roles and the player virtual roles to realize online skill level evaluation. The combination of offline evaluation and online evaluation for battle ability test can realize accurate and objective evaluation of the skill level of the non-player virtual roles and the player virtual roles, and unnecessary waste of human resources is also reduced.

[0052] Optionally, FIG. 2 is a flowchart of a virtual role test evaluation method according to an embodiment of the disclosure, as shown in FIG. 2, the process of performing a battle ability test on a plurality of non-player virtual roles to obtain an initial evaluation result of the non-player virtual roles in S101 can include the following steps S201-S203:

[0053] S201, filtering a plurality of candidate non-player virtual roles based on preset role attribute parameters to obtain a first matching queue.

[0054] The robot template table is preset, and the robot template table includes preset role attribute parameters of the non-player virtual role.

[0055] In some embodiments, the preset role attribute parameter includes: an id (identification) of the non-player virtual role, a professional type, a skill type, a gender, an appearance, an open time period, and the like, wherein the id of the non-player virtual role is for facilitating retrieval of the non-player virtual role. In addition, when the data amount of the preset role attribute parameter is small, the preset role attribute parameter can be stored in an Excel configuration table, and when the data amount of the preset role attribute parameter is large, the preset role attribute parameter can be stored in a MySQL database.

[0056] In addition, from the plurality of candidate non-player virtual roles, the non-player virtual roles matched with the preset role attribute parameter are screened, and the first matching queue includes: a plurality of non-player virtual roles matched with the preset role attribute parameter. For example, the number of the plurality of candidate non-player virtual roles is 1000, and the number of the plurality of non-player virtual roles matched with the preset role attribute parameter is 100.

[0057] S202, determining a first battle formation from the first matching queue according to a preset battle rule.

[0058] The first battle formation is composed of the non-player virtual roles of the two battle parties, wherein the preset battle rule includes a constraint condition of the competitive attribute parameters of the non-player virtual roles of the two battle parties.

[0059] In the embodiments of the present disclosure, the non-player virtual roles of the two battle parties in the first battle formation include: one-party non-player virtual roles and another-party non-player virtual roles, and each-party non-player virtual roles can include: at least one non-player virtual role. Wherein, according to the requirement in the preset battle rule, one first battle formation meeting the preset battle rule can be searched from the first matching queue.

[0060] It should be noted that the preset battle rule refers to a rule description for the battle formation, and in some embodiments, the constraint condition of the competitive attribute parameters of the non-player virtual roles of the two battle parties can include: at least one of the following: a professional constraint of the non-player virtual roles of the two battle parties, a gender constraint of the non-player virtual roles of the two battle parties, and a friend constraint of the non-player virtual roles of the two battle parties.

[0061] The occupation constraint of the non-player virtual characters of the two parties in the battle is used to constrain the number of non-player virtual characters of the same occupation, the combination mode of non-player virtual characters of different occupations in the non-player virtual characters of each party; the gender constraint of the non-player virtual characters of the two parties in the battle is used to constrain the number of non-player virtual characters of the same gender, the combination mode of non-player virtual characters of different genders in the non-player virtual characters of each party; and the friend constraint of the non-player virtual characters of the two parties in the battle is used to constrain the number of friend non-player virtual characters, the combination mode of friend non-player virtual characters and non-friend non-player virtual characters in the non-player virtual characters of each party.

[0062] For example, the occupation constraint of the non-player virtual characters of the two parties in the battle is used to constrain that there is at most one non-player virtual character of a certain occupation in the non-player virtual characters of each party, the number of non-player virtual characters of a certain occupation is equal, and the combination of non-player virtual characters of a certain occupation cannot appear, etc.

[0063] For example, the gender constraint of the non-player virtual characters of the two parties in the battle is used to constrain that the non-player virtual characters of each party are all of a preset number of males, or are all of a preset number of females, or are of a male-female combination type. For the non-player virtual characters of the male-female combination type, it is further necessary to constrain the number of non-player virtual characters of the same gender in the non-player virtual characters of each party. For example, the non-player virtual characters of each party include two male non-player virtual characters and one female non-player virtual character.

[0064] For example, the friend constraint of the non-player virtual characters of the two parties in the battle is used to constrain that the non-player virtual characters of each party are all of a preset number of friend non-player virtual characters, or are all of a preset number of non-friend non-player virtual characters, or are of a friend and non-friend combination type. For the non-player virtual characters of the friend and non-friend combination type, it is further necessary to constrain the number of friend and non-friend non-player virtual characters in the non-player virtual characters of each party. For example, the non-player virtual characters of each party include two friend non-player virtual characters and two non-friend non-player virtual characters.

[0065] In the embodiments of the present disclosure, if a friend is added between two non-player virtual characters, the two non-player virtual characters are friend non-player virtual characters to each other.

[0066] It is worth noting that the battles between non-player virtual characters are initiated in a loop on the server, and before each battle, the battle formation is determined based on preset battle rules.

[0067] S203, performing a battle ability test based on the first battle formation to obtain an initial evaluation result of the non-player virtual character.

[0068] In some implementations, testing combat ability based on the first battle lineup refers to controlling one non-player virtual character in the first battle lineup to fight against another non-player virtual character. The battle results between the non-player virtual characters are obtained, and an initial evaluation result for each non-player virtual character is determined based on these results using a preset algorithm. This preset algorithm can be a preset ability score algorithm, and the initial evaluation result can be the ability score value of the non-player virtual character. For example, the ability score algorithm can use modern statistical learning methods such as the QuickSkill algorithm or classic scoring algorithms such as the Elo algorithm (an algorithm used to evaluate the relative skill level of players or contestants), or it can use a scoring update formula designed by game designers.

[0069] In practical applications, the battle results between non-player virtual characters can be used as the first settlement log. The settlement log is collected to the Hive offline data platform, where it is processed into a format that is easy for the preset algorithm to process. Then, based on the first settlement log, the Hive offline data platform determines the initial evaluation result of the non-player virtual characters.

[0070] The first settlement log may include: the faction of the non-player virtual character IDs of both sides, the win / loss record, and various performance scores. The performance scores vary depending on the game type. For combat-based competitive games, they are usually kills, damage points, damage taken, and healing points, while for sports-based competitive games, they are scores for offense, defense, and assists.

[0071] In addition, after obtaining the initial evaluation results of the non-player virtual characters, the initial evaluation results of the non-player virtual characters can be stored in the Redis database for use in the S102 process.

[0072] In this embodiment of the disclosure, processes S201 to S203 can initialize the skill levels of all non-player virtual characters offline. A large number of matches can be played with non-player virtual characters of different skill levels, collecting performance data for each non-player virtual character in each match. Then, based on this performance data, a statistical learning method is used to continuously update the ability score of each non-player virtual character until the ability scores of all non-player virtual characters converge, yielding the initial evaluation results for the non-player virtual characters.

[0073] Optionally, Figure 3 is a flowchart illustrating a virtual character testing and evaluation method provided in this embodiment of the present disclosure. As shown in Figure 3, the process in S102 above, which tests the combat ability of a non-player virtual character and a player virtual character based on the initial evaluation result of the non-player virtual character and the preset evaluation result of the player virtual character to obtain the first combat result, may include the following steps S301-S305:

[0074] S301. Determine the set of non-player virtual characters from multiple candidate non-player virtual characters.

[0075] Among them, multiple non-player virtual characters are selected from multiple candidate non-player virtual characters in the non-player virtual character pool to form a set of non-player virtual characters.

[0076] S302. Determine the lineup of player virtual characters from multiple candidate player virtual characters.

[0077] It should be noted that during game operation, a large number of players participate, resulting in multiple candidate player virtual characters. These candidate player virtual characters are those awaiting matchmaking, meaning they have not yet formed a faction. These multiple candidate player virtual characters can exist within a pool of player virtual characters.

[0078] Among them, multiple player virtual characters are selected from multiple candidate player virtual characters to form the player virtual character lineup.

[0079] S303. Based on the allocation rules corresponding to the game state, the preset evaluation results of each player virtual character in the player virtual character lineup, and the initial evaluation results of each non-player virtual character in the non-player virtual character set, determine the non-player virtual character lineup from the non-player virtual character set.

[0080] Among them, the number of non-player virtual characters in the non-player virtual character set is greater than the number of player virtual characters in the player virtual character lineup. The non-player virtual characters in the non-player virtual character lineup are a subset of the characters in the non-player virtual character set. The number of non-player virtual characters in the non-player virtual character lineup is roughly equivalent to the number of player virtual characters in the player virtual character lineup.

[0081] Optionally, if the game status indicates that the player virtual character in the player virtual character lineup is the player virtual character corresponding to the novice player, that is, the game status is still a novice game, or the player virtual character in the player virtual character lineup is in a state of losing the battle, then the allocation rule corresponding to the game status is the first allocation rule, which indicates that the player virtual character lineup is allocated a non-player virtual character lineup with a lower battle ability than the player virtual character lineup.

[0082] Furthermore, the first allocation rule can also be called the "giving-a-warm-up" allocation rule. In this case, the initial evaluation results of the non-player virtual character lineup are significantly lower than the preset evaluation results of each player virtual character in the player virtual character lineup. That is, the combat ability of the non-player virtual character lineup is lower than that of the player virtual character lineup, making it easier for the player virtual character lineup to win and improving the gaming experience for the corresponding user. For the "giving-a-warm-up" allocation rule, the non-player virtual character lineup can also be randomly determined from the set of non-player virtual characters.

[0083] Optionally, if the game status indicates that the player's virtual character lineup has been waiting for a battle for a longer period than or equal to a preset time, i.e., the game status is overtime, then the allocation rule corresponding to the game status is the second allocation rule, which indicates that the player's virtual character lineup is allocated a non-player virtual character lineup with combat capabilities equivalent to the player's virtual character lineup.

[0084] In this embodiment, the second allocation rule is an evenly matched allocation rule. In this case, the initial evaluation results of the non-player virtual character lineup are close to the preset evaluation results of each player virtual character in the player virtual character lineup; that is, the difference between the initial evaluation results and the preset evaluation results is within a preset range. This means that the combat capabilities of the non-player virtual character lineup and the player virtual character lineup are comparable, resulting in a more balanced ability level between the participating factions.

[0085] For example, the initial evaluation result of non-player virtual characters in the non-player virtual character lineup being close to the preset evaluation result of each player virtual character in the player virtual character lineup means that the average initial evaluation result of at least one non-player virtual character in the non-player virtual character lineup is the same as or close to the average preset evaluation result of at least one player virtual character in the player virtual character lineup.

[0086] Additionally, if some non-player virtual characters among the matched player virtual characters in the game status indicator have disconnected, then from multiple candidate non-player virtual characters, the non-player virtual characters with the same preset evaluation result as the disconnected virtual characters will be identified, and the identified non-player virtual characters will be used to replace the disconnected virtual characters.

[0087] S304. Determine the second battle lineup based on the lineup of non-player virtual characters and the lineup of player virtual characters.

[0088] The second battle lineup includes: a lineup of non-player virtual characters and a lineup of player virtual characters; the lineup of non-player virtual characters includes at least one non-player virtual character, and the lineup of player virtual characters includes at least one player virtual character.

[0089] Furthermore, the number of non-player virtual characters in the non-player virtual character lineup is roughly the same as the number of player virtual characters in the player virtual character lineup. For example, if there are 2 non-player virtual characters in the non-player virtual character lineup, there are also 2 player virtual characters in the player virtual character lineup; or if there are 5 non-player virtual characters in the non-player virtual character lineup, there are also 5 player virtual characters in the player virtual character lineup.

[0090] S305. Based on the second battle lineup, a battle capability test was conducted to obtain the first battle result.

[0091] The second battle lineup's ability test can be considered a test of human-machine battle ability.

[0092] In some implementations, the non-player virtual characters in the second battle lineup and the player virtual characters in the player virtual character lineup are tested against each other to obtain the first battle result.

[0093] In practical applications, the first battle result is recorded in the second settlement log between the non-player virtual character and the player virtual character. This second settlement log includes: the non-player virtual character's and player virtual character's IDs, their faction affiliation, the win / loss result, and various performance scores. Performance scores vary depending on the game type; for combat-based competitive games, they typically include kill count, damage score, damage taken score, and healing score, while for sports-based competitive games, they include scores for offense, defense, and assists.

[0094] Furthermore, the initial and preset evaluation results are updated based on the first battle results, allowing the ability scores of real players (player virtual characters) and bots (non-player virtual characters) to be updated simultaneously in the online battle environment, gradually achieving alignment between the ability scores of real players and bots. This process is the online alignment process; the offline evaluation stage only involves battles between bots, while the online stage involves a mix of real players and bots. The online alignment evaluation corrects any previous systemic discrepancies in the ability scores between real players and bots.

[0095] Optionally, Figure 4 is a flowchart illustrating a testing and evaluation method for virtual characters provided in this embodiment of the present disclosure. As shown in Figure 4, the process of determining a set of non-player virtual characters from multiple candidate non-player virtual characters in S301 above may include the following steps S401-S402:

[0096] S401. Based on preset character attribute parameters, filter multiple candidate non-player virtual characters to obtain a second matching queue.

[0097] The preset character attribute parameters here are similar to those in S201 above, and will not be repeated here.

[0098] In some implementations, non-player virtual characters that match preset character attribute parameters are selected from a plurality of candidate non-player virtual characters. The second matching queue includes a plurality of non-player virtual characters that match the preset character attribute parameters. For example, the number of candidate non-player virtual characters is 1000, and the number of non-player virtual characters that match the preset character attribute parameters is 120.

[0099] S402. Based on the preset battle rules, determine the set of non-player virtual characters from the second match queue.

[0100] The preset battle rules include constraints on the competitive attribute parameters of both non-player virtual character lineups and player virtual character lineups.

[0101] Optionally, the process of determining the lineup of player virtual characters from multiple candidate player virtual characters in S302 above may include:

[0102] According to the preset battle rules, the lineup of player virtual characters is determined from multiple candidate player virtual characters. The preset battle rules include the constraints on the competitive attribute parameters of the non-player virtual character lineup and the player virtual character lineup.

[0103] Among them, the preset battle rules refer to the rule description for the battle lineup. In some implementations, the preset battle rules may include: the class constraints of non-player virtual characters in the non-player virtual character lineup and player virtual characters in the player virtual character lineup.

[0104] For example, the number of virtual characters of a certain class in both the non-player virtual character lineup and the player virtual character lineup can be at most one; the number of virtual characters of a certain class in both the non-player virtual character lineup and the player virtual character lineup is equal; and a certain combination of virtual characters of a certain class cannot appear.

[0105] Therefore, based on the preset battle rules, a set of non-player virtual characters and a lineup of player virtual characters that meet the preset battle rules can be selected from the second matching queue and multiple candidate player virtual characters, respectively.

[0106] Optionally, Figure 5 is a flowchart illustrating a virtual character testing and evaluation method provided in this embodiment of the present disclosure. As shown in Figure 5, the process in S103 above, which updates the initial evaluation result and the preset evaluation result based on the first battle result to obtain the target evaluation result for the non-player virtual character and the target evaluation result for the player virtual character, may include the following steps S501-S502:

[0107] S501. Obtain the target first battle result from the first battle result when the allocation rule is the second allocation rule.

[0108] The second allocation rule is one where the non-player virtual character lineup and the player virtual character lineup are of similar ability. In other words, the second allocation rule is the aforementioned evenly matched allocation rule.

[0109] In this embodiment, the first battle result includes: the first battle result corresponding to the evenly matched allocation rule, and the first battle result corresponding to the "giving aid" allocation rule. Since the non-player virtual character lineups and player virtual character lineups participating in the battle under the "giving aid" allocation rule are not of equal ability, the first battle result based on the "giving aid" allocation rule is meaningless. However, since the non-player virtual character lineups and player virtual character lineups participating in the battle under the evenly matched allocation rule are of equal ability, updating the initial evaluation result and the preset evaluation result based on the target first battle result corresponding to the evenly matched allocation rule is more reliable.

[0110] In practical applications, the target first battle result corresponding to the second allocation rule is selected from the first battle result, and the second settlement log of the target first battle result is generated as a sample. The second settlement log of the target first battle result is the online settlement log, and the online settlement log is collected from the server in real time to the data middleware Kafka.

[0111] S502. Based on the first battle result, update the initial evaluation result and the preset evaluation result to obtain the target evaluation result for non-player virtual characters and the target evaluation result for player virtual characters.

[0112] Specifically, for real players participating in the competition for the first time, the preset evaluation result of the player's virtual character can be used to initialize a default value; the initial evaluation result of the non-player virtual character is the result obtained based on offline self-play evaluation in S101 above. The initial evaluation result and the preset evaluation result can be stored in the Redis database and retrieved by the player's virtual character ID or the non-player's virtual character ID.

[0113] It should be noted that a preset ability score algorithm can be used to update the initial and preset evaluation results based on the first battle result, resulting in the target evaluation results for both the non-player virtual character and the player virtual character. These target evaluation results can be stored in a Redis database, overwriting the original initial and preset evaluation results. These target evaluation results for the non-player virtual character and the player virtual character can be referred to as the ability score of the non-player virtual character and the ability score of the player virtual character, respectively.

[0114] Optionally, Figure 6 is a flowchart illustrating a testing and evaluation method for virtual characters provided in this embodiment of the present disclosure. As shown in Figure 6, the method may further include the following steps S601-S603:

[0115] S601. Based on the preset evaluation results of the player's virtual character, determine the third battle lineup from multiple candidate player virtual characters.

[0116] During game operation, a large number of players participate, resulting in multiple candidate player virtual characters. These candidate player virtual characters are those awaiting matchmaking, meaning they haven't yet formed factions. These candidate player virtual characters can exist within a pool of player virtual characters.

[0117] It should be noted that the third battle lineup includes virtual characters from both sides, and each side's virtual characters include at least one virtual player character.

[0118] S602. Based on the third battle lineup, a battle ability test was conducted to obtain the second battle result.

[0119] In this embodiment of the disclosure, the virtual characters of both players in the third battle lineup are controlled to conduct a battle ability test to obtain the second battle result.

[0120] S603. Based on the results of the second battle, update the preset evaluation results of the virtual characters of both players in the third battle lineup to obtain the target evaluation results of the virtual characters of both players.

[0121] In some implementations, a third settlement log corresponding to the second battle result is obtained, and the settlement log is collected to the Hive offline data platform. The third settlement log is processed into a format that is easy for the preset algorithm to process on the Hive offline data platform. Then, based on the third settlement log, the preset evaluation results of the virtual characters of both players are updated on the Hive offline data platform using a preset ability score algorithm to obtain the target evaluation results of the virtual characters of both players.

[0122] In addition, the target evaluation results of both players' virtual characters can be stored in a Redis database, overriding the preset evaluation results of both players' virtual characters.

[0123] Optionally, the process in S601 above of determining the third battle lineup from multiple candidate player virtual characters based on the preset evaluation results of the player virtual characters may include:

[0124] Based on the preset evaluation results of the player's virtual character and the second allocation rule, the third battle lineup is determined from multiple candidate player virtual characters.

[0125] The second allocation rule is one where the virtual characters of both players have comparable abilities. This is the same as the aforementioned evenly matched allocation rule.

[0126] In this embodiment of the disclosure, a second allocation rule and a preset battle rule are used to determine the virtual characters of both sides from multiple virtual characters of players. The preset evaluation results of the virtual characters of both sides are comparable. For example, the average preset evaluation result of one side's virtual character may be the same as or similar to the average preset evaluation result of the other side's virtual character.

[0127] It should be noted that the preset evaluation results of the virtual characters of both players in the third battle lineup are updated to obtain the target evaluation results of the virtual characters of both players. This makes it easier to assign non-player virtual characters with similar abilities to player virtual characters in the target allocation of the AI ​​game. In this way, based on the results of the first battle, the target evaluation results of non-player virtual characters and player virtual characters can also be obtained accurately and reliably, thus improving the accuracy of the ability evaluation of non-player virtual characters and player virtual characters.

[0128] In addition, the evaluation results in this embodiment can be updated in real time.

[0129] In summary, the present disclosure provides a method for testing and evaluating virtual characters. It quantifies the comprehensive ability level of non-player virtual characters through a data-driven approach. The main objective of this data-driven approach is to design an ability evaluation process independent of the training process of non-player virtual characters, and to quantify the ability level of non-player virtual characters. The quantified value is called the "ability score of non-player virtual characters", which is the target evaluation result of non-player virtual characters.

[0130] Furthermore, both offline self-play evaluation and online alignment can be highly automated, significantly reducing labor costs. Testing and evaluation of non-player virtual characters does not rely on manual testing and can cover all difficulty levels of non-player virtual characters, thus saving time and effort for game designers and testers. The evaluation results of non-player virtual character ability levels are accurate and highly comparable to the skill levels of player virtual characters, thereby precisely meeting the actual deployment needs in game design.

[0131] This disclosure obtains the relative ability levels of all competitive bots in the game through offline self-play evaluation. The results of this step can be used to help game designers evaluate the rationality of the bot difficulty gradient, and provide a better initial value for the bots' ability scores before the game goes live or is updated, thus ensuring a good human-bot experience in the early stages of the game. The online alignment step goes a step further, correcting the discrepancy between the bots' ability scores and those of real players, aligning the skill levels of bots and real players, thereby enabling more precise control of the human-bot difficulty and optimizing the player's competitive experience.

[0132] The following describes the testing and evaluation apparatus, server, and storage medium for virtual characters used to execute the testing and evaluation method for virtual characters provided in this disclosure. For the specific implementation process and technical effects, please refer to the relevant content of the above-mentioned testing and evaluation method for virtual characters, which will not be repeated below.

[0133] Figure 7 is a schematic diagram of the structure of a virtual character testing and evaluation device provided in an embodiment of this disclosure. As shown in Figure 7, the device includes a testing module 101 and an update module 102.

[0134] The testing module 101 is configured to perform combat ability tests on multiple non-player virtual characters to obtain initial evaluation results for the non-player virtual characters; based on the initial evaluation results of the non-player virtual characters and the preset evaluation results of the player virtual characters, it performs combat ability tests between the non-player virtual characters and the player virtual characters to obtain the first combat result; the updating module 102 is configured to update the initial evaluation results and the preset evaluation results based on the first combat result to obtain the target evaluation results for the non-player virtual characters and the target evaluation results for the player virtual characters.

[0135] In this device, the combat ability of multiple non-player virtual characters is first tested, realizing an automatic offline assessment of the skill level of non-player virtual characters. Then, based on the initial assessment results and preset assessment results obtained from the offline assessment, the combat ability of non-player virtual characters and player virtual characters is tested, realizing an online skill level assessment. The automatic combination of offline and online assessment methods for combat ability testing can achieve accurate and objective assessment of the skill level of non-player virtual characters and player virtual characters, and also reduce unnecessary waste of human resources.

[0136] Optionally, the test module 101 is specifically configured to perform the following actions: filtering multiple candidate non-player virtual characters according to preset character attribute parameters to obtain a first matching queue; determining a first battle lineup from the first matching queue according to preset battle rules, wherein the first battle lineup consists of non-player virtual characters from both sides of the battle, and the preset battle rules include: constraints on the competitive attribute parameters of the non-player virtual characters from both sides of the battle; and conducting a battle ability test based on the first battle lineup to obtain an initial evaluation result of the non-player virtual characters.

[0137] Based on the above modules, the skill levels of all non-player virtual characters can be initialized offline. A large number of matches can be played with non-player virtual characters of different levels to collect performance data of each non-player virtual character in each match. Then, based on this performance data, statistical learning methods are used to continuously update the ability score of each non-player virtual character until the ability scores of all non-player virtual characters converge, thus obtaining the initial evaluation results of the non-player virtual characters. This facilitates the subsequent accurate and objective evaluation of the skill levels of non-player virtual characters and player virtual characters.

[0138] Optionally, the test module 101 is specifically configured to perform the following actions: determining a set of non-player virtual characters from multiple candidate non-player virtual characters; determining a lineup of player virtual characters from multiple candidate player virtual characters; determining a lineup of non-player virtual characters from the set of non-player virtual characters based on the allocation rules corresponding to the game state, the preset evaluation results of each player virtual character in the lineup of player virtual characters, and the initial evaluation results of each non-player virtual character in the set of non-player virtual characters; determining a second battle lineup based on the lineup of non-player virtual characters and the lineup of player virtual characters; and conducting a battle ability test based on the second battle lineup to obtain the first battle result.

[0139] Based on the aforementioned modules, human-computer interaction is tested to assess combat capabilities. The initial and preset evaluation results are then updated based on the first battle outcome, ensuring that the ability scores of both real players (player virtual characters) and bots (non-player virtual characters) are updated simultaneously in the online battle environment, gradually achieving alignment between their ability scores. Furthermore, since online battles involve a mix of real players and bots, online alignment evaluation corrects any previous systemic discrepancies in ability scores between real players and bots, thus facilitating accurate and objective assessment of the skill levels of both non-player virtual characters and player virtual characters.

[0140] Optionally, if the game status indicates that the player virtual character in the player virtual character lineup is the player virtual character corresponding to the novice player, or that the player virtual character in the player virtual character lineup is in a state of defeat in battle, then the allocation rule corresponding to the game status is the first allocation rule, which indicates that the player virtual character lineup is allocated a non-player virtual character lineup with combat ability lower than the player virtual character lineup; if the game status indicates that the player virtual character lineup has been waiting for a battle for a period of time greater than or equal to a preset time, then the allocation rule corresponding to the game status is the second allocation rule, which indicates that the player virtual character lineup is allocated a non-player virtual character lineup with combat ability equivalent to the player virtual character lineup.

[0141] Based on the above embodiments, the first allocation rule is that the combat ability of the non-player virtual character lineup is lower than that of the player virtual character lineup, making it easier for the player virtual character lineup to win and improving the gaming experience for the corresponding user. The second allocation rule is an evenly matched allocation rule, ensuring that the combat abilities of the non-player virtual character lineup and the player virtual character lineup are comparable, thus creating a more balanced skill level among the participating factions.

[0142] Optionally, the test module 101 is specifically configured to perform the following actions: filtering multiple candidate non-player virtual characters according to preset character attribute parameters to obtain a second matching queue; and determining a set of non-player virtual characters from the second matching queue according to preset battle rules, wherein the preset battle rules include: constraints on the competitive attribute parameters of the non-player virtual character lineup and the player virtual character lineup.

[0143] Optionally, the test module 101 is specifically configured to determine the lineup of player virtual characters from multiple candidate player virtual characters according to preset battle rules. The preset battle rules include constraints on the competitive attribute parameters of the non-player virtual character lineup and the player virtual character lineup.

[0144] Optionally, the update module 102 is specifically configured to retrieve the target first battle result from the first battle result when the allocation rule is the second allocation rule, wherein the second allocation rule is an allocation rule in which the abilities of the non-player virtual character lineup and the player virtual character lineup are equivalent; based on the target first battle result, the initial evaluation result and the preset evaluation result are updated to obtain the target evaluation result of the non-player virtual character and the target evaluation result of the player virtual character.

[0145] Based on the above modules, the non-player virtual character lineups and player virtual character lineups participating in the battle are of similar ability when using the evenly matched allocation rule (second allocation rule). It is more reliable to update the initial evaluation results and preset evaluation results based on the first battle result corresponding to the evenly matched allocation rule.

[0146] Optionally, the above-mentioned device may further include a determining module, a first testing module, and a first updating module, wherein the determining module is configured to determine a third battle lineup from multiple candidate player virtual characters based on the preset evaluation results of the player virtual characters; the first testing module is configured to perform a battle ability test based on the third battle lineup to obtain a second battle result; and the first updating module is configured to update the preset evaluation results of the player virtual characters of both sides in the third battle lineup based on the second battle result to obtain the target evaluation results of the player virtual characters of both sides.

[0147] Optionally, the determining module is specifically configured to determine a third battle lineup from multiple candidate player virtual characters based on the preset evaluation results of the player virtual characters and the second allocation rule, wherein the second allocation rule is an allocation rule in which the abilities of the player virtual characters on both sides are comparable.

[0148] Based on the above modules, the preset evaluation results of the virtual characters of both players in the third battle lineup are updated to obtain the target evaluation results of the virtual characters of both players. This makes it easier to assign non-player virtual characters with similar abilities to player virtual characters in the target allocation of human-machine games. In this way, based on the results of the first battle, the target evaluation results of non-player virtual characters and player virtual characters can also be obtained accurately and reliably, thus improving the accuracy of the ability evaluation of non-player virtual characters and player virtual characters.

[0149] In summary, the virtual character testing and evaluation device provided in this embodiment first tests the combat ability of multiple non-player virtual characters, realizing automatic offline evaluation of the skill level of non-player virtual characters; then, based on the initial evaluation results and preset evaluation results obtained from the offline evaluation, it tests the combat ability of non-player virtual characters against player virtual characters, realizing online skill level evaluation. By automatically adopting a combination of offline and online evaluation for combat ability testing, it can achieve accurate and objective evaluation of the skill level of non-player virtual characters and player virtual characters, and also reduces unnecessary waste of human resources.

[0150] It should be explained that the above-mentioned device is used to execute the method provided in the foregoing embodiments, and its implementation principle and technical effect are similar, and the rest will not be repeated here.

[0151] These modules can be one or more integrated circuits configured to implement the above methods, such as one or more Application Specific Integrated Circuits (ASICs), one or more digital signal processors (DSPs), or one or more Field Programmable Gate Arrays (FPGAs). Alternatively, when a module is implemented using processing element scheduler code, the processing element can be a general-purpose processor, such as a Central Processing Unit (CPU) or other processor capable of calling program code. Furthermore, these modules can be integrated together as a system-on-a-chip (SOC).

[0152] Figure 8 is a schematic diagram of the structure of a server provided in an embodiment of the present disclosure. As shown in Figure 8, the server includes a processor 201 and a memory 202.

[0153] The memory 202 is used to store programs, and the processor 201 calls the programs stored in the memory 202 to execute the above method embodiments. The specific implementation and technical effects are similar, and will not be described in detail here.

[0154] For example, the method may include: testing the combat capabilities of multiple non-player virtual characters to obtain initial evaluation results of the non-player virtual characters; testing the combat capabilities of the non-player virtual characters against the player virtual characters based on the initial evaluation results of the non-player virtual characters and the preset evaluation results of the player virtual characters to obtain a first combat result; and updating the initial evaluation results and the preset evaluation results based on the first combat result to obtain a target evaluation result for the non-player virtual characters and a target evaluation result for the player virtual characters.

[0155] This method first tests the combat capabilities of multiple non-player virtual characters, achieving an automatic offline assessment of their skill levels. Then, based on the initial and preset assessment results obtained from the offline assessment, it tests the combat capabilities of non-player virtual characters against player virtual characters, achieving an online skill level assessment. By automatically combining offline and online assessments for combat capability testing, it can achieve an accurate and objective assessment of the skill levels of both non-player and player virtual characters, while also reducing unnecessary waste of human resources.

[0156] Optionally, multiple non-player virtual characters are tested for their combat abilities to obtain initial evaluation results. This includes: filtering multiple candidate non-player virtual characters based on preset character attribute parameters to obtain a first matchmaking queue; determining a first battle lineup from the first matchmaking queue according to preset battle rules, wherein the first battle lineup consists of non-player virtual characters from both sides, and the preset battle rules include: constraints on the competitive attribute parameters of the non-player virtual characters from both sides; and testing the combat abilities of the first battle lineup to obtain initial evaluation results for the non-player virtual characters.

[0157] In this embodiment, the skill levels of all non-player virtual characters can be initialized offline. A large number of games can be played with non-player virtual characters of different levels to collect performance data of each non-player virtual character in each game. Then, based on these performance data, statistical learning methods are used to continuously update the ability score of each non-player virtual character until the ability scores of all non-player virtual characters converge, thus obtaining the initial evaluation results of the non-player virtual characters. This facilitates the subsequent accurate and objective evaluation of the skill levels of non-player virtual characters and player virtual characters.

[0158] Optionally, based on the initial evaluation results of the non-player virtual characters and the preset evaluation results of the player virtual characters, a combat ability test is conducted between the non-player virtual characters and the player virtual characters to obtain a first combat result, including: determining a set of non-player virtual characters from multiple candidate non-player virtual characters; determining a lineup of player virtual characters from multiple candidate player virtual characters; determining a lineup of non-player virtual characters from the set of non-player virtual characters based on the allocation rules corresponding to the game state, the preset evaluation results of each player virtual character in the player virtual character lineup, and the initial evaluation results of each non-player virtual character in the set of non-player virtual characters; determining a second combat lineup based on the lineup of non-player virtual characters and the lineup of player virtual characters; and conducting a combat ability test based on the second combat lineup to obtain the first combat result.

[0159] In this embodiment, a battle ability test is conducted between humans and machines. Based on the first battle result, the initial evaluation result and the preset evaluation result are updated, allowing the ability scores of real players (player virtual characters) and robots (non-player virtual characters) to be updated simultaneously in the online battle environment, gradually achieving alignment between the ability scores of real players and robots. Furthermore, since online battles involve a mix of real players and robots, online alignment evaluation can correct any previous systemic discrepancies in the ability scores of real players and robots, thereby facilitating accurate and objective evaluation of the skill levels of both non-player virtual characters and player virtual characters.

[0160] Optionally, if the game status indicates that the player virtual character in the player virtual character lineup is the player virtual character corresponding to the novice player, or that the player virtual character in the player virtual character lineup is in a state of defeat in battle, then the allocation rule corresponding to the game status is the first allocation rule, which indicates that the player virtual character lineup is allocated a non-player virtual character lineup with combat ability lower than the player virtual character lineup; if the game status indicates that the player virtual character lineup has been waiting for a battle for a period of time greater than or equal to a preset time, then the allocation rule corresponding to the game status is the second allocation rule, which indicates that the player virtual character lineup is allocated a non-player virtual character lineup with combat ability equivalent to the player virtual character lineup.

[0161] In this embodiment, the first allocation rule is that the combat ability of the non-player virtual character lineup is lower than that of the player virtual character lineup, making it easier for the player virtual character lineup to win and improving the gaming experience for the corresponding user. The second allocation rule is an evenly matched allocation rule, ensuring that the combat abilities of the non-player virtual character lineup and the player virtual character lineup are comparable, thus creating a more balanced skill level among the participating factions.

[0162] Optionally, determining a set of non-player virtual characters from multiple candidate non-player virtual characters includes: filtering multiple candidate non-player virtual characters according to preset character attribute parameters to obtain a second matching queue; and determining a set of non-player virtual characters from the second matching queue according to preset battle rules, wherein the preset battle rules include: constraints on the competitive attribute parameters of the non-player virtual character lineup and the player virtual character lineup.

[0163] Optionally, determining the lineup of player virtual characters from multiple candidate player virtual characters includes: determining the lineup of player virtual characters from multiple candidate player virtual characters according to preset battle rules, wherein the preset battle rules include: constraints on the competitive attribute parameters of non-player virtual character lineups and player virtual character lineups.

[0164] Optionally, based on the first battle result, the initial evaluation result and the preset evaluation result are updated to obtain the target evaluation result for non-player virtual characters and the target evaluation result for player virtual characters. This includes: obtaining the target first battle result from the first battle result when the allocation rule is the second allocation rule, wherein the second allocation rule is an allocation rule in which the abilities of the non-player virtual character lineup and the player virtual character lineup are equivalent; and updating the initial evaluation result and the preset evaluation result based on the target first battle result to obtain the target evaluation result for non-player virtual characters and the target evaluation result for player virtual characters.

[0165] In this embodiment, the non-player virtual character lineup and player virtual character lineup participating in the battle using the evenly matched allocation rule (second allocation rule) are of comparable ability. Updating the initial evaluation result and the preset evaluation result based on the target first battle result corresponding to the evenly matched allocation rule is more reliable.

[0166] Optionally, the above method further includes: determining a third battle lineup from multiple candidate player virtual characters based on the preset evaluation results of the player virtual characters; conducting a battle ability test based on the third battle lineup to obtain a second battle result; and updating the preset evaluation results of the player virtual characters in the third battle lineup based on the second battle result to obtain the target evaluation results of the player virtual characters in both sides.

[0167] Optionally, based on the preset evaluation results of the player's virtual character, a third battle lineup is determined from multiple candidate player virtual characters, including: based on the preset evaluation results of the player's virtual character and a second allocation rule, a third battle lineup is determined from multiple candidate player virtual characters, wherein the second allocation rule is an allocation rule in which the abilities of the player virtual characters on both sides are comparable.

[0168] In the above embodiment, the preset evaluation results of the virtual characters of both players in the third battle lineup are updated to obtain the target evaluation results of the virtual characters of both players. This makes it easier to assign non-player virtual characters with similar abilities to player virtual characters in the target allocation of human-machine games. In this way, based on the results of the first battle, the target evaluation results of non-player virtual characters and player virtual characters can also be obtained accurately and reliably, thereby improving the accuracy of the ability evaluation of non-player virtual characters and player virtual characters.

[0169] In summary, the embodiments of this disclosure provide a method for testing and evaluating virtual characters. First, multiple non-player virtual characters are tested for their combat capabilities, achieving an automatic offline evaluation of the skill level of non-player virtual characters. Then, based on the initial evaluation results and preset evaluation results obtained from the offline evaluation, the combat capabilities of non-player virtual characters and player virtual characters are tested, achieving an online skill level evaluation. By automatically adopting a combination of offline and online evaluation for combat capability testing, accurate and objective evaluation of the skill levels of non-player virtual characters and player virtual characters can be achieved, while also reducing unnecessary waste of human resources.

[0170] Optionally, this disclosure also provides a program product, such as a computer-readable storage medium, including a program that, when executed by a processor, performs the above-described method embodiments.

[0171] For example, the method may include: testing the combat capabilities of multiple non-player virtual characters to obtain initial evaluation results of the non-player virtual characters; testing the combat capabilities of the non-player virtual characters against the player virtual characters based on the initial evaluation results of the non-player virtual characters and the preset evaluation results of the player virtual characters to obtain a first combat result; and updating the initial evaluation results and the preset evaluation results based on the first combat result to obtain a target evaluation result for the non-player virtual characters and a target evaluation result for the player virtual characters.

[0172] This method first tests the combat capabilities of multiple non-player virtual characters, achieving an automatic offline assessment of their skill levels. Then, based on the initial and preset assessment results obtained from the offline assessment, it tests the combat capabilities of non-player virtual characters against player virtual characters, achieving an online skill level assessment. By automatically combining offline and online assessments for combat capability testing, it can achieve an accurate and objective assessment of the skill levels of both non-player and player virtual characters, while also reducing unnecessary waste of human resources.

[0173] Optionally, multiple non-player virtual characters are tested for their combat abilities to obtain initial evaluation results. This includes: filtering multiple candidate non-player virtual characters based on preset character attribute parameters to obtain a first matchmaking queue; determining a first battle lineup from the first matchmaking queue according to preset battle rules, wherein the first battle lineup consists of non-player virtual characters from both sides, and the preset battle rules include: constraints on the competitive attribute parameters of the non-player virtual characters from both sides; and testing the combat abilities of the first battle lineup to obtain initial evaluation results for the non-player virtual characters.

[0174] In this embodiment, the skill levels of all non-player virtual characters can be initialized offline. A large number of games can be played with non-player virtual characters of different levels to collect performance data of each non-player virtual character in each game. Then, based on these performance data, statistical learning methods are used to continuously update the ability score of each non-player virtual character until the ability scores of all non-player virtual characters converge, thus obtaining the initial evaluation results of the non-player virtual characters. This facilitates the subsequent accurate and objective evaluation of the skill levels of non-player virtual characters and player virtual characters.

[0175] Optionally, based on the initial evaluation results of the non-player virtual characters and the preset evaluation results of the player virtual characters, a combat ability test is conducted between the non-player virtual characters and the player virtual characters to obtain a first combat result, including: determining a set of non-player virtual characters from multiple candidate non-player virtual characters; determining a lineup of player virtual characters from multiple candidate player virtual characters; determining a lineup of non-player virtual characters from the set of non-player virtual characters based on the allocation rules corresponding to the game state, the preset evaluation results of each player virtual character in the player virtual character lineup, and the initial evaluation results of each non-player virtual character in the set of non-player virtual characters; determining a second combat lineup based on the lineup of non-player virtual characters and the lineup of player virtual characters; and conducting a combat ability test based on the second combat lineup to obtain the first combat result.

[0176] In this embodiment, a battle ability test is conducted between humans and machines. Based on the first battle result, the initial evaluation result and the preset evaluation result are updated, allowing the ability scores of real players (player virtual characters) and robots (non-player virtual characters) to be updated simultaneously in the online battle environment, gradually achieving alignment between the ability scores of real players and robots. Furthermore, since online battles involve a mix of real players and robots, online alignment evaluation can correct any previous systemic discrepancies in the ability scores of real players and robots, thereby facilitating accurate and objective evaluation of the skill levels of both non-player virtual characters and player virtual characters.

[0177] Optionally, if the game status indicates that the player virtual character in the player virtual character lineup is the player virtual character corresponding to the novice player, or that the player virtual character in the player virtual character lineup is in a state of defeat in battle, then the allocation rule corresponding to the game status is the first allocation rule, which indicates that the player virtual character lineup is allocated a non-player virtual character lineup with combat ability lower than the player virtual character lineup; if the game status indicates that the player virtual character lineup has been waiting for a battle for a period of time greater than or equal to a preset time, then the allocation rule corresponding to the game status is the second allocation rule, which indicates that the player virtual character lineup is allocated a non-player virtual character lineup with combat ability equivalent to the player virtual character lineup.

[0178] In this embodiment, the first allocation rule is that the combat ability of the non-player virtual character lineup is lower than that of the player virtual character lineup, making it easier for the player virtual character lineup to win and improving the gaming experience for the corresponding user. The second allocation rule is an evenly matched allocation rule, ensuring that the combat abilities of the non-player virtual character lineup and the player virtual character lineup are comparable, thus creating a more balanced skill level among the participating factions.

[0179] Optionally, determining a set of non-player virtual characters from multiple candidate non-player virtual characters includes: filtering multiple candidate non-player virtual characters according to preset character attribute parameters to obtain a second matching queue; and determining a set of non-player virtual characters from the second matching queue according to preset battle rules, wherein the preset battle rules include: constraints on the competitive attribute parameters of the non-player virtual character lineup and the player virtual character lineup.

[0180] Optionally, determining the lineup of player virtual characters from multiple candidate player virtual characters includes: determining the lineup of player virtual characters from multiple candidate player virtual characters according to preset battle rules, wherein the preset battle rules include: constraints on the competitive attribute parameters of non-player virtual character lineups and player virtual character lineups.

[0181] Optionally, based on the first battle result, the initial evaluation result and the preset evaluation result are updated to obtain the target evaluation result for non-player virtual characters and the target evaluation result for player virtual characters. This includes: obtaining the target first battle result from the first battle result when the allocation rule is the second allocation rule, wherein the second allocation rule is an allocation rule in which the abilities of the non-player virtual character lineup and the player virtual character lineup are equivalent; and updating the initial evaluation result and the preset evaluation result based on the target first battle result to obtain the target evaluation result for non-player virtual characters and the target evaluation result for player virtual characters.

[0182] In this embodiment, the non-player virtual character lineup and player virtual character lineup participating in the battle using the evenly matched allocation rule (second allocation rule) are of comparable ability. Updating the initial evaluation result and the preset evaluation result based on the target first battle result corresponding to the evenly matched allocation rule is more reliable.

[0183] Optionally, the method further includes: determining a third battle lineup from multiple candidate player virtual characters based on the preset evaluation results of the player virtual characters; conducting a battle ability test based on the third battle lineup to obtain a second battle result; and updating the preset evaluation results of the player virtual characters in the third battle lineup based on the second battle result to obtain the target evaluation results of the player virtual characters in both sides.

[0184] Optionally, based on the preset evaluation results of the player's virtual character, a third battle lineup is determined from multiple candidate player virtual characters, including: based on the preset evaluation results of the player's virtual character and a second allocation rule, a third battle lineup is determined from multiple candidate player virtual characters, wherein the second allocation rule is an allocation rule in which the abilities of the player virtual characters on both sides are comparable.

[0185] In the above embodiment, the preset evaluation results of the virtual characters of both players in the third battle lineup are updated to obtain the target evaluation results of the virtual characters of both players. This makes it easier to assign non-player virtual characters with similar abilities to player virtual characters in the target allocation of human-machine games. In this way, based on the results of the first battle, the target evaluation results of non-player virtual characters and player virtual characters can also be obtained accurately and reliably, thereby improving the accuracy of the ability evaluation of non-player virtual characters and player virtual characters.

[0186] In summary, the embodiments of this disclosure provide a method for testing and evaluating virtual characters. First, multiple non-player virtual characters are tested for their combat capabilities, achieving an automatic offline evaluation of the skill level of non-player virtual characters. Then, based on the initial evaluation results and preset evaluation results obtained from the offline evaluation, the combat capabilities of non-player virtual characters and player virtual characters are tested, achieving an online skill level evaluation. By automatically adopting a combination of offline and online evaluation for combat capability testing, accurate and objective evaluation of the skill levels of non-player virtual characters and player virtual characters can be achieved, while also reducing unnecessary waste of human resources.

[0187] In the several embodiments provided in this disclosure, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0188] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0189] Furthermore, the functional units in the various embodiments of this disclosure can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in a combination of hardware and software functional units.

[0190] The integrated units implemented as software functional units described above can be stored in a computer-readable storage medium. These software functional units, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods of the various embodiments of this disclosure. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0191] The above are merely preferred embodiments of this disclosure and are not intended to limit this disclosure. Various modifications and variations can be made to this disclosure by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A method for testing and evaluating virtual roles, the method comprising: testing the fighting capabilities of a plurality of non-player virtual roles to obtain initial evaluation results of the non-player virtual roles; testing the fighting capabilities of the non-player virtual roles and a player virtual role according to the initial evaluation results of the non-player virtual roles and preset evaluation results of the player virtual role to obtain a first fighting result; updating the initial evaluation results and the preset evaluation results according to the first fighting result to obtain target evaluation results of the non-player virtual roles and target evaluation results of the player virtual role.

2. The method of claim 1, wherein, The testing of the fighting capabilities of a plurality of non-player virtual roles to obtain initial evaluation results of the non-player virtual roles comprises: screening a plurality of candidate non-player virtual roles according to preset role attribute parameters to obtain a first matching queue; determining a first fighting lineup from the first matching queue according to preset fighting rules, wherein the first fighting lineup is composed of non-player virtual roles of both fighting sides, and the preset fighting rules include constraint conditions for competitive attribute parameters of non-player virtual roles of both fighting sides; testing the fighting capabilities based on the first fighting lineup to obtain the initial evaluation results of the non-player virtual roles.

3. The method of claim 1, wherein, The testing of the fighting capabilities of the non-player virtual roles and a player virtual role according to the initial evaluation results of the non-player virtual roles and preset evaluation results of the player virtual role to obtain a first fighting result comprises: determining a set of non-player virtual roles from a plurality of candidate non-player virtual roles; determining a player virtual role lineup from a plurality of candidate player virtual roles; determining a non-player virtual role lineup from the set of non-player virtual roles according to an allocation rule corresponding to a game state, the preset evaluation results of the player virtual roles in the player virtual role lineup, and the initial evaluation results of the non-player virtual roles in the set of non-player virtual roles; determining a second fighting lineup according to the non-player virtual role lineup and the player virtual role lineup; testing the fighting capabilities based on the second fighting lineup to obtain the first fighting result.

4. The method of claim 3, wherein, If the game state indicates that a player virtual role in the player virtual role lineup is a player virtual role corresponding to a novice player or a player virtual role in the player virtual role lineup is in a fighting failure state, an allocation rule corresponding to the game state is a first allocation rule, and the first allocation rule indicates that the player virtual role lineup is allocated the non-player virtual role lineup with fighting capabilities lower than those of the player virtual role lineup. If the game state indicates that a length of time for which the player virtual role lineup waits for fighting is greater than or equal to a preset length of time, an allocation rule corresponding to the game state is a second allocation rule, and the second allocation rule indicates that the player virtual role lineup is allocated the non-player virtual role lineup with fighting capabilities comparable to those of the player virtual role lineup.

5. The method of claim 3, wherein, The determination of a set of non-player virtual roles from a plurality of candidate non-player virtual roles comprises: According to a preset role attribute parameter, the plurality of candidate non-player virtual roles are screened to obtain a second matching queue; According to a preset battle rule, the non-player virtual role set is determined from the second matching queue, wherein the preset battle rule includes a constraint condition of competitive attribute parameters of the non-player virtual role lineup and the player virtual role lineup.

6. The method of claim 3, wherein, The player virtual role lineup is determined from the plurality of candidate player virtual roles, including: According to a preset battle rule, the player virtual role lineup is determined from the plurality of candidate player virtual roles, wherein the preset battle rule includes a constraint condition of competitive attribute parameters of the non-player virtual role lineup and the player virtual role lineup.

7. The method of claim 3, wherein, The initial evaluation result and the preset evaluation result are updated according to the first battle result to obtain a target evaluation result of the non-player virtual role and a target evaluation result of the player virtual role, including: From the first battle result, a target first battle result is obtained when the distribution rule is a second distribution rule, wherein the second distribution rule is a distribution rule in which the non-player virtual role lineup and the player virtual role lineup have equivalent abilities; According to the target first battle result, the initial evaluation result and the preset evaluation result are updated to obtain a target evaluation result of the non-player virtual role and a target evaluation result of the player virtual role.

8. The method of claim 1, wherein, The method further includes: According to the preset evaluation result of the player virtual role, a third battle lineup is determined from a plurality of candidate player virtual roles; A battle ability test is performed based on the third battle lineup to obtain a second battle result; According to the second battle result, the preset evaluation result of the player virtual roles in the third battle lineup is updated to obtain a target evaluation result of the player virtual roles.

9. The method of claim 8, wherein, According to the preset evaluation result of the player virtual role, a third battle lineup is determined from a plurality of candidate player virtual roles, including: According to the preset evaluation result of the player virtual role and a second distribution rule, the third battle lineup is determined from the plurality of candidate player virtual roles, wherein the second distribution rule is a distribution rule in which the player virtual roles have equivalent abilities.

10. A virtual role test evaluation device, the device comprising: A test module configured to perform a battle ability test on a plurality of non-player virtual roles to obtain an initial evaluation result of the non-player virtual roles; According to the initial evaluation result of the non-player virtual roles and a preset evaluation result of player virtual roles, a battle ability test is performed on the non-player virtual roles and the player virtual roles to obtain a first battle result; An update module configured to update the initial evaluation result and the preset evaluation result according to the first battle result to obtain a target evaluation result of the non-player virtual roles and a target evaluation result of the player virtual roles.

11. A server comprising: A memory and a processor, the memory storing a computer program executable by the processor, the processor executing the computer program to implement the test evaluation method of the virtual role according to any one of claims 1-9. 12.A computer readable storage medium, the storage medium storing a computer program, the computer program being read and executed to implement the test evaluation method of the virtual role according to any one of claims 1-9.

Images