KR20060020004A - Game difficulty control method for user interacting using artificial intelligence technique - Google Patents

Abstract

TECHNICAL FIELD The present invention relates to electronic information technology, and more particularly, to the field of game development, and more particularly, to a method for controlling user interaction game difficulty using artificial intelligence. An object of the present invention is to provide a user interactive game difficulty control method that can effectively adjust the difficulty of the game according to the user's game play proficiency. In the present invention, by using the genetic algorithm (Genetic Algorithm) of the artificial intelligence technique to determine the user's propensity and adaptation to the game progress by adjusting the difficulty of the game to match the user's level in real time accordingly In addition, the game is developed in a user-friendly manner to provide easy difficulty for beginners and more difficult for experienced users. In this case, it is possible to increase the user's interest and extend the life cycle of the game.

Game, Difficulty Control, Artificial Intelligence, User Interaction, Genetic Algorithm

Description

GAME DIFFICULTY CONTROL METHOD FOR USER INTERACTING USING ARTIFICIAL INTELLIGENCE TECHNIQUE}             

1 is a flow chart illustrating a process of adjusting a user difficulty game difficulty using an artificial intelligence technique according to an embodiment of the present invention.

2 is a diagram illustrating the order of pieces generated in the Tetris game.

3 illustrates a Tetris gameboard.

4 is a graph of B / T (number of free blocks per piece occurrence) according to goodness of fit.

Figure 5a is a graph showing the results for the number of lines removed by the user in the experiment for the random users to apply the present invention.

5B is a graph showing the total number of pieces generated while a user plays in an experiment with a random user applying the present invention.

Figure 6a is a graph showing the number of lines removed by the user in the experiments for the novice group and the skilled group applying the present invention.

FIG. 6B is a graph showing the total number of pieces generated while a user plays in an experiment involving a novice group and a skilled group applying the present invention. FIG.

7 is a graph showing the total number of pieces generated while a user plays in an experiment in which the mutation rate is changed by applying the present invention.

TECHNICAL FIELD The present invention relates to electronic information technology, and more particularly, to the field of game development, and more particularly, to a method for controlling user interaction game difficulty using artificial intelligence.

Recently, the game industry, which has been in the spotlight due to the rapid growth in size, is focused on research for improving high definition graphics and networking. In the case of high-definition graphics, the processing power or performance has been greatly improved, and now it is seen that the limit of technological development is being reached. Therefore, attention is focused on the networking of the game, which is another important factor in the game.

Currently, most young adults are quite familiar with games, and the life cycle of games is getting shorter and shorter. This is usually the more difficult the user wants to be when they are used to the game, but the game provides users who are more familiar with the game by providing difficulty that is uniformly adjusted according to saved difficulty scenarios without considering the user's game skill. It is caused by halting the fun of the game.

Artificial intelligence techniques are applied in the game development stage to compensate for the reduction of game life cycle. However, most of the AI techniques used in games have been used to set more realistic situations, to make the game more realistic, or to avoid simple repetition, and AI techniques are used for game play. It does not directly contribute to controlling the difficulty of the game because it is not directly involved.

Although games should be easier to reach the public and should be developed in a way that can be enjoyed by more people, the difficulty of adjusting the difficulty of the existing game is only for users to understand, so they are not familiar with the game. A user who has been left out of the game or who has become accustomed to the game soon has a positive result. In addition, in terms of game developers, since many scenarios must be developed according to the user's skill level in consideration of the user's game skill level, the difficulty of investing a lot of time and money in difficulty control was entailed.

The present invention has been proposed to solve the above problems of the prior art, and an object of the present invention is to provide a user interaction game difficulty control method that can effectively control the difficulty of the game according to the user's game play proficiency.

According to an aspect of the present invention for achieving the above technical problem, the step of determining the adaptability to the game of the user proceeding the game by using a genetic algorithm, and until the game end request, the determined of the user There is provided a user interactive game difficulty adjustment method comprising the step of applying a game adaptability to the difficulty of the game to proceed with the game.

Preferably, determining the adaptability of the user to the game, the step of loading the initial gene for the initial game difficulty setting; Constructing an initialization generation that is a collection of initial game difficulty control elements using the loaded initial gene; Applying the configured generation to proceed with the game; And determining the fitness of the user according to the fitness function while the game is in progress.

Preferably, the step of applying the game adaptability of the user to the difficulty of the game to proceed with the game, checking whether the difficulty adjustment is necessary in response to the result of the determination of the adaptability; Forming a new generation by applying an operator as a result of adjusting the degree of difficulty; And loading a new gene for controlling the difficulty of the game according to the new generation.

Preferably, the step of proceeding the game by applying the game adaptability of the user to the difficulty of the game, the step of checking whether there is a game end request, and as a result of the check, by applying the configured generation according to the absence of the game end request The method further includes returning to the step of proceeding with the game.

Further, the step of applying the game adaptability of the user to the difficulty of the game to play the game, the game is applied by using the existing generation until the end of the game request as the difficulty adjustment is not necessary, as a result of the determination of the adaptability Further comprising the step of proceeding.

On the other hand, according to another aspect of the invention, the system having a processor for executing a game, the function of loading the initial gene for setting the initial game difficulty; A function of constructing an initialization generation which is a set of initial game difficulty control elements using the loaded initial gene; Applying a configured generation to play the game; Determining the fitness of the user according to the fitness function while playing the game; Checking whether difficulty adjustment is necessary in response to the result of the fitness determination; As a result of the adaptability determination, a function of configuring a new generation by applying an operator as a difficulty adjustment is needed; And a computer-readable recording medium having recorded thereon a program for realizing a function of loading a new gene for controlling the difficulty of a game according to the new generation.

In the present invention, by using the genetic algorithm (Genetic Algorithm) of the artificial intelligence technique to determine the user's propensity and adaptation to the game progress by adjusting the difficulty of the game to match the user's level in real time accordingly In addition, the game is developed in a user-friendly manner to provide easy difficulty for beginners and more difficult for experienced users. In this case, it is possible to increase the user's interest and extend the life cycle of the game.

Hereinafter, preferred embodiments of the present invention will be introduced in order to enable those skilled in the art to more easily implement the present invention.

1 is a flowchart illustrating a process of adjusting a user interaction game difficulty using an artificial intelligence method according to an embodiment of the present invention.

Referring to FIG. 1, the process of adjusting a user difficulty game difficulty using an artificial intelligence method according to the present embodiment includes a process (A) of determining an adaptability to a game of a user who proceeds a game using a genetic algorithm. According to the determination result, the process of adjusting the difficulty of the game to match the level of the user in real time (B).

Hereinafter, the game difficulty adjustment process according to the present embodiment including the game adaptation determination process (A) and the difficulty adjustment process (B) will be described in more detail.

First, an initial gene for setting a difficulty level of a game is initially loaded (S10), and an initial generation that is a set of initial game difficulty control elements is configured using the loaded initial gene (S11).

Subsequently, the game proceeds by applying the configured initialization generation (S12). As described above, the adaptability of the user is determined according to the fitness function in the game (S13).

Subsequently, it is checked whether difficulty adjustment is necessary in response to the adaptation determination result (S14).

If, as a result of adaptation determination, difficulty adjustment is necessary, a new generation is configured by applying an operator (S15), and a new gene for loading difficulty adjustment is loaded according to the new generation (S16).

Thereafter, the game proceeds while repeatedly performing the game progress step (S12) to the new gene loading step (S16) until the game end request is made. That is, it is checked whether there is a game end request (S17), and if there is no game end request, the process returns to the game progress step S12, and if there is a game end request, the game ends (S18).

On the other hand, as a result of the adaptation determination (S14), when difficulty control is not necessary, a new generation configuration step (S15) and a new gene loading step (S16) is skipped.

The genetic algorithm applied to the present invention has a fitness that can be represented as an environment to which one individual (Individual) wants to adapt, that is, a problem to be solved in the inheritance of nature, It is based on Gene, which is a candidate for year, and Generation, which is a collection of candidates for that year.

And, as operators of genetic algorithms, well-adapted solutions survive and poorly-adapted solutions can be eliminated, improved by either a parent selected with high fit, or by a given mother, which is likely to be inherited. Reproductions produce a new generation of genes, crossovers to select random positions during stinging, and random mutations.

If you apply the genetic algorithm to adjust the difficulty of the game, the genetic algorithm uses the difficulty control elements of the game as a gene, such as increasing the number of enemies in the game or increasing the speed of the game. The user adaptability evaluation factor of is used as the fitness function.

In general, genes are used to determine which factors control overall difficulty at the beginning of game production, and the fitness function is mainly determined by reflecting the intention of the creator or producer of the game.

As described above, the gene and the fitness function according to the present invention can be applied to most genres such as flight shooting games, action games, role-playing games, puzzle games, and the like. As follows.

First, in the flying shooting game, the factor controlling the difficulty level of the game is to increase the speed of the enemy, the speed of the game, or increase the number of enemies. Therefore, in the case of the gallag, a typical flight shooting game, the genetic factor is C = [speed , Missile], and the gene is the change of the genetic factor per 10 units of time, represented by G (t) = {C0, C1, C2, ..., C9}, and the fitness function is the amount of change in the score obtained by the player over time This results in a change in enemy population per unit time (density of the enemy).

Next, in the case of Bank Panic, a typical action game, the visitor's visit order is determined during the game, so if the user is smart and repeats the game several times, the visitor's visit order is memorized. The genetic factor of the action game is the character (strength or guest) appearing in the three doors, that is, C = [Pl, Pc, Pr], and the gene becomes the change of the genetic factor per 10 units of time. (t) = {C0, C1, C2, ..., C9}, and the fitness function is a change in score obtained by a user for a predetermined time or a ratio of intensity among characters visited.

On the other hand, Diablo, one of the representative role-playing games, appears at a certain number of enemies on a certain stage. If there is a limited number of enemies, the difficulty of the game may vary depending on how close they are, that is, the cohesion of the enemies. Therefore, in role-playing games, there are no genetic factors, and the gene is the average distance between the enemies, that is, the cohesion of the enemies, and the fitness function is the variation in the experience (score) acquired by the player per unit time or the variation in the number of enemies per unit time. .

Finally, in the case of Tetris, the oldest puzzle game in the game's history, each person has a different level of difficulty and ease depending on the sequence of pieces that appear in the game. The gene is in the order in which each fragment occurs, and the fitness function uses the density of the entire blank block, taking into account the clean line.

 Hereinafter, an example in which the present invention is applied to Tetris, which is considered to be the most apparent game among the genres, will be described in more detail.

Tetris is a game in which a score is obtained by combining pieces of each shape generated through a random number function to generate more lines. The difficulty of the game varies according to the order of the generated pieces. The underlying side of the plate must be flat to form a permutation of flat pieces.

Here, the degree of difficulty of permutation of the pieces is quite variable and entirely depends on the situation and the user, even the same user is very different depending on the level of skill, the present invention uses the permutation of the fragments generated by the genetic factor.

In the present invention, as an element to evaluate the adaptability of the user to the game, the density of the empty blocks up to the last line including the already removed line is used as the adaptability of the user.

In addition, as shown in Figure 3, the gene encodes a sequence of four fragments to be generated and its adaptability, one generation consists of eight encoded genes, and the selection method is close to the desired adaptability value of the developer. Two genes of goodness-of-fit values are selected.

Therefore, when the game starts, the module driven to execute the game initially loads the initial gene for setting the difficulty of the game (S10), and initializes generation, which is a set of initial game difficulty control elements using the loaded initial gene. After configuring (S11), the game proceeds by applying the configured initialization generation (S12).

In general, the Tetris game generates an integer from 0 to 6 through a random function to generate a corresponding fragment, as shown in FIG. 3, and the fitness function is represented by the following equation, which represents a constant value.

Where 'B' is the number of blank blocks shown on the screen, 'C' is the number of removed lines already removed by forming a line, and 'L' is the number of final lines of pieces currently stacked on the game board. .

In FIG. 3, blocks and pieces become a basic unit constituting the game board of Tetris, and the entire game board is composed of 12 * 23 blocks. When the number of occurrences of a piece is 'T', it can be seen that the relationship shown in Equation 2 below is established using Equation 1 above.

 Here, 'F' represents a goodness-of-fit function that is a constant.

4 is a graph of B / T (number of empty blocks per piece generation) according to the goodness of fit, and since the goodness-of-fit function is fixed, B / T converges to one constant (C, Constant).

Accordingly, it can be seen that the Tetris game represented by the graph proceeds in a direction of keeping the number of empty blocks per piece generation allowance constant, and the game is based on the number of blocks generated during the game by the user. The number of blocks is kept constant.

In the present invention, a fragment is generated by implementing a function 'GA (I)' to replace 'rand ()', which is a random number generation function used in a conventional Tetris game, and generates an input parameter of the function 'GA (I)'. Are the last lines, the clean lines, and the blank blocks, which are used to find the goodness of fit and to determine the user level and output the corresponding pieces.

Accordingly, the module driven to execute the game while the game is in progress adapts the user to whether it is equal to or less than a predetermined reference density based on the density of empty blocks up to the last line including the line already removed according to the fitness function. If the difficulty is necessary, the operator may select a selection, cross, mutation, etc. to maintain a constant number of empty blocks in the game board with respect to the number of blocks generated during the game. To form a new generation. In addition, according to a new generation, the game proceeds by loading and applying a new gene for controlling the difficulty of the game until the end of the game is required.

Then, the module driven to progress the game implements the game by using a difficulty control element which is a gene suitable for the user's level, wherein the difficulty control element is inserted into the logic part of the game to continuously maintain the user during the game. Since the state of the game and the difficulty of the game is adjusted, it is also applicable to pre-made games.

As shown in FIG. 1, in the game to which the genetic algorithm is applied, a separate game end condition does not exist until the user requests the end of the game.

Hereinafter, a result of performing a simulation on each user having a different game skill level for the game to which the present invention is applied will be described.

The simulation is divided into three methods, the first is the experiment for random users and the result of comparing the original game and the game to which the present invention is applied, and the second is the experiment comparing the difference between the results by dividing the user into a novice group and a skilled group. Finally, the experiment was carried out by changing the mutation rate among factors that influence the performance of the genetic algorithm of the present invention.

First, in the case of experiments involving random users, the original game and the game to which the present invention was applied were directly conducted for eight random persons. The game is designed to keep empty blocks at a ratio of 3.8% and 5.7% of the numbers.

Figure 5a is a graph showing the results of the number of lines removed by the user in the experiment for the random user applying the present invention, Figure 5b is a user play in the experiment for the random user applying the present invention This is a graph showing the total number of fragments that occurred during the process.

In the drawing, the left side of the bar graph represents the original game, and the middle and the right side represent the game results for games with 3.8% and 5.7% respectively, with the number of lines removed by the user or the total number of pieces generated during the user's play. Can be used as a measure of the level of the user.

Comparing the averages in each of the graphs of FIGS. 5A and 5B, the results are lower than those of the original game, and it can be seen that the user is more difficult when the game is applied with a suitability of 5.7% rather than 3.8%.

In addition, if you look at the variance on the right side of the two graphs, it can be seen that the variance between users gradually decreases in the game to which the genetic algorithm is applied, which means that the skill gap between users is somewhat reduced.

In addition, it can be seen that the user's level is converged to a specific difficulty level through the result of the dispersion, and the result proceeds in the direction of gradually decreasing by the goodness of fit.

Figure 6a is a graph showing the results of the number of lines removed by the user in the experiment for the novice group and the skilled group applying the present invention, Figure 6b is applied to the novice group and skilled group applying the present invention In the experiment, it is a graph showing the total number of pieces that occurred while the user played.

Here, for the original game, 'users A, B, and C' are users who can be judged as novices with less than 35 clean lines and less than 150 total pieces. , E, F, G 'are skilled users with more than 55 removal lines and more than 200 total pieces.

Looking at the average value of the total pieces in the two graphs, it can be seen that the results of the novice group tend to increase gradually in the game to which the present invention is applied, and the skilled group tend to decrease gradually. As a result, in the results of the game having a goodness of fit of 5.7%, it can be seen that the results of the novice group and the skilled group show almost similar values.

Here, each graph is drawn from two perspectives, the removal line and the total piece, but the result of the removal line is more reliable than the total piece. This is because the elimination line is directly related to the score of the game and the total number of pieces is related to the play time.

Finally, the case of an experiment in which the mutation rate is changed among the performance left and right factors of the genetic algorithm is described.

7 is a graph showing the total number of pieces generated while the user plays in an experiment in which the mutation rate is changed by applying the present invention.

The genetic algorithm of the present invention has no termination condition due to the characteristics of the game, and is distinguished from the general genetic algorithm independently of the hybridization and mutation operations.

In this experiment, mutations were generated at 5%, 10%, and 15% of the total 6 users, respectively, and as shown in FIG. This is low, confirming the findings of other papers that a good fit of 10% is the best mutation rate.

Although the technical idea of the present invention has been described in detail according to the above preferred embodiment, it should be noted that the above-described embodiment is for the purpose of description and not of limitation. In addition, those skilled in the art will understand that various embodiments are possible within the scope of the technical idea of the present invention.

For example, the above-described embodiment has been described using Tetris, which is a representative puzzle game, as an example. However, the present invention also applies to games of other genres such as flight shooting games, action games, and role-playing games.

The present invention described above uses a genetic algorithm, which is one of artificial intelligence techniques, to adjust a game's difficulty level to a user's level in real time according to a user's propensity and adaptability to a game, thereby making the game user-friendly. By developing this, easy difficulty for beginners, more difficult difficulty for experienced users, and thus increase the interest of the game for the user, thereby increasing the life cycle of the game.

Claims (6)

Determining the adaptability of the user to the game using the genetic algorithm, Proceeding to the game by applying the determined game adaptability of the user to the difficulty of the game until a game end request is made; User interaction game difficulty adjustment method comprising a. The method of claim 1, The determining of the adaptability to the game of the user, Loading an initial gene for initial game difficulty setting; Constructing an initialization generation that is a collection of initial game difficulty control elements using the loaded initial gene; Applying the configured generation to proceed with the game; And Determining the degree of adaptability of the user according to the fitness function as the game progresses. The method of claim 2, The process of applying the game adaptability of the user to the difficulty of the game proceeds, Checking whether difficulty adjustment is necessary in response to the result of the fitness determination; Forming a new generation by applying an operator as a result of adjusting the degree of difficulty; And And loading a new gene for controlling a difficulty level of the game according to the new generation. The method of claim 3, The process of applying the game adaptability of the user to the difficulty of the game proceeds, Checking if there is a game end request, The method of claim 1, further comprising the step of returning to the step of proceeding the game by applying the configured generation in accordance with the request for no end of the game. The method of claim 3, The step of applying the game adaptability of the user to the difficulty of the game to proceed with the game, As a result of the determination of the degree of adaptation, the difficulty adjustment is further required, and further comprising the step of proceeding the game by using the existing generation until the end of the game request. In a system having a processor for running a game, Loading initial genes for initial game difficulty setting; A function of constructing an initialization generation which is a set of initial game difficulty control elements using the loaded initial gene; Applying a configured generation to play the game; Determining the fitness of the user according to the fitness function while playing the game; Checking whether difficulty adjustment is necessary in response to the result of the fitness determination; As a result of the adaptability determination, a function of configuring a new generation by applying an operator as a difficulty adjustment is needed; And Loading a new gene for controlling the difficulty of the game according to the new generation A computer-readable recording medium having recorded thereon a program for realizing this.

Images