CN111185015A

CN111185015A - Method for optimizing ten-player online competitive game matching mechanism

Info

Publication number: CN111185015A
Application number: CN201911304653.8A
Authority: CN
Inventors: 万国春; 米健
Original assignee: Tongji University
Current assignee: Tongji University
Priority date: 2019-12-17
Filing date: 2019-12-17
Publication date: 2020-05-22
Anticipated expiration: 2039-12-17
Also published as: CN111185015B

Abstract

The invention provides a method for optimizing the matching mechanism of a ten-player online competitive game, and relates to the optimization technology of the game matching mechanism. This method collects the last 30 game information of each player who has entered between 5 seconds and 60 seconds of matching time, and uses 1‑D convolution, maximum function, random sampling, function mapping, arithmetic matching and other algorithms to make one player's game The average level is in a set interval, and the probability of conflict between one player's career choice is minimized. Using the method of the present invention, on the one hand, the game experience of the player is increased, the limited game time of the player is fully utilized, and the purpose of the game serving entertainment is realized. On the other hand, the application of the present invention promotes the development and expansion of the ten-player online competitive game in the game market, which has certain promotion significance for the game industry, and also has positive significance for the harmonious development of human society at a deeper level.

Description

Method for optimizing ten-player online competitive game matching mechanism

Technical Field

The invention relates to the field of game matching mechanism optimization.

Technical Field

With the acceleration of the rhythm of life, the fast-paced online competitive game becomes a part of the daily life of people, particularly teenagers, and the proper game is beneficial to physical and psychological health. However, the current matching mechanism is imperfect, and the problems that the game levels of the players of the two parties are often in a large gap and the players compete for occupation of the characters exist, so that the countercheck experience is poor, the purpose of the game for serving life is deviated, and the advance of the game industry is hindered.

At present, the online competitive game has good development prospect, and the crowd base number must be increased rapidly along with the time. Therefore, how to optimize the matching mechanism, solve the phenomenon that the difference between the strength of the players of the two parties is large, reduce the probability of professional selection conflict of the player characters, and improve the game experience of the players becomes a problem which needs to be solved urgently by game developers.

Disclosure of Invention

The invention aims to provide an algorithm method for optimizing a matching mechanism of a ten-player online competitive game, which can effectively improve the confrontation experience of players, increase the interestingness of the competitive game and promote the development of the game industry.

In order to achieve the above purposes, the invention adopts the technical scheme that:

an algorithmic method for optimizing a ten player online competitive game matching mechanism, comprising the steps of:

1) collection of player gaming information

Collecting the game information of the nearest N fields of each player entering at a certain time, wherein the game information refers to two aspects, namely the score of the nearest N fields of games and the role occupation selection condition of the nearest N fields of games; supplied to step 2);

2) scored 1-D convolution

Carrying out convolution operation on the N field scores score of each player by using 1 row, N columns and 1-D templates respectively, wherein the coefficients of the templates correspond to the N field scores score of each player one by one, so as to obtain convolution operation results equal to the number of matched people in the step 1), and recording the convolution results as G; after all players perform the same operation, each matched player in the step 1) has a G value; provided to step 3), step 6);

3) section setting of score

Randomly sampling all matched players in the step 2) to obtain G values of ten players, namely extracting ten G values, carrying out 1-D convolution on the ten G values by utilizing the template in the figure 4, recording the convolution result as A, and setting A +/-20% as a scoring interval; supplied to step 6);

4) function mapping for role occupations

Respectively executing the same operation on all the players matched in the step 1), performing function mapping on the selection of N game role occupations of each player in the step 1), wherein the mapping function is in the form of t ═ E (role occupations), wherein E is a functional relation, t is a mapping result, t of each occupation in the five types of occupations has uniqueness, counting the same t, and obtaining the number d corresponding to each t, namely a table of fig. 5t-d, wherein all the players matched in the step 1) have a respective table of fig. 5 t-d; supplied to step 5);

5) generation of player combinations

Obtaining the maximum value d of the d (d1, d2, d3, d4, d5) value of each player in the step 4) through a maximum value function_maxAnd t is one of 000, 001, 010, 011 and 100 and is marked as t_dmax(ii) a Arithmetic matching five different t_dmaxTraversing all matched players in the step 1) for one player combination to generate a player combination of one group of five persons, namely obtaining a primary player combination, and providing the primary player combination for the step 6) for screening;

6) final player combination derivation

Recording the G value of each group of players in the preliminary player combination obtained in the step 5) as G1, G2, G3, G4 and G5, and calculating G_ave(G1+ G2+ G3+ G4+ G5)/5, if G_aveIn step 3), the interval is set, and the set of player combinations is the final player combination.

In step 5), the arithmetic matching algorithm realizes five different t_dmaxBy the following steps:

step 5.2.1, matching t of all players matched in 1)_dmaxForming a queue, the length of the queue is consistent with the number of players, setting a mark p to point to t of the first player in the queue_dmax。

Step 5.2.2, take out t of the first player by p_dmaxT of the first player, denoted t1, reserved t1_dmaxDisappear from the queue, then p moves one bit backwards, pointing to t of the second player_dmax。

Step 5.2.3, the second name is taken out by reusing pT of player_dmaxIf t1 is equal to t of the second player_dmaxOtherwise, t of the second player is reserved_dmaxT2 and t1 form a combination, denoted t2, of the second player_dmaxAlso disappear from the queue; otherwise, t of the second player is set_dmaxThe home position of the queue is retired. P is shifted back by one bit.

And 5.2.4, repeating the step 5.2.3 to obtain t3, t4 and t5, wherein 5 players corresponding to t1, t2, t3, t4 and t5 are a group of player combinations.

Step 5.2.4-1 if p points to t of the last player in the queue_dmaxI.e. the first traversal is finished, it is found that a group of player combinations cannot be formed, then all the processes of the present invention are executed again by the player who has been taken out, the remaining players in the queue and the next player who enters the matching, and step 5.2.5, step 5.2.6 and step 6 are not executed); if at least one set of player combinations is generated after the first traversal, then steps 5.2.5, 5.2.6, and 6) continue.

Step 5.2.5, now again point p move to t of the first player of the remaining players in the queue_dmaxAnd repeating the steps 5.2.1 to 5.2.4 to achieve the aim of traversing all players.

Step 5.2.6, when p points to t of the last player in the queue_dmaxWhen it is found that a group of t1, t2, t3, t4, t5 combinations is no longer available, all the processes of the present invention are executed again for the player who has taken out, the players who remain in the queue, the player combinations which do not satisfy the scoring interval in step 3) in step 6), and the next group of players who enter the matching.

The invention can effectively reduce the difference of the average game level of the players of the two parties, reduce the probability of conflict of professional choices of the role of the player of one party and improve the game experience of the players of the two parties. The significance of the invention is as follows: on one hand, for the individual player, the body and mind are more fully relaxed in a limited time, the game experience of the player is increased, the industrial purpose of the game for serving life is met, and the construction pace of the harmonious society is accelerated to a certain extent. On the other hand, the crowd base of the multiplayer online competitive game is necessarily and continuously expanded along with the time. The use of the invention promotes the development and growth of the games in the game market and indirectly promotes the development of the whole game industry.

Drawings

FIG. 1 is a schematic flow chart of the operation of the present invention

FIG. 2 is a 1-D template coefficient plot for row 1 and column 30 of an embodiment

FIG. 3 is a schematic diagram of a 1-row 30-column 1-D template of an embodiment

FIG. 4 is a 1-row 10-column filter template of an embodiment

FIG. 5 is a schematic diagram of t-d of an embodiment

FIG. 6 is a schematic flow chart of step 5) of the present invention

Detailed Description

Examples

1) collection of player gaming information

Collecting game information of a plurality of N fields (in the embodiment, 30 fields are set for N) which enter each player between certain matching time (in the embodiment, 5S-60S), wherein the game information refers to two aspects, namely scoring score of the game of the N fields (in the embodiment, 30 fields) and character occupation selection condition of the game of the N fields (in the embodiment, 30 fields); supplied to step 2);

2) scored 1-D convolution

Carrying out convolution operation on 30 field scores score of each player by using 1 row and 30 columns of 1-D templates in the figure 3, wherein the coefficients of the templates correspond to the 30 field scores score of each player one by one, so as to obtain a convolution operation result equal to the number of matched people in the step 1), and recording the convolution result as G; after all players perform the same operation, each matched player in the step 1) has a G value; provided to step 3), step 6);

3) section setting of score

4) function mapping for role occupations

Respectively executing the same operation on all the players matched in the step 1), performing function mapping on the selection of thirty game role occupations of each player in the step 1), wherein the mapping function is in the form of t ═ E (role occupations), wherein E is a functional relation, t is a mapping result, t of each of the five occupations has uniqueness, counting the same t, and obtaining the number d corresponding to each t, namely a table of fig. 5t-d, wherein all the players matched in the step 1) have a respective table of fig. 5 t-d; supplied to step 5);

5) generation of player combinations

6) final player combination derivation

Each step is described in detail below

In step 2), performing convolution operation of scores of each player: the size of the template of fig. 3 should be consistent with the game field, i.e., 30. In addition, according to the characteristic that the correlation degree between the game field closer in time and the current game level is larger, the one-to-one correspondence between the coefficient of the template in fig. 3 and the 30-field score of each player in the step 1) refers to the one-to-one correspondence between the coefficient from large to small of the template in fig. 3 and the 30-field score of each player from near to far in time.

The specific method comprises the following steps:

the value of G resulting from the template convolution of FIG. 2 should reflect the current game level of each matching player of step 1). It is logical that the score correlation degree is higher when the current game level is closer to the time point, and the score correlation degree is lower when the current game level is farther from the time point. To implement the above logic, the 30-field score of each player in step 1) is recorded as 1 st field, 2 nd field, n.n.n.30 th field according to the sequence of time points from far to near, and corresponds to the template coefficients f (1), f (2).. f (n)... f (30) in fig. 3, and the value of f (n) is determined by the following formula:

f(n)＝(1/2)f(n-1)+(1/4)f(n-2)+(1/8)f(n-3)+......+(1/2^n-1) f (1) +1(f (1) ═ 1, n ═ 2,3,4.. 30), the values of the template coefficients in fig. 2, i.e. in fig. 3, are finally determined.

score (n) represents the score of the nth field, then for step 1) each matched player G ═ (1/(f (1) + f (2) +. ·. · f (30))) (f (1) × score (1) + f (2) × score (2) +. ·.

Step 3) significance:

the convolution result a in step 3) represents the average level of all matched players in step 1), and to achieve this, the following steps can be performed:

step 3.1, recording step 3) the G values of the ten players are G0, G1,. to.. G9, and the coefficients of the templates of fig. 4 are a0, a1 and a2.. to.. a9, respectively;

3.2, G0 and A0, G1 and A1.... G9 and A9 correspond to each other one by one;

step 3.3, normalizing the template coefficients of FIG. 4: g0+ G1+, + G9,

determining the template coefficients of fig. 4 by calculating a0 ═ G0/T, a1 ═ G1/T.. a9 ═ G9/T;

step 3.4, after which a is determined by: a × G0 × a0+ G1 × a1+. a9 × a 9.

Step 4):

the character professions can be divided into five types in total, which is consistent with the number of players on one side of a ten-player online competitive game. The embodiment is provided for ten-player online competitive games, players are divided into two camps, each of the camps is five, each player has a game task, and the game tasks correspond to five different professions.

In step 4), t is coded in a three-bit binary manner, and five different t respectively correspond to 000, 001, 010, 011 and 100 and correspond to d1, d2, d3, d4 and d 5.

t takes 000001010011100 five three-digit binary numbers to encode, and the five binary numbers correspond to five roles, so the corresponding relation E of the function independent variable and the dependent variable is single-to-single.

All players matched in the step 1) are respectively provided with a t-d table matched with the players per se shown in fig. 5, the use frequency of each occupation of the players in the last thirty times is reflected exactly, the larger the value of d is, the more favored the players are in the corresponding occupation, and the smaller the value of d is, the more contradictory the players are in the corresponding occupation.

Step 5):

the arithmetic matching is characterized in that all t consistent with the number of matched players in the step 1) can be matched_dmaxIn matching five different t_dmaxCombined as a group of players and up to the point where the remaining players fail to satisfy five different t' s_dmaxThis condition is up to.

Suppose now that step 1) matches 101 players, each of them d_maxCorrespond to respective t_dmax(d_maxThe corresponding t is denoted as t_dmax) That is, there are 101 t in total_dmax，

At this time at 101 t_dmaxIn, there may not even be a way to find five different t_dmaxOne group, it is also possible to find two and three groups but at most twenty groups, as this is necessarily limited by the total number of approaches 101. At least one player can not participate in the combination in any way, and then the players which cannot be combined and the next round of matched players are required to perform all processes again.

Number of players matched in step 1) and t generated in step 5)_dmaxIs consistent, the arithmetic matching algorithm realizes five different t_dmaxBy the following steps:

And 5. step 5.2.2, take t of the first player with p_dmaxT of the first player, denoted t1, reserved t1_dmaxDisappear from the queue, then p moves one bit backwards, pointing to t of the second player_dmax。

Step 5.2.3, take t of the second player again with p_dmaxIf t1 is equal to t of the second player_dmaxOtherwise, t of the second player is reserved_dmaxT2 and t1 form a combination, denoted t2, of the second player_dmaxAlso disappear from the queue; otherwise, t of the second player is set_dmaxThe home position of the queue is retired. P is shifted back by one bit.

The maximum function is of the form d_maxF (d1, d2, d3, d4, d5) for obtaining the maximum value or one of the maximum values among d1, d2, d3, d4, d5 of each player. F is a function corresponding relation, and the algorithm comprises the following steps:

step 5.1.1, first d_maxD1, i.e., assume that d1 is the maximum or one of the maxima.

Step 5.1.2, mixing d_maxIf d2 is greater than d, as compared to d2_maxThen let d_maxD2 otherwise_maxThe value is unchanged.

Repeat step 5.1.2, d_maxComparing with d3, d4 and d5 in sequence to finally obtain the maximum value or one of the maximum values d of d1, d2, d3, d4 and d5_max。

For example, a player has a t-d table as shown in FIG. 5, where 5 d values are recorded as d1, d2, d3, d4, d5, and the maximum value or one of the maximum values of the 5 d values is recorded as d_max

That should determine how to d_maxIs there?

Assuming that the player's d-value is 45665 respectively,

suppose d_max4, 4 corresponds to d1, d_maxComparison with d2, i.e. 4 and 5, since 5>4, so update d_maxLet d_max＝5；d_maxComparison with d3, i.e. 5 and 6 (because d_maxHas been updated) because 6>5, continuously updating d_max，d_max＝6；d_maxCompare with d4, i.e. 6 and 6, because 6 ═ 6, d is not updated_max，d_maxIs still 6; d_maxComparison with d5, i.e. 6 and 5, since 5<6, do not update d_max，d_maxIs still 6;

thus d_maxWith 6, i.e. d3, corresponding to t being 010.

Claims

1. an algorithmic method for optimizing ten online competitive game matching mechanism, is characterized in that, comprises the following steps:

1) Collection of player game information

Collect the game information of each player who entered at a certain time in the last N games, where the game information refers to two aspects, one is the score of the last N games, and the other is the character occupation selection of the last N games; provide it to step 2 );

2) 1-D convolution for scoring

Convolution operation is performed on each player's N-field rating score with a 1-row N-column 1-D template, the coefficients of the template are in one-to-one correspondence with each player's N-field rating score, and it is drawn that matches in step 1). The result of the convolution operation with the same number of people, denote the result of the convolution as G; after all players perform the same operation, each matching player in step 1) has a G value; provide it to step 3), step 6);

3) The interval setting of the score

The G values of ten players are randomly sampled from all matching players in step 2), that is, ten G values are extracted, and the ten G values are subjected to 1-D convolution using the template in Figure 4, and the convolution result is denoted as A, Set A±20% as the scoring interval; provide to step 6);

4) Function mapping of role occupation

Perform the same operation on all players matched in step 1) respectively, and perform function mapping on the choice of each player's role occupation in N games in step 1). The mapping function is in the form of t=E (role occupation), where E is the functional relationship , t is the mapping result, the t of each occupation in the five occupations is unique, and the same t is counted to obtain the number d corresponding to each t, that is, the table t-d in Figure 5, all players matched in step 1) have a respective Figure 5t-d table; provided to step 5);

5) Generation of player combinations

Obtain the maximum value d _max of each player's d (d1, d2, d3, d4, d5) value in step 4) through the maximum value function, and its corresponding t is one of 000, 001, 010, 011, and 100, denoted Be t _dmax ; Arithmetic matching five different t _dmax is a player combination, traverse step 1) all matching players, generate the player combination of a group of five, i.e. obtain preliminary player combination, provide step 6) screening;

6) The final player combination is obtained

Denote the G value of each group of players in the preliminary player combination obtained in step 5) as G1, G2, G3, G4, G5, and calculate _Gave = (G1+G2+G3+G4+G5)/5, if _Gave is in the middle interval of step 3), then this group of players is the final player combination.

2. method as claimed in claim 1, is characterized in that, in step 5),

The arithmetic matching algorithm realizes the combination of five different t _dmax by the following steps:

Step 5.2.1, form a queue with the t _dmax of all players matched in 1), the length of the queue is the same as the number of players, and set a flag p to point to the t _dmax of the first player in the queue;

Step 5.2.2, use p to take out the t _dmax of the first player, record it as t1, keep t1, the t _dmax of the first player disappears from the queue, and then p moves backward one place, pointing to the t of the second player _dmax ;

Step 5.2.3, use p again to take out the t _dmax of the second player. If t1 is different from the t _dmax of the second player, keep the t _dmax of the second player and record it as t2. t2 and t1 form a combination. The second player's t _dmax is also removed from the queue; otherwise, the second player's t _dmax is returned to its original position in the queue. Shift p backward by one;

Step 5.2.4, repeat step 5.2.3 to get t3, t4, t5, then 5 players corresponding to t1, t2, t3, t4, t5 are a group of players;

Step 5.2.4-1 If p points to the t _dmax of the last player in the queue, that is, at the end of the first traversal, it is found that a group of players cannot be formed, then the players that have been taken out, the remaining players in the queue and the next batch The players who have entered the match re-execute all the processes of the present invention, and do not execute steps 5.2.5, 5.2.6 and 6); if at least one group of player combinations is generated after the first traversal, continue to execute steps 5.2.5 and 5.2.5. 5.2.6 and step 6);

Step 5.2.5, move p again to the t _dmax of the first player among the remaining players in the queue, and repeat the above steps 5.2.1 to 5.2.4 to achieve the goal of traversing all players;

Step 5.2.6, when p points to the t _dmax of the last player in the queue, it is found that a group of t1, t2, t3, t4, t5 combinations can no longer be obtained, then the players that have been taken out, the remaining players in the queue, and the steps In 6), the combination of players that do not meet the scoring interval in step 3) and the next batch of players who enter the matching process re-execute all the processes of the present invention.