CN113077274A

CN113077274A - Data flow control method and advertisement data flow control method

Info

Publication number: CN113077274A
Application number: CN202010006958.7A
Authority: CN
Inventors: 林叶; 张富
Original assignee: Shanghai Jiatou Internet Technology Group Co ltd
Current assignee: Shanghai Jiatou Internet Technology Group Co ltd
Priority date: 2020-01-03
Filing date: 2020-01-03
Publication date: 2021-07-06

Abstract

The invention discloses a data flow control method and an advertisement data flow control method, which comprise the following steps: (1) counting traffic consumption data of different nodes at different positions; (2) counting association thresholds of different nodes at the positions; (3) judging whether the traffic consumption data weights of different nodes at different positions are positively correlated with the correlation threshold weight, if so, setting a task threshold, and distributing the task threshold by the traffic consumption data weights; otherwise, executing the step (4); (4) and (4) recalculating the traffic consumption data weights of different nodes at different positions based on a preset algorithm, and returning to the step (3). As described above, the data traffic control method and the advertisement data traffic control method of the present invention solve the problems that the advertisement procurement traffic delivery effect cannot be guaranteed to be stable, the advertisement distribution period is slow, and the efficiency is low.

Description

Data flow control method and advertisement data flow control method

Technical Field

The invention relates to the field of network real-time flow distribution, in particular to a data flow control method and an advertisement data flow control method.

Background

At present, advertisement purchase flow is mostly the manual optimization mode, and the experience of advertisement putting personnel is relied on to handle, because unified standard does not exist at present, and a lot of rely on the training to put in the optimization, and an advertisement putting personnel need cultivate the cycle comparatively long, because all manual handling in addition, put in the effect and can not guarantee stably, and it is comparatively slow, inefficiency according to data operating personnel at calculation distribution advertisement cycle simultaneously.

In addition, the advertisement traffic distribution is an average distribution control method, that is, all traffic control is distributed to consumers on average according to the proportion of the total traffic and purchased, and since the demand of each consumer for advertisement traffic at each time is different, the traffic required by the consumers may not be purchased according to the conventional traffic distribution mode, and the issued traffic has no value to the consumers and causes waste to both consumers.

Disclosure of Invention

In order to overcome the defects of the prior art, a data flow control method and an advertisement data flow control method are provided, which are used for solving the problems that the current advertisement purchasing flow delivery effect cannot be guaranteed to be stable, meanwhile, the advertisement distribution period is relatively slow in calculation and distribution according to data operating personnel, the efficiency is low, flow control is averagely distributed to consumers according to the proportion of total flow for purchasing, and because the demand of each consumer for advertisement flow at each moment is different, the flow required by the consumers can not be purchased according to the traditional flow distribution mode, and the found flow has no value to the consumers and wastes the two parties.

In order to achieve the above object, the present invention provides a data traffic control method, which comprises the following steps:

(1) counting traffic consumption data of different nodes at different positions;

(2) counting association thresholds of different nodes at the positions;

(3) judging whether the traffic consumption data weights of different nodes at different positions are positively correlated with the correlation threshold weight, if so, setting a task threshold, and distributing the task threshold by the traffic consumption data weights;

otherwise, executing the step (4);

(4) and (4) recalculating the traffic consumption data weights of different nodes at different positions based on a preset algorithm, and returning to the step (3).

In a further improvement, the predetermined algorithm comprises a genetic algorithm, a simulated annealing algorithm, a random search algorithm or a traversal algorithm.

In a further improvement, the step of recalculating the traffic consumption data weights for different nodes in the location based on a genetic algorithm comprises:

(1) respectively taking the traffic consumption data and the associated threshold values of different advertisement positions as a gene;

(2) using a combination of traffic consumption data and associated thresholds for different advertisement positions as a chromosome;

(3) distributing genes of the chromosomes, setting the value range of the genes of the chromosomes to be 0-100 and weighting the corresponding flow consumption data to be 0-100%;

(4) calculating scores according to genes of the chromosomes, wherein the scores are (traffic consumption data of an A advertisement position + traffic consumption data of a B advertisement position + …)/(traffic consumption data of the A advertisement position + traffic consumption data of the B advertisement position + …), recording the highest scores and the genes of the corresponding chromosomes, and if the calculation times reach a preset value, outputting the genes of the chromosomes corresponding to the highest scores; and (4) if the calculated times do not reach the preset value, returning to the step (3).

In a further refinement, said traffic consumption data comprises: advertisement slot information, user information, and media information sent to the consumers.

The invention also provides an advertisement data flow control method, which comprises the following steps:

(1) counting traffic consumption data of different advertisement positions;

(2) counting the click rate of the advertisement positions with different positions;

(3) judging whether the traffic consumption data weights of the advertisement positions with different positions are positively correlated with the click rate weight, if so, setting an expected click rate, and distributing the expected click rate according to the traffic consumption data weights;

otherwise, executing the step (4);

(4) and (4) recalculating the traffic consumption data weight of the advertisement positions with different positions based on a preset algorithm, and returning to the step (3).

In a further improvement, the step of recalculating the traffic consumption data weights for different ad slots in a location based on a genetic algorithm comprises:

(1) respectively taking the flow consumption data and the click rate of different advertisement positions as a gene;

(2) taking the combination of the traffic consumption data and the click rate of different advertisement positions as a chromosome;

(4) calculating scores according to genes of the chromosomes, wherein the scores are (flow consumption data of an advertisement position A + click rate of an advertisement position B + …)/(flow consumption data of an advertisement position A + flow consumption data of an advertisement position B + …), recording the highest scores and the genes of the corresponding chromosomes, and outputting the genes of the chromosomes corresponding to the highest scores if the calculation times reach a preset value; and (4) if the calculated times do not reach the preset value, returning to the step (3).

A further improvement is that the traffic consumption data includes ad spot information, user information and media information sent to the consumers.

(1) counting traffic consumption data of different advertisement positions;

(2) counting the purchase rate of the advertisement positions in different positions;

(3) judging whether the traffic consumption data weights of the advertisement positions with different positions are positively correlated with the purchase rate weight, if so, setting the query rate, and distributing the query rate according to the traffic consumption data weights;

otherwise, executing the step (4);

(1) respectively taking the flow consumption data and the purchase rate of different advertisement positions as a gene;

(2) using the combination of the traffic consumption data and the purchase rate of different advertisement positions as a chromosome;

(4) calculating and scoring according to genes of the chromosomes, wherein the score is (traffic consumption data of the advertisement position A. gene. A. Purchase rate of the advertisement position + B. advertisement position' B. gene. B. Purchase rate of the advertisement position + …)/(traffic consumption data of the advertisement position A + B. advertisement position + …), recording the highest score and the genes of the corresponding chromosomes, and outputting the genes of the chromosomes corresponding to the highest score if the calculation times reach a preset value; and (4) if the calculated times do not reach the preset value, returning to the step (3).

The invention relates to a data flow control method and an advertisement data flow control method, which are used for counting flow consumption data of different nodes at different positions; counting association thresholds of nodes at different positions, judging whether the traffic consumption data weights of the nodes at different positions are positively correlated with the association threshold weights, if so, setting a task threshold, and distributing the task threshold by the traffic consumption data weights; otherwise, based on a preset algorithm, recalculating the traffic consumption data weights of different nodes at different positions, setting a task threshold, and distributing the task threshold by the traffic consumption data weights.

Drawings

Fig. 1 is a binary sequence diagram of a chromosome in an advertisement traffic distribution method disclosed in an embodiment of the present invention;

FIG. 2 is a binary sequence diagram of cross chromosomes in an advertisement traffic distribution method disclosed in an embodiment of the present invention;

FIG. 3 is a binary sequence diagram of a gene mutation chromosome in an advertisement traffic distribution method disclosed in an embodiment of the present invention;

FIG. 4 is a diagram showing a chromosome arrangement code map in an advertisement traffic distribution method according to an embodiment of the present invention;

fig. 5 is a flowchart illustrating an advertisement traffic distribution method according to an embodiment of the present invention.

Detailed Description

The following description of the embodiments of the present invention is provided for illustrative purposes, and other advantages and effects of the present invention will become apparent to those skilled in the art from the present disclosure.

Please refer to fig. 1 to 5. It should be understood that the structures, ratios, sizes, and the like shown in the drawings and described in the specification are only used for matching with the disclosure of the specification, so as to be understood and read by those skilled in the art, and are not used to limit the conditions under which the present invention can be implemented, so that the present invention has no technical significance, and any structural modification, ratio relationship change, or size adjustment should still fall within the scope of the present invention without affecting the efficacy and the achievable purpose of the present invention. In addition, the terms "upper", "lower", "left", "right", "middle" and "one" used in the present specification are for clarity of description, and are not intended to limit the scope of the present invention, and the relative relationship between the terms and the terms is not to be construed as a scope of the present invention.

Referring to fig. 1 to 5, the present invention provides a data flow control method, as shown in fig. 1, including the following steps:

(2) counting correlation threshold values of different nodes at different positions;

otherwise, executing the step (4);

In the data flow control method provided by this embodiment, the preset algorithm includes a genetic algorithm, a simulated annealing algorithm, a random search algorithm, or a traversal algorithm.

In the data flow control method provided in this embodiment, the step of recalculating the flow consumption data weights of different nodes in different positions based on a genetic algorithm includes:

In the data flow control method provided in this embodiment, the flow consumption data includes: advertisement slot information, user information, and media information sent to the consumers.

(1) counting traffic consumption data of different advertisement positions;

(2) counting the click rate of different advertisement positions;

otherwise, executing the step (4);

In the advertisement data traffic control method provided in this embodiment, the step of recalculating the traffic consumption data weights of advertisement slots at different positions based on a genetic algorithm includes:

In the advertisement data flow control method provided by this embodiment, the flow consumption data includes advertisement slot information, user information, and media information sent to the consumer.

(1) counting traffic consumption data of different advertisement positions;

(2) counting the purchase rate of different advertisement positions;

otherwise, executing the step (4);

(4) calculating and scoring according to genes of the chromosomes, wherein the score is (traffic consumption data of the advertisement position A. gene. A. Purchase rate of the advertisement position + B. advertisement position' B. gene. B. Purchase rate of the advertisement position + …)/(traffic consumption data of the advertisement position A + B. advertisement position + …), recording the highest score and the genes of the corresponding chromosomes, and outputting the genes of the chromosomes corresponding to the highest score if the calculation times reach a preset value; and (4) if the calculation times do not reach the preset value, returning to the step (3), and presetting a QPS query rate threshold, and if the sum of the query rates of different advertisement positions is greater than the query rate threshold, not scoring, and returning to the step (3).

Specifically, the traffic consumption data of advertisement positions with different positions are counted; counting the click rate of different advertisement positions;

judging whether the traffic consumption data weights of the advertisement positions with different positions are positively correlated with the click rate weight, if so, setting an expected click rate, and distributing the expected click rate according to the traffic consumption data weights; otherwise, based on a preset algorithm, recalculating the traffic consumption data weights of the advertisement positions at different positions, and then setting an expected click rate to distribute the expected click rate according to the traffic consumption data weights; counting traffic consumption data of different advertisement positions; counting the purchase rate of different advertisement positions; and judging whether the traffic consumption data weights of the advertisement positions with different positions are positively correlated with the purchase rate weights, if so, setting the query rate, distributing the query rate by the traffic consumption data weights, otherwise, recalculating the traffic consumption data weights of the advertisement positions with different positions based on a preset algorithm, then setting the query rate, and distributing the query rate by the traffic consumption data weights.

Among them, Genetic Algorithm (GA) is an Evolutionary Computing (Evolutionary Computing) Algorithm, and belongs to a part of artificial intelligence technology. The genetic algorithm was originally invented and improved by John Holland and his students, and was derived from a simulation of the theory of evolution of da vinci species. In the evolution process of species, good genes are kept and bad genes are eliminated in order to adapt to the environment, so that the genes of the species are the genes with the best fitness under the current natural environment through the change of genes of many generations. The algorithm is widely used in optimization and search for finding the optimal solution (or an approximation of the optimal solution), and its most important steps include cross (crossover) and mutation (mutation).

All organisms are composed of cells, each containing the same chromosome (chromosome). A chromosome is composed of a string of DNA, and we can simply represent an individual organism as a chromosome. Each chromosome contains a gene, which is composed of multiple DNAs. Each gene controls the expression of a trait in an individual, such as the color of the eye, the identity of the eyelids, etc. During species reproduction, crossover occurs first, and chromosomes from parents undergo division and recombination to form chromosomes of offspring. Later, the offspring has a certain probability of gene mutation, i.e., the gene at a certain position on the chromosome changes with a certain probability. Thereafter, both the crossover and mutation steps were repeated for each generation. For each offspring, we can measure its fitness in some way. The better fitness of an individual, the greater the probability of selection in the next crossover, the easier its gene will be passed on to the next generation. Thus, the offspring will be more and more well adapted until a stable value is converged.

In optimization problems where there are always many possible solutions, it is desirable to find an optimal solution that is more adaptive (i.e., has a larger or smaller value for the objective function) than the other possible solutions. Each feasible solution is a "biological individual" that can be represented as a point and fitness in the state space. Each solution is a coded sequence, which has been binary coded for example, and each solution is a binary sequence. Thus each chromosome is a binary sequence. Genetic algorithms begin with a set of feasible solutions, called a population, from which chromosomes are randomly selected for crossover to generate the next generation. This approach would be better suited for the next generation than the previous generation. The procedure of the genetic algorithm is as follows:

1. initializing and randomly generating n chromosomes;

2. calculating a function value, namely fitness, of each chromosome according to the objective function;

3. the next generation is generated:

randomly selecting two chromosomes, wherein the probability that the chromosome with better fitness is selected is higher;

crossing the two chromosomes according to a certain crossing probability to generate two new offspring; if no crossover occurs, the offspring is the two chromosomes themselves;

making the offspring generate gene mutation according to a certain mutation probability;

putting the offspring into a new population;

4.

steps

2 and 3 are repeated using a new output until a termination condition is reached.

The termination condition may be that the maximum number of iterations is reached, or that the fitness difference of the optimal chromosome of successive generations is less than a threshold. The above algorithmic description may not be intuitive enough, and we illustrate it. Each chromosome is a binary sequence, provided that the solution can be represented in binary code. As shown in fig. 1, each chromosome is a 16-bit binary sequence assuming a sequence length of 16.

First, we randomly generate a population, and assuming that the population size is 20, there are 20 binary sequences with length of 16. Fitness of each chromosome is calculated and then two chromosomes are selected to be crossed as shown in fig. 2. The lower panel shows the recombination after the chromosome is broken at the upper position in FIG. 6, and the broken position can be randomly selected. Of course, there may be more than one fracture site. Specific crossing modes can be selected according to specific problems to improve the performance of the algorithm.

Then, a gene mutation occurs in a randomly selected gene on the chromosome of the offspring, and the position of the mutation is randomly selected as shown in FIG. 3. Moreover, gene mutations do not occur in every offspring, but only with a certain probability. For binary coding, the way in which a gene is mutated is by bitwise inversion.

The above examples are for binary coding, which can be used if the maximum and minimum values of the unary function within a certain interval are solved. For example, the solving function f (x) is the minimum value of x + sin (3x) + cos (3x) in the interval [0,6 ]. Assuming we need the minimum point x to retain 4 decimals, the solution interval is discretized into 60000 numbers. Because 2{15} <60000<2{16}, a 16-bit binary number is needed to represent the 60000 possible solutions. Where 0x0000 represents 0, 0x0001 represents 0.0001, and so on. For this example, the remove code is given at the end of the text.

However, binary coding cannot be used in the ordering problem, and permutation coding (permutation encoding) should be employed. For example, there are two chromosomes in fig. 4.

And (3) crossing: a crossover point is randomly chosen from which to break the two chromosomes. The front part of chromosome a is grouped into the front part of offspring 1, then chromosome B is scanned, and if genes not contained in offspring 1 appear, they are added sequentially to offspring 1. Similarly, the anterior part of chromosome B constitutes the anterior part of progeny 2, and chromosome A is scanned to obtain the posterior part of progeny 2. Note that the way of intersection is various, and only one of them is mentioned here.

(1 5 3 2 6|4 7 9 8)+(8 5 6 7 2|3 1 4 9)＝>(1 5 3 2 6 8 7 4 9)+(8 5 6 7 2 1 3 4 9)

Mutation: for one chromosome, the two genes were randomly chosen to interchange positions. For example, the 3 rd gene and the 2 nd last gene are interchanged:

(1 5 3 2 6 8 7 4 9)＝>(1 5 4 2 6 8 7 3 9)

in practical genetic algorithms, a few elites (elites) in the previous generation are retained, i.e. the chromosomes with the best fitness in the previous generation are added to the postings of the offspring, and the chromosomes with the worst fitness in the offspring are removed. With this strategy, if an optimal solution is generated in a certain iteration, the optimal solution can be retained until the end of the iteration.

In summary, the advertisement traffic distribution method of the present invention solves the problems that the advertisement purchasing traffic putting effect cannot be guaranteed to be stable at present, meanwhile, the advertisement distribution period is relatively slow in calculation and distribution according to data operating personnel, the efficiency is low, and the traffic control is averagely distributed to consumers for purchasing according to the proportion of the total traffic, and since the demand of each consumer for advertisement traffic at each moment is different, the traffic required by the consumers cannot be purchased according to the previous traffic distribution mode, and the issued traffic has no value to the consumers, which causes waste to both consumers. Reasonable design and is suitable for production, popularization and application. Therefore, the invention effectively overcomes various defects in the prior art and has high industrial utilization value.

The foregoing embodiments are merely illustrative of the principles and utilities of the present invention and are not intended to limit the invention. Any person skilled in the art can modify or change the above-mentioned embodiments without departing from the spirit and scope of the present invention. Accordingly, it is intended that all equivalent modifications or changes which can be made by those skilled in the art without departing from the spirit and technical spirit of the present invention be covered by the claims of the present invention.

Claims

1. A data flow control method is characterized by comprising the following steps:

(2) counting association thresholds of different nodes at the positions;

otherwise, executing the step (4);

2. The data flow control method of claim 1, wherein the pre-set algorithm comprises a genetic algorithm, a simulated annealing algorithm, a random search algorithm, or a traversal algorithm.

3. The data traffic control method of claim 2, wherein the step of recalculating the traffic consumption data weights for different nodes at different locations based on a genetic algorithm comprises:

4. The data flow control method of claim 2, wherein the flow consumption data comprises: advertisement slot information, user information, and media information sent to the consumers.

5. An advertisement data flow control method is characterized by comprising the following steps:

(1) counting traffic consumption data of different advertisement positions;

otherwise, executing the step (4);

6. The advertisement data traffic control method of claim 5, wherein the step of recalculating the traffic consumption data weights for different ad slots in different positions based on a genetic algorithm comprises:

7. The advertisement data flow control method of claim 5, wherein the flow consumption data includes advertisement slot information, user information, and media information transmitted to a consumer.

8. An advertisement data flow control method is characterized by comprising the following steps:

(1) counting traffic consumption data of different advertisement positions;

otherwise, executing the step (4);

9. The advertisement data traffic control method of claim 8, wherein the step of recalculating the traffic consumption data weights for different ad slots in different positions based on a genetic algorithm comprises:

10. The advertisement data flow control method of claim 8, wherein the flow consumption data includes advertisement slot information, user information, and media information transmitted to a consumer.