Intelligent decision management system and method based on big data

Abstract

The invention discloses an intelligent decision management system and method based on big data, and relates to the technical field of intelligent breeding. Performing growth stage division according to historical beef cattle growth data to obtain a historical individual growth curve of each growth stage; aggregating the historical individual growth curves to obtain a reference individual growth curve; the matching degree between the current growth curve of the target beef cattle and the growth curve of the reference body is obtained, the growth curve needed by the target beef cattle is determined according to user requirements, the target beef cattle is fed, and comprehensive analysis is conducted on the growth cycle historical data and real-time data of the beef cattle; analyzing a corresponding relation between the individual growth curve of the beef cattle and historical data; and the marketing time of the beef cattle is determined according to the corresponding relation, and the marketing strategy is dynamically adjusted and the proper marketing time is selected according to influence factors such as market supply and demand dynamics, farm resource allocation and the like, so that pasture benefit maximization is realized, and market competitiveness is guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of intelligent breeding technology, and in particular to an intelligent decision-making management system and method based on big data. Background Art

[0002] In recent years, with the improvement of people's living standards and changes in dietary structure, the demand for meat has gradually increased, especially the demand for beef. Therefore, the beef cattle breeding industry has developed rapidly and the market scale has continued to expand. With the development of science and technology, beef cattle breeding technology has also been continuously improving. For example, through breeding, nutritional regulation, disease prevention and control and other means, the growth rate and meat quality of beef cattle have been improved. At present, the growth curve of beef cattle is based on the group measurement of beef cattle, but there are differences between individual beef cattle. In addition, due to changes in demand, users may need to make timely adjustments to the breeding methods of beef cattle. At this time, the growth of beef cattle will be affected by the original breeding methods. For this reason, how to make intelligent decisions on beef cattle breeding methods based on user needs has become an urgent problem to be solved. Summary of the invention

[0003] The purpose of the present invention is to provide an intelligent decision-making management system and method based on big data to solve the problems raised in the prior art.

[0004] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: an intelligent decision-making management system based on big data, comprising a data acquisition module, a data storage module, a growth analysis module and a user interaction module; the output end of the data acquisition module is connected to the input end of the data storage module, so as to obtain the growth data of beef cattle and send it to the data storage module; the data storage module is interconnected with the growth analysis module, so as to store historical beef cattle growth data and corresponding feeding data; the growth analysis module is interconnected with the user interaction module, so as to analyze the growth conditions of target beef cattle at different stages and determine the feeding method of target beef cattle according to user needs; the user interaction module is used to obtain user demand information.

[0005] The growth analysis module also includes a growth stage division unit, an individual growth curve aggregation unit, a similarity analysis unit and a growth time analysis unit; the growth stage division unit is used to divide the growth of beef cattle into K growth stages, K being the number of beef cattle growth stages; the individual growth curve aggregation unit is used to aggregate the curves of historical beef cattle growth data at each growth stage; the similarity analysis unit is used to determine a target individual growth curve similar to the target beef cattle; the growth time analysis unit is used to combine the target individual growth curves of K growth stages to determine the growth time of the target beef cattle. The growth stage division unit divides the historical beef cattle growth data into K growth stages through the Kmeans unsupervised classification algorithm, and adds the beef cattle growth data with adjacent growth times to the same classification cluster. The growth time analysis unit obtains the duration of each growth stage in the target individual growth curve of the target beef cattle, combines the target individual growth curves of K different growth stages, and obtains the growth time of the target beef cattle; according to the user's demand for growth time, determines the target individual growth curve combination that meets the requirements, and obtains the feeding method corresponding to the target individual growth curve combination. The user interaction module also includes an input module and an output module; the input module is used to input the user's requirements for the growth time of beef cattle; and the output module is used to output the feeding method of beef cattle.

[0006] To achieve the above object, the present invention provides the following technical solution: an intelligent decision-making management method based on big data, comprising the following steps:

[0007] S11, obtaining historical beef cattle growth data;

[0008] S12, dividing the growth stages according to the historical growth data of beef cattle, and obtaining the historical individual growth curve of each growth stage;

[0009] S13, aggregate the historical individual growth curves to obtain the reference individual growth curve;

[0010] S14, obtaining the matching degree between the current growth curve of the target beef cattle and the growth curve of the reference body, determining the growth curve required by the target beef cattle according to user needs, and feeding the target beef cattle.

[0011] In step S12, the growth stage division is performed according to the historical beef cattle growth data to obtain the historical individual growth curve of each growth stage, and the following steps are also included:

[0012] S21, obtain the historical beef cattle growth vector X from the historical beef cattle growth data, X = (x 1 ,x 2 ,x 3 ) The growth vector includes the cumulative growth x 1 , Growth time x2 and growth rate x 3 , where the growth rate is obtained by subtracting the cumulative growth of the next growth vector with adjacent growth time from the cumulative growth of the previous growth vector;

[0013] S22, randomly selecting K growth vectors as initial centroids according to the growth time of the growth vectors, each centroid forming a cluster; where K is the number of growth stages of beef cattle;

[0014] S23, calculate the distance D from each vector to the centroid, Where Δx 1 , Δx 2 and Δx 3 is the cumulative growth amount, growth time and growth speed of the vector and the distance from the center of mass, a 1 、a 2 and a 3 The weights of the accumulated growth amount, growth time and growth speed are assigned to the cluster with the nearest centroid;

[0015] S24, for each cluster, calculate the average value of all vectors in the cluster in terms of cumulative growth amount, growth time and growth speed, and use the average value as the new centroid;

[0016] S25, repeating steps S23 and S24, when the centroid change is less than the first threshold, stopping the iteration and entering step S26;

[0017] S26, analyze the growth time distribution of the vectors in the K clusters, and determine whether the concentration of the growth time of the vectors in each cluster meets the requirements: arrange the growth time of the vectors in the cluster in descending order, calculate the difference between the growth time of the adjacent next item and the previous item, obtain the average value of all the differences, compare the average value with the second threshold value, if the average value is not greater than the second threshold value, then the concentration of the growth time of the vectors in the cluster meets the requirements, and enter step S27; if the average value is greater than the second threshold value, then the concentration of the growth time of the vectors in the cluster does not meet the requirements, and adjust the weight a 1 、a 2 and a 3 , return to step S23;

[0018] S27, obtain the difference in growth time of the vectors in the cluster in step S26, determine that all differences are equal to the growth time of the minimum unit, and among all the determined growth times, obtain the minimum value min and the maximum value max of the growth time, and obtain the growth stage [min, max] corresponding to the cluster and the duration max-min of the growth stage; the minimum unit is the sampling period for generating historical beef cattle growth data.

[0019] Since the growth stage of beef cattle is continuous, the growth stage of beef cattle can be reflected according to the cumulative growth amount, growth time and growth rate; when a certain error is allowed, vectors with continuous growth time are divided into the same cluster; in the growth process of beef cattle, data is obtained according to a fixed sampling period, such as collecting data once a day. Then, in a growth stage of beef cattle, after the growth time is arranged in descending order, the difference between the adjacent next item and the previous item is 1; by analogy to the classification result, the growth time of the vectors in the cluster is arranged in descending order, the difference between the adjacent next item and the previous item is calculated, and the average value of all differences is obtained, and the average value is used as a standard for measuring errors; at the same time, the growth time difference generated by the normal vector in the cluster is also 1 day, and the growth time difference generated by the error vector in the cluster is not 1 day; the starting point and end point of the growth stage are determined by the normal vector, so that the error vector will not affect the stage division.

[0020] In step S13, the historical individual growth curves are aggregated to obtain a reference individual growth curve, and the following steps are also included:

[0021] The historical individual growth curve is obtained according to the historical beef cattle growth data, and the historical individual growth curve of each growth stage is obtained according to the growth stage division result. The difference between any two curves with similar duration in the same growth stage is analyzed, and the root mean square error is used to characterize the difference between the two curves. The root mean square error between two curves with similar duration in the same growth stage is calculated. If the root mean square error is less than the threshold value L, the two curves are merged into one and the original curve is deleted. The threshold L is set automatically. If the root mean square error is greater than the threshold value L, the two curves are not merged and both curves are retained. After aggregation, several reference individual growth curves of K stages are obtained.

[0022] After each beef cattle is slaughtered, a growth curve can be generated, which will produce more growth curves. As the number of farmed beef cattle increases, the number of growth curves will also increase. Therefore, it is necessary to aggregate the growth curves to achieve the purpose of reducing the reference growth curves. Since the aggregation is based on the stage, the reference growth curve of the previous stage after aggregation can correspond to more than one reference growth curve of the next stage.

[0023] In step S14, the step of obtaining the matching degree between the current growth curve of the target beef cattle and the growth curve of the reference individual further comprises the following steps:

[0024] Obtain the current growth curve of the target beef cattle, divide the current growth data of the target beef cattle into growth stages, determine the current growth stage of the target beef cattle and record it as r;

[0025] Set variable b to 1;

[0026] S51, calculating the similarity between the growth curve of the target beef cattle at the b-th growth stage and the growth curve of the reference individual, if the similarity is not less than a set value, adding the growth curve of the reference individual at the b-th growth stage and the corresponding growth curve of the reference individual at the subsequent growth stage to the set U; if the similarity is not less than the set value, not adding them to the set U;

[0027] S52, determine whether b is greater than r, if not, proceed to step S53, otherwise proceed to step S54;

[0028] S53, adding 1 to the value of b, calculating the similarity between the growth curve of the target beef cattle at the b-th growth stage and the growth curve of the reference individual in the set U, if the similarity is not less than the set value, retaining the growth curve of the reference individual in the set U; if the similarity is not less than the set value, removing the corresponding reference individual growth curve at the subsequent growth stage of the reference individual growth curve at the b-th growth stage in the set U, and returning to step S52;

[0029] S54, extracting the reference individual growth curve in the set U to obtain the target individual growth curve of the target beef cattle.

[0030] Considering the individual specificity of the target beef cattle, the subsequent growth curve of the target beef cattle is determined according to the growth curve that the target beef cattle have experienced. At this time, if the user needs to speed up the marketing of beef cattle, a combination with a short growth time can be selected. If the marketing time needs to be extended, a combination with a longer growth time can be selected.

[0031] In step S14, determining the growth curve required by the target beef cattle according to the user's needs specifically includes the following steps:

[0032] In the target individual growth curve of the target beef cattle, the duration of each growth stage is obtained, and the target individual growth curves of K different growth stages are combined to obtain the growth time of the target beef cattle; according to the user's demand for growth time, a target individual growth curve combination that meets the requirements is determined, and the beef cattle are fed according to the feeding method corresponding to the target individual growth curve combination.

[0033] Compared with the prior art, the beneficial effects of the present invention are: comprehensive analysis of historical data and real-time data of the growth cycle of beef cattle, analysis of the correspondence between the individual growth curve of beef cattle and the historical data; determination of the time for the release of beef cattle according to the correspondence, and dynamic adjustment of the release strategy according to market supply and demand dynamics, farm resource allocation and other influencing factors, selection of the appropriate release time, and adjustment of the breeding method of beef cattle; so as to maximize the benefits of the pasture, reduce the loss of income caused by human decision-making, improve the economic benefits of beef cattle breeding, reduce the breeding costs, and ensure market competitiveness. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 This is a schematic diagram of the structure of an intelligent decision-making management system based on big data in the present invention;

[0035] Figure 2 The present invention is a flowchart of an intelligent decision-making management method based on big data. DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0037] Example: Figure 1 As shown, the present invention provides a technical solution, an intelligent decision-making management system based on big data, comprising a data acquisition module, a data storage module, a growth analysis module and a user interaction module; the output end of the data acquisition module is connected to the input end of the data storage module, for acquiring the growth data of beef cattle and sending it to the data storage module; the data storage module and the growth analysis module are interconnected, for storing historical beef cattle growth data and corresponding feeding data; the growth analysis module and the user interaction module are interconnected, for analyzing the growth conditions of target beef cattle through stages, and determining the feeding method of target beef cattle according to user needs; the user interaction module is used to obtain user demand information.

[0038] The growth analysis module also includes a growth stage division unit, an individual growth curve aggregation unit, a similarity analysis unit and a growth time analysis unit; the growth stage division unit is used to divide the growth of beef cattle into K growth stages, K being the number of beef cattle growth stages; the individual growth curve aggregation unit is used to aggregate the curves of historical beef cattle growth data at each growth stage; the similarity analysis unit is used to determine a target individual growth curve similar to the target beef cattle; the growth time analysis unit is used to combine the target individual growth curves of K growth stages to determine the growth time of the target beef cattle. The growth stage division unit divides the historical beef cattle growth data into K growth stages through the Kmeans unsupervised classification algorithm, and adds the beef cattle growth data with adjacent growth times to the same classification cluster. The growth time analysis unit obtains the duration of each growth stage in the target individual growth curve of the target beef cattle, combines the target individual growth curves of K different growth stages, and obtains the growth time of the target beef cattle; according to the user's demand for growth time, determines the target individual growth curve combination that meets the requirements, and obtains the feeding method corresponding to the target individual growth curve combination. The user interaction module also includes an input module and an output module; the input module is used to input the user's requirements for the growth time of beef cattle; and the output module is used to output the feeding method of beef cattle.

[0039] Embodiment: The present invention provides a technical solution, an intelligent decision-making management method based on big data, comprising the following steps:

[0040] S11, obtaining historical beef cattle growth data;

[0041] S12, dividing the growth stages according to the historical growth data of beef cattle to obtain the historical individual growth curve of each growth stage; specifically including steps S21 to S27:

[0042] S21, obtain the historical beef cattle growth vector X from the historical beef cattle growth data, X = (x 1 ,x 2 ,x 3 ) The growth vector includes the cumulative growth x 1 , Growth time x 2 and growth rate x 3 , where the growth rate is obtained by subtracting the cumulative growth of the next growth vector with adjacent growth time from the cumulative growth of the previous growth vector;

[0043] S22, randomly select K growth vectors as initial centroids according to the growth time of the growth vector, and each centroid forms a cluster; wherein K is the number of growth stages of beef cattle; the growth stages of beef cattle can be divided into five stages: calf period, weaning period, rack period, late intensive fattening period and late fattening period, and K is 5 at this time;

[0044] S23, calculate the distance D from each vector to the centroid, Where Δx 1 , Δx 2 and Δx 3 is the cumulative growth amount, growth time and growth speed of the vector and the distance from the center of mass, a 1 、a 2 and a 3 The weights of the accumulated growth amount, growth time and growth speed are assigned to the cluster with the nearest centroid;

[0045] S24, for each cluster, calculate the average value of all vectors in the cluster in terms of cumulative growth amount, growth time and growth speed, and use the average value as the new centroid;

[0046] S25, repeating steps S23 and S24, when the centroid change is less than the first threshold, stopping the iteration and entering step S26;

[0047] S26, analyze the growth time distribution of the vectors in the K clusters, and determine whether the concentration of the growth time of the vectors in each cluster meets the requirements: arrange the growth time of the vectors in the cluster in descending order, calculate the difference between the growth time of the adjacent next item and the previous item, obtain the average value of all the differences, compare the average value with the second threshold value, if the average value is not greater than the second threshold value, then the concentration of the growth time of the vectors in the cluster meets the requirements, and enter step S27; if the average value is greater than the second threshold value, then the concentration of the growth time of the vectors in the cluster does not meet the requirements, and adjust the weight a 1 、a 2 and a 3 , return to step S23; if the concentration of the growth time does not meet the requirements, the weight a 2 Increase, and correspondingly reduce the weight a 1 and a 3 .

[0048] S27, obtain the difference in growth time of the vectors in the cluster in step S26, determine that all differences are equal to the growth time of the minimum unit, and among all the determined growth times, obtain the minimum value min and the maximum value max of the growth time, and obtain the growth stage [min, max] corresponding to the cluster and the duration max-min of the growth stage; the minimum unit is the sampling period for generating historical beef cattle growth data.

[0049] S13, aggregating historical individual growth curves to obtain a reference individual growth curve, including the following steps:

[0050] The historical individual growth curve is obtained according to the historical beef cattle growth data, and the historical individual growth curve of each growth stage is obtained according to the growth stage division result. The difference between any two curves with similar duration in the same growth stage is analyzed, and the root mean square error is used to characterize the difference between the two curves. The root mean square error between two curves with similar duration in the same growth stage is calculated. If the root mean square error is less than the threshold value L, the two curves are merged into one and the original curve is deleted. The threshold L is set automatically. If the root mean square error is greater than the threshold value L, the two curves are not merged and both curves are retained. After aggregation, several reference individual growth curves of K stages are obtained.

[0051] Since aggregation is performed based on stages, the reference growth curve of the previous stage after aggregation can correspond to more than one reference growth curve of the next stage. For example, there are historical growth data of 5 beef cattle, and the growth curves in the 5 growth stages are recorded as (L11 L12 L13 L14 L15), (L21 L22 L23 L24 L25), (L31 L32 L33 L34L35), …. When L11 and L21 meet the aggregation standard and L12 and L22 do not meet the aggregation standard, L11 and L21 are aggregated. At this time, L12 and L22 do not meet the aggregation standard, and the aggregated curves of L11 and L21 correspond to L12 and L22 at the same time.

[0052] S14, obtaining the matching degree between the current growth curve of the target beef cattle and the growth curve of the reference body, including the following steps:

[0053] Obtain the current growth curve of the target beef cattle, divide the current growth data of the target beef cattle into growth stages, determine the current growth stage of the target beef cattle and record it as r;

[0054] Set variable b to 1;

[0055] S51, calculating the similarity between the growth curve of the target beef cattle at the b-th growth stage and the growth curve of the reference individual, if the similarity is not less than a set value, adding the growth curve of the reference individual at the b-th growth stage and the corresponding growth curve of the reference individual at the subsequent growth stage to the set U; if the similarity is not less than the set value, not adding them to the set U;

[0056] S52, determine whether b is greater than r, if not, proceed to step S53, otherwise proceed to step S54;

[0057] S53, adding 1 to the value of b, calculating the similarity between the growth curve of the target beef cattle at the b-th growth stage and the growth curve of the reference individual in the set U, if the similarity is not less than the set value, retaining the growth curve of the reference individual in the set U; if the similarity is not less than the set value, removing the corresponding reference individual growth curve at the subsequent growth stage of the reference individual growth curve at the b-th growth stage in the set U, and returning to step S52;

[0058] S54, extracting the reference individual growth curve in the set U to obtain the target individual growth curve of the target beef cattle.

[0059] The similarity can be determined by dynamic time planning or root mean square error. When the root mean square error is used, the minimum value of the root mean square error is calculated to measure the similarity, that is, on the coordinate axis, the growth curve of the target beef cattle at the experienced stage is translated, and the root mean square error between the reference individual curve is calculated. The horizontal axis of the coordinate axis is the growth time, and the vertical axis can be the cumulative growth of the beef cattle. After obtaining the minimum value of the root mean square error, the root mean square error is reversed, such as taking the reciprocal, to obtain the similarity.

[0060] When the target beef cattle are in the weaning period, the curve obtained by aggregating L11 and L21 is recorded as JL11. If the similarity between the target beef cattle and JL11 in the calf period is not less than the set value, JL11 and the corresponding subsequent reference individual curves are added to the set U, including (JL11 L12 L13 L14 L15), (JL 11L22 L23 L24 L25). If the similarity between the target beef cattle and L31 in the calf period is less than the set value, the reference individual curves (L31 L32 L33 L34 L35) are not added to the set U. Since L12 and L22 are not aggregated, it is necessary to perform a separate similarity judgment on whether (L12 L13 L14 L15) and (L22 L23L24 L25) continue to be retained in the set U. After the similarity judgment of all stages is completed, the reference individual growth curve in the set U is the target individual growth curve.

[0061] Determine the growth curve required by the target beef cattle according to user needs, including the following steps:

[0062] In the target individual growth curve of the target beef cattle, the duration of each growth stage is obtained, and the target individual growth curves of K different growth stages are combined to obtain the growth time of the target beef cattle; according to the user's demand for growth time, a target individual growth curve combination that meets the requirements is determined, and the beef cattle are fed according to the feeding method corresponding to the target individual growth curve combination.

[0063] The overall process is as follows: for the target beef cattle in the weaning period, they have gone through the complete calf period and part of the weaning period, and can obtain the growth curve of the calf period and the growth curve of the part of the weaning period. First, according to the growth curve of the calf period, a reference individual growth curve similar to the growth curve of the calf period can be obtained, because each stage will have an impact on the growth of beef cattle, and the growth curve of the calf period will affect the subsequent growth of beef cattle. First, the reference individual growth curve is screened through the growth curve of the calf period;

[0064] Next, since there is still a partial weaning growth curve of the target beef cattle, the beef cattle are in the calf period from January to April, and enter the weaning period after April. When the target beef cattle has experienced 15 days in the weaning period and has not yet entered the rack period, the 15-day growth curve of the target beef cattle is obtained. After the calf period screening, the reference growth curve of the weaning period contained in the set U is a complete growth curve. Due to the different feeding methods of historical beef cattle, there are differences between individuals, so the duration of the growth curves of different reference individuals is also different. At this time, the similarity of the weaning period can be obtained by calculating the partial root mean square error. First, the target beef cattle The growth curve of the cow during the weaning period is aligned with the starting point of the growth curve of the reference individual during the weaning period, and the root mean square error of 15 days is calculated. After reversal, the similarity can be obtained. Since there is no strict definition standard for the growth stage of beef cattle, for example, when the cumulative growth volume is similar, the target beef cattle is judged to be in the calf period, and the historical individual may be judged to be in the weaning period. At this time, the growth curve of the target beef cattle can be translated, that is, the starting point of the growth curve of the target beef cattle during the weaning period is aligned with the starting point of the growth curve of the reference individual during the weaning period a few days after the starting point, and the root mean square error is calculated again; the calculation formula of the root mean square error is: Where RMSE is the root mean square error, y1 α is the αth cumulative growth of the target beef cattle on the growth curve during the weaning period, y2 α is the αth cumulative growth amount aligned on the growth curve of the reference individual at weaning period;

[0065] After screening the reference individual growth curves through the growth curves of the calf period and the weaning period, the target individual growth curves of the weaning period, the rack period, the late intensive fattening period and the late fattening period similar to the target beef cattle are obtained. At the same time, the feeding methods of historical beef cattle corresponding to the target individual growth curve and the growth time of historical beef cattle in the weaning period, the rack period, the late intensive fattening period and the late fattening period can be obtained from the historical data; through the analysis of the market conditions, when the market conditions are in a downward state, it is necessary to speed up the slaughter of beef cattle to reduce losses. For this purpose, the target individual growth curves of the weaning period, the rack period, the late intensive fattening period and the late fattening period with the shortest growth time can be selected according to the corresponding Feeding is carried out according to the feeding method; if the reference individual growth curve is not screened, and only historical data is referred to, the feeding method with the shortest time to market is selected, which may not achieve the desired effect, because the calf period and weaning period experienced by the target beef cattle will affect the growth of the beef cattle; at the same time, when the farm needs to market the beef cattle at a specified time, it can choose to freely combine the duration of the weaning period, rack period, late intensive fattening period and late fattening period according to the target individual growth curve, and select the feeding method that meets the time requirements for feeding. When there are multiple feeding methods that meet the requirements, the feeding method with the least resource consumption can be selected to achieve the purpose of saving resources.

[0066] It will be apparent to those skilled in the art that the invention is not limited to the details of the exemplary embodiments described above and that the invention can be implemented in other specific forms without departing from the spirit or essential features of the invention. Therefore, the embodiments should be considered exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended claims rather than the foregoing description, and it is intended that all variations falling within the meaning and scope of the equivalent elements of the claims be included in the invention. Any reference numeral in a claim should not be considered as limiting the claim to which it relates.

Claims

1. An intelligent decision-making management method based on big data, characterized in that: The following steps are involved: S11, obtaining historical beef cattle growth data; S12, dividing the growth stages according to the historical growth data of beef cattle to obtain the historical individual growth curve of each growth stage; S13, aggregate the historical individual growth curves to obtain the reference individual growth curve; S14, obtaining the matching degree between the current growth curve of the target beef cattle and the growth curve of the reference body, determining the growth curve required by the target beef cattle according to user needs, and feeding the target beef cattle.

2. According to the big data-based intelligent decision-making management method of claim 1, it is characterized in that: In step S12, the growth stage division is performed according to the historical beef cattle growth data to obtain the historical individual growth curve of each growth stage, and the following steps are also included: S21, obtaining a historical beef cattle growth vector X from historical beef cattle growth data, X=(x1, x2, x3), wherein the growth vector includes a cumulative growth amount x1, a growth time x2, and a growth speed x3, wherein the growth speed is obtained by subtracting the cumulative growth amount of a subsequent growth vector with adjacent growth times from the cumulative growth amount of a previous growth vector; S22, randomly selecting K growth vectors as initial centroids according to the growth time of the growth vectors, each centroid forming a cluster; where K is the number of growth stages of beef cattle; S23, calculate the distance D from each vector to the centroid, Where Δx1, Δx2 and Δx3 are the cumulative growth, growth time and growth rate of the vector and the distance from the centroid, and a1, a2 and a3 are the weights of the cumulative growth, growth time and growth rate; the vector is assigned to the cluster with the nearest centroid; S24, for each cluster, calculate the average value of all vectors in the cluster in terms of cumulative growth amount, growth time and growth speed, and use the average value as the new centroid; S25, repeating steps S23 and S24, when the centroid change is less than the first threshold, stopping the iteration and entering step S26; S26, analyzing the growth time distribution of the vectors in the K clusters, and determining whether the concentration of the growth time of the vectors in each cluster meets the requirements: arranging the growth time of the vectors in the cluster in descending order, calculating the difference between the growth time of the adjacent next item and the previous item, obtaining the average value of all the differences, and comparing the average value with the second threshold value. If the average value is not greater than the second threshold value, the concentration of the growth time of the vectors in the cluster meets the requirements, and the process proceeds to step S27; if the average value is greater than the second threshold value, the concentration of the growth time of the vectors in the cluster does not meet the requirements, and the weights a1, a2 and a3 are adjusted, and the process returns to step S23; S27, obtain the difference in growth time of the vectors in the cluster in step S26, determine that all differences are equal to the growth time of the minimum unit, and among all the determined growth times, obtain the minimum value min and the maximum value max of the growth time, and obtain the growth stage [min, max] corresponding to the cluster and the duration max-min of the growth stage; the minimum unit is the sampling period for generating historical beef cattle growth data.

3. The intelligent decision-making management method based on big data according to claim 2 is characterized in that: In step S13, the historical individual growth curves are aggregated to obtain a reference individual growth curve, and the following steps are also included: The historical individual growth curve is obtained according to the historical beef cattle growth data, and the historical individual growth curve of each growth stage is obtained according to the growth stage division result. The difference between any two curves with similar duration in the same growth stage is analyzed, and the root mean square error is used to characterize the difference between the two curves. The root mean square error between two curves with similar duration in the same growth stage is calculated. If the root mean square error is less than the threshold value L, the two curves are merged into one and the original curve is deleted. The threshold L is set automatically. If the root mean square error is greater than the threshold value L, the two curves are not merged and both curves are retained. After aggregation, several reference individual growth curves of K stages are obtained.

4. The intelligent decision-making management method based on big data according to claim 3 is characterized in that: In step S14, the step of obtaining the matching degree between the current growth curve of the target beef cattle and the growth curve of the reference individual further comprises the following steps: Obtain the current growth curve of the target beef cattle, divide the current growth data of the target beef cattle into growth stages, determine the current growth stage of the target beef cattle and record it as r; Set variable b to 1; S51, calculating the similarity between the growth curve of the target beef cattle at the b-th growth stage and the growth curve of the reference individual, if the similarity is not less than a set value, adding the growth curve of the reference individual at the b-th growth stage and the corresponding growth curve of the reference individual at the subsequent growth stage to the set U; if the similarity is not less than the set value, not adding them to the set U; S52, determine whether b is greater than r, if not, proceed to step S53, otherwise proceed to step S54; S53, adding 1 to the value of b, calculating the similarity between the growth curve of the target beef cattle at the b-th growth stage and the growth curve of the reference individual in the set U, if the similarity is not less than the set value, retaining the growth curve of the reference individual in the set U; if the similarity is not less than the set value, removing the corresponding reference individual growth curve at the subsequent growth stage of the reference individual growth curve at the b-th growth stage in the set U, and returning to step S52; S54, extracting the reference individual growth curve in the set U to obtain the target individual growth curve of the target beef cattle.

5. The intelligent decision-making management method based on big data according to claim 4 is characterized in that: In step S14, determining the growth curve required by the target beef cattle according to the user's needs specifically includes the following steps: In the target individual growth curve of the target beef cattle, the duration of each growth stage is obtained, and the target individual growth curves of K different growth stages are combined to obtain the growth time of the target beef cattle; according to the user's demand for growth time, a target individual growth curve combination that meets the requirements is determined, and the beef cattle are fed according to the feeding method corresponding to the target individual growth curve combination.

6. An intelligent decision-making management system based on big data, characterized in that: It comprises a data acquisition module, a data storage module, a growth analysis module and a user interaction module; the output end of the data acquisition module is connected with the input end of the data storage module, so as to obtain the growth data of beef cattle and send it to the data storage module; the data storage module is connected with the growth analysis module, so as to store the historical growth data of beef cattle and the corresponding feeding data; the growth analysis module is connected with the user interaction module, so as to analyze the growth conditions of target beef cattle in different stages and determine the feeding method of target beef cattle according to the user's demand; the user interaction module is used to obtain the user's demand information.

7. The intelligent decision-making management system based on big data according to claim 6 is characterized in that: The growth analysis module also includes a growth stage division unit, an individual growth curve aggregation unit, a similarity analysis unit and a growth time analysis unit; the growth stage division unit is used to divide the growth of beef cattle into K growth stages, K is the number of growth stages of beef cattle; the individual growth curve aggregation unit is used to aggregate the curves of historical beef cattle growth data at each growth stage; the similarity analysis unit is used to determine a target individual growth curve similar to the target beef cattle; the growth time analysis unit is used to combine the target individual growth curves of K growth stages to determine the growth time of the target beef cattle.

8. The intelligent decision-making management system based on big data according to claim 7 is characterized in that: The growth stage division unit divides the historical beef cattle growth data into K growth stages through the Kmeans unsupervised classification algorithm, and adds the beef cattle growth data with adjacent growth times into the same classification cluster.

9. The intelligent decision-making management system based on big data according to claim 7 is characterized in that: The growth time analysis unit obtains the duration of each growth stage in the target individual growth curve of the target beef cattle, combines the target individual growth curves of K different growth stages, and obtains the growth time of the target beef cattle; determines a target individual growth curve combination that meets the requirements according to the user's demand for growth time, and obtains a feeding method corresponding to the target individual growth curve combination.

10. The intelligent decision management system based on big data according to claim 6, characterized in that: The user interaction module also includes an input module and an output module; the input module is used to input the user's requirements for the growth time of beef cattle; and the output module is used to output the feeding method of beef cattle.

Images