Mine operation scheduling system based on big data

Abstract

The present invention relates to the field of intelligent scheduling technology, specifically to a mining operation scheduling system based on big data, including an operation feature construction module, an intensity label generation module, a density mutation identification module, a rhythm type determination module, and an operation time sequence sorting module. The present invention forms a load feature distribution structure between multiple mining areas by constructing transportation equipment operation behavior data and extracting the mean combination of load change and residence time, and realizes the division of operation intensity labels between mining areas by using the similarity measurement mechanism between normalized feature pairs, which can realize the quantitative expression of the operation load distribution state, further combines the equipment trajectory information with the distribution sequence to construct the mining area density time series, determines the operation rhythm change node by identifying the density mutation point, and classifies the time period rhythm type based on the dual variables of operation intensity and operation density, thereby accurately dividing the mining area operation rhythm type within different time periods.

Description

Technical Field

[0001] The present invention relates to the field of intelligent scheduling technology, and in particular to a mining operation scheduling system based on big data. Background Art

[0002] The field of intelligent scheduling technology mainly involves the relevant technologies of optimizing the configuration and dynamic adjustment of job tasks through computer systems under multi-resource, multi-task and multi-constraint conditions.

[0003] Among them, mine operation scheduling refers to the time and resource arrangement for mining, transportation, loading and unloading and other operations within the mine area, and realizes operation scheduling and resource allocation through static rule calculation.

[0004] Existing technologies rely solely on static rules for scheduling and resource allocation, lacking real-time summarization and modeling of actual behavioral data from multi-source transportation equipment. This makes it difficult to reflect the load differences and rhythm fluctuations exhibited by different mining areas in actual operations. Dynamic factors such as sudden load changes or changes in equipment density cannot be captured in a timely manner when the operating rhythm changes. For example, when a mining area experiences a sudden high-frequency event of concentrated equipment residence or unloading, static rules fail to respond to changes in equipment density within the time period, resulting in a disconnect between scheduling results and actual on-site operations, affecting resource allocation efficiency, and even causing scheduling conflicts or transportation delays, restricting the practical value and intelligent adaptability of the scheduling system in a dynamic environment. Summary of the Invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and propose a mining operation scheduling system based on big data.

[0006] In order to achieve the above purpose, the present invention adopts the following technical solution: a mining operation scheduling system based on big data, the system includes:

[0007] The operation feature construction module obtains big data on the operation behavior of multi-source transportation equipment, including statistics on the load changes and residence time of transportation equipment in each mining area, and constructs an operation load feature set;

[0008] An intensity label generation module takes the operation load feature set as input features, constructs the load feature space of each mining interval through a spectral clustering algorithm, and divides the corresponding operation intensity labels to obtain an operation intensity label set;

[0009] The density mutation identification module obtains the entire trajectory of the mine scheduling and the distribution records of each transportation equipment in the mine, constructs the mine density time series, identifies the density mutation points in the sequence, and obtains the operation density mutation point sequence;

[0010] The rhythm type determination module divides the rhythm type of each mining area operation time period based on the operation intensity label set and the operation density mutation point sequence to obtain an operation rhythm division result.

[0011] The improvements of the present invention are that the workload feature set includes the number of data dimensions and the number of regional features; the workload intensity label set includes the number of label categories, intensity division standards, and label allocation rules; the workload density mutation point sequence includes the number of mutation points, identification time range, and mutation change amplitude; and the workload rhythm division results include the number of rhythm segments, rhythm distribution type, and rhythm level identification.

[0012] The present invention is improved in that the operation feature construction module includes:

[0013] The data collection submodule obtains the transport equipment number, mining area number, timestamp sequence, cargo load value sequence, and transport equipment positioning information. It classifies the transport equipment records according to the mining area number, extracts the cargo load changes and residence time of the transport equipment in the mining area, and generates mining area operation behavior data.

[0014] The feature calculation submodule calls the mining area operation behavior data, calculates the average value of the cargo load change and residence time of the transportation equipment in the mining area, and organizes the two mean values ​​corresponding to the mining area number to generate an operation load value pair;

[0015] The feature normalization submodule calls the workload value pair, performs normalization processing on the mean value of cargo load change and the mean value of residence time corresponding to each mining area, and generates a workload feature set.

[0016] The present invention is improved in that the intensity label generation module includes:

[0017] The spatial construction submodule constructs a two-dimensional feature combination that characterizes the differences between mining areas based on the mean value of the mining area load change and the mean value of the residence time in the workload feature set, and establishes a mining area load feature distribution structure based on the feature combination of all mining areas;

[0018] The clustering submodule, based on the distribution structure of the mining area load characteristics, regards each mining area as a node in the feature space, uses the Euclidean distance between mining area feature pairs as the basis for similarity between nodes, constructs a similarity connection graph for clustering processing, groups the mining area nodes using the spectral clustering algorithm, and generates mining area load clustering results;

[0019] The label generation submodule sets the operation intensity level range according to the mining area load clustering results and the combined level of the normalized mean value of cargo load change and the mean value of residence time corresponding to each mining area, and generates an operation intensity label set.

[0020] The present invention is improved in that the density mutation recognition module includes:

[0021] The trajectory acquisition submodule obtains the transportation equipment trajectory data from the mining area scheduling records and the distribution records of the transportation equipment in the mining area, classifies and organizes the appearance of the transportation equipment in each mining area in chronological order, and generates a mining area equipment distribution trajectory set;

[0022] The sequence construction submodule calls the mining area equipment distribution trajectory set, calculates the number of transportation equipment in each mining area in each time period based on the time axis, arranges all equipment quantity records in chronological order, and generates a mining area density time series;

[0023] The mutation identification submodule calls the mining area density time series, calculates the slope of the density value change direction according to adjacent time periods, makes a joint judgment based on the local density standard deviation ratio, identifies the trend mutation position in the sequence based on the judgment result, uses the CUSUM cumulative sum control chart algorithm to determine the mutation node, and generates an operation density mutation point sequence.

[0024] The present invention is improved in that the rhythm type determination module includes:

[0025] The label matching submodule calls the operation intensity label set, matches each operation intensity label to the corresponding mining area according to the mining area number, identifies the intensity information for rhythm identification in each mining area, and generates a mining area intensity label mapping result;

[0026] The time division submodule calls the mining area intensity label mapping result and the operation density mutation point sequence, identifies adjacent mutation time nodes in each mining area, divides the time period between adjacent time nodes into continuous operation time periods, organizes the start and end time of each time period and the mining area number to which it belongs, and generates a mining area operation time period structure;

[0027] The type determination submodule calls the mining area operation time period structure, combines the intensity label of the corresponding mining area with the density change direction in each time period, determines the rhythm type of each time period according to the rhythm classification rules, and generates the operation rhythm division result.

[0028] The present invention is improved in that it further includes an operation time sequence sorting module, which optimizes and sorts the priorities of the operation time segments between each mining area according to the operation rhythm division result to obtain the mining area operation scheduling result;

[0029] The mining area operation scheduling result includes a time sequence number, a scheduling coverage range, and a sorting priority level.

[0030] The present invention is improved in that the job timing sorting module includes:

[0031] The weight construction submodule calls the operation rhythm division result, extracts the rhythm type information of each mining area operation time segment, obtains the time priority weight parameter in the scheduling plan, sets the scheduling scoring standard corresponding to the rhythm level, and generates the operation rhythm weight parameter set;

[0032] The priority calculation submodule calls the operation rhythm weight parameter set, performs score calculation on each mining area operation time segment according to the rhythm type and time priority weight, and generates mining area time segment priority information;

[0033] The sorting output submodule calls the priority information of the mining area time segment, performs scheduling order sorting according to the priority value, uses the path selection algorithm to process the sorting logic, combines all mining area operation time segments in the sorting order, and generates the mining area operation scheduling result.

[0034] Compared with the prior art, the advantages and positive effects of the present invention are:

[0035] In the present invention, by constructing the operation behavior data of transportation equipment and extracting the mean combination of load change and residence time, a load characteristic distribution structure is formed between multiple mining areas, and the operation intensity label division between mining areas is realized by the similarity measurement mechanism between normalized feature pairs, which can realize the quantitative expression of the operation load distribution state, and further combine the equipment trajectory information and distribution sequence to construct the mining area density time series, and determine the operation rhythm change node by identifying the density mutation point. On the basis of the dual variables of operation intensity and operation density, the time period rhythm type is classified, so as to accurately divide the mining area operation rhythm type in different time periods, and establish a scheduling sorting scoring rule based on the rhythm distribution and priority weight parameters of each time segment, finally realizing the optimization of operation timing sorting, improving the rhythm matching degree and priority rationality of scheduling arrangements, and enhancing the responsiveness of multi-mining area scheduling strategies to dynamic operation behaviors and the rhythm coordination of resource allocation. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is a system module diagram of the present invention;

[0037] Figure 2 It is a system framework diagram of the present invention;

[0038] Figure 3 This is a schematic diagram of the operation feature construction module of the present invention;

[0039] Figure 4 This is a schematic diagram of the intensity label generation module of the present invention;

[0040] Figure 5 Schematic diagram of the density mutation recognition module of the present invention;

[0041] Figure 6Schematic diagram of a rhythm type determination module of the present invention;

[0042] Figure 7 Schematic diagram of the job timing sorting module of the present invention. DETAILED DESCRIPTION

[0043] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0044] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", etc., indicating directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings and are only for the convenience of describing the present invention and simplifying the description. They do not indicate or imply that the devices or elements referred to must have a specific direction, be constructed and operate in a specific direction, and therefore should not be understood as limiting the present invention. In addition, in the description of the present invention, the meaning of "plurality" is two or more, unless otherwise clearly and specifically defined.

[0045] See also Figure 1 The present invention provides a technical solution: a mining operation scheduling system based on big data, the system includes:

[0046] The operation feature construction module obtains big data on the operation behavior of multi-source transportation equipment, including statistics on the load changes and residence time of transportation equipment in each mining area, and constructs an operation load feature set;

[0047] The intensity label generation module takes the operation load feature set as input features, constructs the load feature space of each mining interval through the spectral clustering algorithm, and divides the corresponding operation intensity labels to obtain the operation intensity label set;

[0048] The density mutation identification module obtains the entire trajectory of the mine scheduling and the distribution records of each transportation equipment in the mine, constructs the mine density time series, identifies the density mutation points in the sequence, and obtains the operation density mutation point sequence;

[0049] The rhythm type determination module divides the rhythm type of each mining area's operating time period based on the operating intensity label set and the operating density mutation point sequence, and obtains the operating rhythm classification result;

[0050] The workload feature set includes the number of data dimensions and the number of regional features; the workload intensity label set includes the number of label categories, intensity division criteria, and label allocation rules; the workload density mutation point sequence includes the number of mutation points, identification time range, and mutation change amplitude; the workload rhythm division results include the number of rhythm segments, rhythm distribution type, and rhythm level identification.

[0051] See also Figure 2 and Figure 3 , the job feature building modules include:

[0052] The data collection submodule obtains the transport equipment number, mining area number, timestamp sequence, cargo load value sequence, and transport equipment positioning information. It classifies the transport equipment records according to the mining area number, extracts the cargo load changes and residence time of the transport equipment in the mining area, and generates mining area operation behavior data.

[0053] First, call the transport equipment sensor and dispatch platform data, extract the transport equipment number, positioning information, timestamp, mining area number and cargo load value one by one, confirm whether it contains clear mining area number and positioning data for each data, filter and delete the data with missing positioning or wrong number, and then sort in ascending order by equipment number and timestamp, group the entries whose positioning coordinates fall within the same mining area number range in the continuous records, and define the single operation cycle of the equipment in the mining area. Then extract the load value change in each cycle, calculate the instantaneous point of loading start and unloading completion, and compare the records before and after. The load value is used to identify the moments of significant load increase and decrease. For example, if the load jumps from 0 tons to 48 tons, the point can be determined to be the loading point. Combined with the timestamp difference between the loading point and the unloading point, the residence time is calculated. For example, if loading occurs at 10:12:30 and unloading occurs at 10:24:15, the residence time is 11 minutes and 45 seconds. The load change value of each operation cycle is the maximum value minus the minimum value. For example, if the load in this cycle goes from 0 to 48 tons, the change value is 48 tons. Finally, all equipment records are classified according to the mining area number, and the single load change and corresponding residence time of each equipment in each mining area are sorted out.

[0054] The feature calculation submodule calls the mining area operation behavior data, calculates the average value of the cargo load change and residence time of the transportation equipment according to the mining area, and organizes the two mean combinations corresponding to the mining area number to generate the operation load value pair;

[0055] First, read the operation behavior dataset generated by the data acquisition submodule. For each mining area number, retrieve all transportation equipment operation records belonging to that number. Within this subset, classify by equipment number and extract all load change values ​​and residence time values ​​for each equipment in that mining area. Then calculate the mean, where the load change mean is the sum of the load differences of all equipment operation cycles divided by the number of cycles. For example, if equipment A has five operation cycles in mining area K1, and its load changes are 48, 50, 47, 49, and 48 tons respectively, the mean is 48.4 tons. Similarly, the mean residence time is the sum of all residence times of the equipment in the mining area divided by the number of operations. If the residence time of the above cycles is 12, 13, 12, 11, and 14 minutes, the mean is 12.4 minutes. Then, combine the two means calculated for all equipment in the mining area to describe the average operating load status corresponding to the mining area number. The result for each mining area is a value pair containing "average load change" and "average residence time".

[0056] The feature normalization submodule calls the workload value pair and performs normalization processing on the mean change of cargo load and the mean residence time corresponding to each mining area to generate a workload feature set;

[0057] Receive the workload value pairs of all mining areas, that is, the mean load change and the mean residence time of each mining area. First, count the maximum and minimum values ​​of these two dimensions for all mining areas. For example, if the maximum load change is 52 tons and the minimum is 42 tons, then normalize the mean load change of any mining area: if a mining area is 48 tons, the normalized value is (48-42) / (52-42)=0.6. Similarly, process the mean residence time. If the maximum residence time is 15 minutes and the minimum is 9 minutes, the normalized value of a mining area with a residence time of 12 minutes is (12-9) / (15-9)=0.5. Finally, obtain the workload feature set of all mining areas, each item is a normalized (load change value, residence time value) combination.

[0058] See also Figure 2 and Figure 4 , the intensity label generation module includes:

[0059] The spatial construction submodule constructs a two-dimensional feature combination that characterizes the differences between mining areas based on the mean load change and the mean residence time of the mining areas in the workload characteristic cluster, and establishes the mining area load characteristic distribution structure based on the feature combination of all mining areas;

[0060] First, read each set of characteristic value pairs according to the mining area number. The pair consists of the normalized load change mean and the residence time mean. Each set of values ​​represents the standardized operating load characteristics of a mining area. The characteristic combination is used as a two-dimensional coordinate point representation. Each point constitutes a position node in the characteristic space. Let The normalized feature combination of a mining area is represented by the coordinate point ,in represents the mean value of the normalized load change in the mining area, It represents the normalized mean residence time of the mining area. For example, if the original load change mean of mining area K1 is 48 tons and the maximum and minimum range is 42 tons to 52 tons, then its normalized value is:

[0061] ;

[0062] For example, if the average K1 dwell time is 12 minutes and the maximum and minimum intervals are 9 to 15 minutes, then:

[0063] ;

[0064] Therefore, the point of K1 in the two-dimensional feature space is Then calculate the two-dimensional coordinate points of all mining areas in turn, and form a point set of all points ,in Represents the number of mining areas. After completing the mapping of all feature points, pairwise distance calculation is performed to construct a spatial structure relationship diagram. Euclidean distance is used as the similarity measurement method between two mining areas. The mining area and The characteristic points of the mining areas are and , then the Euclidean distance between them The formula is:

[0065] ;

[0066] in, : No. The mean change of normalized load in each mining area, : No. The average normalized residence time of each mining area, : No. The mean change of normalized load in each mining area, : No. The average normalized residence time of each mining area, : No. The first and The characteristic space distance between mining areas.

[0067] For example, suppose the feature points of K1 are , the point of K2 is , then the Euclidean distance between the two is: ;

[0068] This calculation is performed for each pair of mining areas, resulting in a A symmetric matrix with the main diagonal elements , indicating that the distance between itself is zero, and the non-diagonal elements represent the degree of load difference between different mining areas. The characteristic spatial structure grid diagram between mining areas is constructed through this distance matrix. The nodes in the grid are mining areas, and the edge weights are the distance values ​​between nodes.

[0069] The clustering submodule, based on the distribution structure of mining area load characteristics, regards each mining area as a node in the feature space, uses the Euclidean distance between mining area feature pairs as the basis for similarity between nodes, constructs a similarity connection graph for clustering processing, groups mining area nodes using the spectral clustering algorithm, and generates mining area load clustering results;

[0070] First, extract the coordinate points of all mining areas in the normalized feature space , build contains The feature space graph structure of each node represents a mining area. Similarity edges are generated between any two nodes based on their feature distances. The similarity is calculated using Euclidean distance. The smaller the distance, the closer the features are. An undirected weighted graph is initially constructed. ,in It is a collection of mining areas. is an edge set, the edge weight is determined by the similarity value between nodes, and the similarity function is set to the Gaussian kernel function:

[0071] ;

[0072] in, :Indicates the The mining area and The similarity weight value between mining areas is in the range of ; :Indicates the With the Euclidean distance between mining areas; : is the scale control parameter of the Gaussian kernel function, which determines the degree of attenuation of distance to similarity. It is recommended to set it to all For example, if , is 0.3, then: .

[0073] Repeat this calculation for all pairs of mining areas to obtain the similarity matrix ,in, : is the symmetric similarity matrix between mining areas, with dimension , each element represents the characteristic similarity between two mining areas; :Indicates the first Rank The elements of the column, i.e. the mining area With mining area Similarity value of : Indicates traversing all mining area numbers.

[0074] Then construct the degree matrix , where each element is the sum of the similarities between the corresponding mining area and all other mining areas: : is a diagonal matrix with dimension , used to record the total connectivity weight of each mining area; :Indicates the The degree value of a mining area in the similarity graph is the sum of its edge weights with all other mining areas; the summation symbol :For all The similarity of each mining area is summed up. Indicates the total number of mining areas involved in the calculation.

[0075] After the calculations are complete, the Laplacian matrix is ​​constructed. This process involves subtracting each element in the similarity matrix from each main diagonal element in the degree matrix. In other words, the diagonal value in a row represents the total similarity of the mining area, which is then paired and subtracted from the mining area similarity values ​​represented by each column in the row to obtain each element of the new matrix, ultimately forming a symmetrical Laplacian matrix.

[0076] The matrix is ​​input into the spectral clustering algorithm, and the process is as follows:

[0077] 1. Perform eigenvalue decomposition on the constructed Laplace matrix and extract The eigenvectors corresponding to the smallest eigenvalues ​​are concatenated into a characteristic matrix ,in, : represents the feature matrix after dimensionality reduction; : represents the total number of mining areas, the matrix has OK; : represents the number of target clusters, the matrix has List; :Indicates the Feature representation of each mining area in dimensionality reduction space;

[0078] 2. Matrix Each row vector in is normalized to form a matrix ,in, : is the standardized feature matrix, with the same dimension ; Each row is processed as a unit vector to eliminate the amplitude difference; 3. Each row of is used as an input sample and clustered using the Kmeans clustering algorithm: The input data is , that is, the characteristic vector of each mining area; the number of clusters It means that all mining areas are divided into three types of operating load groups.

[0079] If there is a total mining areas, numbered from K1 to K6, the calculated similarity matrix for:

[0080] ;

[0081] Among them, the first row represents the similarity between K1 and the rest of the mining areas, and the first column value 0 indicates no connection with itself; the second row represents the similarity between K2 and other mining areas; the remaining rows are represented by ellipsis, following the symmetry of the matrix; the matrix obtained by spectral clustering is After inputting the Kmeans algorithm for clustering, the following clustering results might be obtained: K1 and K2 are classified as the first category, K3 and K4 are classified as the second category, and K5 and K6 are classified as the third category. Ultimately, each mining area is assigned a clear category number, indicating the workload cluster category to which it belongs, forming a complete mining area load clustering result.

[0082] The label generation submodule sets the operation intensity level range based on the mining area load clustering results and the normalized combined level of the mean value of cargo load change and the mean value of residence time corresponding to each mining area, and generates an operation intensity label set;

[0083] After receiving the mining area load clustering results, firstly, according to the mining area numbers included in each category, the normalized load change mean and normalized residence time mean of the corresponding mining area are extracted one by one to form the feature value set under the category. The average value of the two features is calculated for each category. For example, the first category contains 3 mining areas, and their normalized load change means are 0.70, 0.60, and 0.65, respectively, and their residence time means are 0.40, 0.35, and 0.38, respectively. The average load change of this category is The average residence time is , repeat the same steps for all cluster categories to obtain the representative feature pairs of each category, and then set the work intensity level classification standard according to the interval position of the combination mean. The specific classification rules are as follows: when the normalized load change mean ≥ 0.66 and the residence time mean ≤ 0.33, it is classified as "high intensity"; when the load change mean ∈ [0.33, 0.66) and the residence time mean ∈ (0.33, 0.66], it is classified as "medium intensity"; when the load change mean < 0.33 and the residence time mean > 0.66, it is classified as "low intensity"; the remaining boundary combinations are approximately classified according to the degree of proximity, and finally each cluster category is assigned a corresponding intensity level label to form an work intensity label set, such as category 1 is "high intensity", category 2 is "medium intensity", category 3 is "low intensity", etc., to achieve a clear conversion from clustering results to intensity identification.

[0084] See also Figure 2 and Figure 5 , the density mutation recognition module includes:

[0085] The trajectory acquisition submodule obtains the transportation equipment trajectory data from the mining area scheduling records and the distribution records of the transportation equipment in the mining area, classifies and organizes the appearance of the transportation equipment in each mining area in chronological order, and generates a mining area equipment distribution trajectory set;

[0086] First, read the historical transportation equipment scheduling records from the mine scheduling platform, which include the unique number of the transportation equipment, positioning time, mine identification and trajectory coordinate information. After sorting all equipment records in ascending order by the time field, classify the equipment that appears in the same mine during the same time period. For example, at a certain time point of 10:00, the trajectory points of equipment T1, T2, and T5 appear in mine A, then it is considered that there are 3 equipment stationed in mine A at that time. Then, by comparing the trajectory records at consecutive moments, the change behavior of the equipment entering or leaving the mine is identified. Construct the equipment presence status of each mining area in each time slice, then aggregate them by mining area number, and record the appearance of equipment in each mining area in different time periods in a sequential manner. For example: Mining area B has 2, 3, and 2 devices at 10:00, 10:05, 10:10, etc., then the equipment distribution trajectory of this mining area is a time-sequential recursive equipment number trajectory line. Repeat the above steps and process all mining areas to finally form a mining area equipment distribution trajectory set. Each mining area corresponds to a distribution trajectory line consisting of the number of equipment and time points.

[0087] The sequence construction submodule calls the mining area equipment distribution trajectory set, calculates the number of transportation equipment in each mining area within each time period based on the time axis, arranges all equipment quantity records in chronological order, and generates a mining area density time series;

[0088] After reading the set of equipment distribution trajectories in the mining area, all time points are divided into uniform time intervals, for example, every 5 minutes is a time period. The equipment trajectory data recorded in all mining areas in each time period are traversed one by one, and the number of equipment in each mining area in the time period is counted. For example, in the time period from 10:00 to 10:05, equipment numbered T2, T3, and T4 appeared in mining area C, and the number of equipment was 3. In the next period from 10:05 to 10:10, T3 and T5 appeared, and the number of equipment was 2. The equipment values ​​counted in each time period are arranged in chronological order to form a numerical sequence. This sequence is the operation density time series of mining area C. This process is repeated for all mining areas, and finally multiple mining area density time series are generated. Each sequence reflects the change process of the density of transportation equipment in the mining area in a continuous time period.

[0089] The mutation identification submodule calls the mining area density time series, calculates the slope of the density value change direction according to adjacent time periods, and makes a joint judgment based on the local density standard deviation ratio. Based on the judgment results, it identifies the trend mutation location in the sequence, uses the CUSUM cumulative sum control chart algorithm to determine the mutation node, and generates an operation density mutation point sequence;

[0090] After receiving the density time series of each mining area, we first perform difference processing on the adjacent time periods in each sequence in chronological order, calculate the change amplitude and direction of the number of transport equipment per unit time, and form a change slope sequence. Assuming the time interval is 5 minutes, each slope can be understood as the difference in the change of the number of equipment between two time points divided by the time interval. For example, if the number of equipment in a mining area is 5 at 10:00 and 3 at 10:05, the slope is , recorded as a negative change; then, using three adjacent time periods as a sliding window, the standard deviation of the number of devices in each window is calculated and compared with the mean of the sequence segment where the current window is located. This ratio is called the local density fluctuation rate, which is used to measure whether the local fluctuation is significant.

[0091] Then, we enter the mutation detection phase and use the CUSUM (CumulativeSum) algorithm to determine the cumulative deviation. The core formula is as follows:

[0092] ;

[0093] in, :Indicates the CUSUM cumulative sum at the moment; :The previous moment The cumulative sum of CUSUM is initially set to 0; :Indicates the The number of devices in each time period; : represents the average number of devices in the early stable period of the time series; : is the sensitivity coefficient, usually taken as half of : Used to reset the accumulated value to zero to avoid misjudgment caused by long-term accumulation of small fluctuations.

[0094] Assume that the average number of equipment in the first five time periods of a mining area is , then suppose , starting from the sixth segment, the number of devices in the sixth segment is 8, then:

[0095] ;

[0096] The number of devices in the seventh segment is 7, and the number continues to accumulate:

[0097] ;

[0098] If after this Continuously increasing and exceeding the threshold , for example, , then determine the The moment is the density mutation point. It is usually determined by the experience of the maximum single point deviation in the stable sequence. For example, if the maximum deviation in the first five segments is 1.5, then Round up to 5. This algorithm, combined with the directional consistency of the slope sign and the local standard deviation ratio, can identify the location of trend mutations and ultimately output a sequence of activity density mutation points, recording the specific time period index location where the mutation occurred.

[0099] See also Figure 2 and Figure 6 , the rhythm type determination module includes:

[0100] The label matching submodule calls the operation intensity label set, matches each operation intensity label to the corresponding mining area according to the mining area number, identifies the intensity information used for rhythm identification in each mining area, and generates the mining area intensity label mapping result;

[0101] Read the operation intensity label set, where each record contains an intensity level label and the mine area number to which it belongs. The label content is annotated to the corresponding mine area through a matching operation, that is, the mine area number is used as the index field, and the intensity label is attached to the mine area data structure as an attribute value. For example: if the mine area number is K03, it obtains the "high intensity" label after cluster analysis, then the label will be bound to K03 in this module, identifying K03 as a high-intensity operation area. This operation performs a one-time matching on all mine areas to form a collection result containing each mine area number and its intensity information, which is used in the subsequent rhythm recognition process to finally generate the mine area intensity label mapping result.

[0102] The time division submodule calls the mining area intensity label mapping results and the operation density mutation point sequence, identifies adjacent mutation time nodes in each mining area, divides the time period between adjacent time nodes into continuous operation time periods, organizes the start and end times of each time period and the mining area number to which it belongs, and generates the mining area operation time period structure;

[0103] Read the mining area intensity label mapping results and the operation density mutation point sequence, process each mutation sequence separately according to the mining area number, pair the mutation nodes of each mining area in chronological order, and divide the time period between every two adjacent mutation nodes into a continuous operation cycle. For example: if there are mutation point times 10:00, 10:25, and 10:55 in the density mutation sequence of mining area K05, then the divided operation time periods are 10:00 to 10:25 and 10:25 to 10:55 respectively. Each segment is bound to the mining area number K05 as the belonging identifier. After sorting, it is summarized into a set of structured data records. The fields contain start time, end time and mining area number. All mining area data are combined to form the mining area operation time period structure.

[0104] The type determination submodule calls the mining area operation time period structure, combines the intensity label of the corresponding mining area with the density change direction in each time period, and determines the rhythm type of each time period according to the rhythm classification rules to generate the operation rhythm classification result;

[0105] The system receives the mining area operation time period structure and combines the intensity label of the mining area to which the time period belongs with the operation density change trend within the section to identify the rhythm type. For each operation time data, the intensity label matching its mining area number is first retrieved, and then the density slope value of the time period is calculated to determine whether the density is rising, falling or stable within the period. For example, if the number of equipment in the section increases from 5 to 9 and it is a "high-intensity" mining area, it can be judged as a "rapid growth" rhythm; on the contrary, if the number of equipment decreases from 8 to 4, it can be judged as an "intensity release" rhythm; if the number of equipment basically remains at 6 and fluctuates by no more than 1, it is judged as "stable operation"; for medium-intensity mining areas, a relatively mild fluctuation judgment interval is applied. If the density increase is less than 2 units, it is still considered "stable"; if there is no obvious trend change in low-intensity mining areas, it is marked as an "inefficient operation" rhythm type. The system completes the rhythm type classification of all operation time periods based on this judgment rule.

[0106] See also Figure 2 and Figure 7 ,It also includes an operation time sequence sorting module, which optimizes and sorts the priority of the operation time segments between each mining area according to the operation rhythm division results, and obtains the mining area operation scheduling results;

[0107] The mining area operation scheduling results include time sequence number, scheduling coverage, and sorting priority level;

[0108] The job timing sequencing module includes:

[0109] The weight construction submodule calls the operation rhythm division results, extracts the rhythm type information of each mining area operation time segment, obtains the time priority weight parameters in the scheduling plan, sets the scheduling scoring standards corresponding to the rhythm level, and generates the operation rhythm weight parameter set;

[0110] The system reads the work rhythm classification results and extracts rhythm type information based on the mining area number corresponding to each work time segment. It also loads the time priority weight configuration corresponding to each time segment from the scheduling system parameter table. For example, the schedule table defines 06:00 to 10:00 as the "high-efficiency segment" with a priority weight of 3, 10:00 to 16:00 as the "normal segment" with a priority weight of 2, and 16:00 to 20:00 as the "low-pressure segment" with a priority weight of 1. A correlation is then established between rhythm types and scheduling scores. Based on on-site work analysis and historical task feedback, the following scoring criteria are established: when equipment density increases rapidly and is accompanied by a high-intensity label, the rhythm is classified as "rapid growth" with a basic score of 5 points; if equipment density remains stable overall without significant fluctuations, it is classified as "stable operation" with a basic score of 3 points; and when equipment density continues to decrease, indicating a withdrawal trend, it is classified as "intensity release" with a basic score of 2 points. This scoring rule is based on the mapping of work rhythm to scheduling urgency. Rapid growth indicates resource concentration and requires a quick response, while intensity release indicates a reduced load and allows for appropriate delays.

[0111] The priority calculation submodule calls the operation rhythm weight parameter set, calculates the score of each mining area operation time segment based on the rhythm type and time priority weight, and generates the mining area time segment priority information;

[0112] After receiving the scoring criteria and weights, a score is calculated for each time period. This calculation uses a weighted product method, multiplying the basic rhythm score by the time priority weight to form a comprehensive score. For example, if a mining area is identified as having a "rapid growth" rhythm from 7:30 AM to 8:30 AM, with a corresponding basic score of 5 and a "high efficiency" period with a time weight of 3, the final score is 5 times 3, which equals 15. Another mining area is identified as having a "stable operation" rhythm from 11:00 AM to 12:00 PM, with a basic score of 3 and a time weight of 2, resulting in a final score of 6. A period identified as having an "intensity release" rhythm with a basic score of 2 and a time weight of 3 also results in a final score of 6. To prevent either the rhythm score or the time weight from dominating the scoring, a maximum score of 20 is set, with any excess score clipped. If two time periods have the same final score, the start times are compared, with the earlier start being prioritized. If the start times match, the durations are compared, with the shorter duration being prioritized. The priority calculation process is uniformly executed in all time segments to generate complete mine area time segment priority information. The record fields include mine area number, time period start and end, rhythm type, time weight, basic score and final priority score.

[0113] The sorting output submodule calls the priority information of the mining area time segment, performs scheduling order sorting according to the priority value, uses the path selection algorithm to process the sorting logic, combines all the mining area operation time segments in the sorting order, and generates the mining area operation scheduling result;

[0114] After reading the above priority information, all time segments are sorted from high to low according to the final score value to form a preliminary scheduling execution order. During the sorting process, if there are multiple time segments with the same score, they are fine-tuned according to the backup order rules. For example, the time segment with an earlier job start time is prioritized. If the distinction is still not possible, the segment with a shorter job duration is selected as the priority to reduce the spread of scheduling delays. After the sorting is completed, the connection relationship between each time segment is integrated, and the path selection algorithm framework is used to evaluate the physical connection order of the time segment combination. For example, when there is a large spatial distance between two high-priority segments or equipment switching is difficult, the execution order is temporarily adjusted according to the system rules or other low-priority segments are inserted to fill the gaps in the scheduling path. Finally, all mining areas and all operation time segments are integrated into a scheduling execution path, indicating the scheduling order and the operation time, rhythm type and mine area number of each segment. The final generated mining area operation scheduling result is submitted to the execution system for use.

[0115] The above are merely preferred embodiments of the present invention and do not limit the present invention in any other form. Any technician familiar with the profession may use the technical content disclosed above to change or modify it into an equivalent embodiment with equivalent changes and apply it to other fields. However, any simple modification, equivalent change and modification made to the above embodiment based on the technical essence of the present invention without departing from the content of the technical solution of the present invention shall still fall within the scope of protection of the technical solution of the present invention.

Claims

1. A mining operation scheduling system based on big data, characterized by: The system comprises: The operation feature construction module obtains big data on the operation behavior of multi-source transportation equipment, including statistics on the load changes and residence time of transportation equipment in each mining area, and constructs an operation load feature set; An intensity label generation module takes the operation load feature set as input features, constructs the load feature space of each mining interval through a spectral clustering algorithm, and divides the corresponding operation intensity labels to obtain an operation intensity label set; The density mutation identification module obtains the entire trajectory of the mine scheduling and the distribution records of each transportation equipment in the mine, constructs the mine density time series, identifies the density mutation points in the sequence, and obtains the operation density mutation point sequence; A rhythm type determination module, based on the operation intensity label set and the operation density mutation point sequence, divides the rhythm type of each mining area operation time period to obtain an operation rhythm division result; The operation timing sorting module optimizes and sorts the priorities of the operation time segments between each mining area according to the operation rhythm division result to obtain the mining area operation scheduling result.

2. The mining operation scheduling system based on big data according to claim 1 is characterized in that: The workload feature set includes the number of data dimensions and the number of regional features; the workload intensity label set includes the number of label categories, intensity division criteria, and label allocation rules; the workload density mutation point sequence includes the number of mutation points, identification time range, and mutation change amplitude; the workload rhythm division result includes the number of rhythm segments, rhythm distribution type, and rhythm level identification.

3. The mining operation scheduling system based on big data according to claim 1 is characterized in that: The job feature construction module includes: The data collection submodule obtains the transport equipment number, mining area number, timestamp sequence, cargo load value sequence, and transport equipment positioning information. It classifies the transport equipment records according to the mining area number, extracts the cargo load changes and residence time of the transport equipment in the mining area, and generates mining area operation behavior data. The feature calculation submodule calls the mining area operation behavior data, calculates the average value of the cargo load change and residence time of the transportation equipment in the mining area, and organizes the two mean values ​​corresponding to the mining area number to generate an operation load value pair; The feature normalization submodule calls the workload value pair, performs normalization processing on the mean value of cargo load change and the mean value of residence time corresponding to each mining area, and generates a workload feature set.

4. The mining operation scheduling system based on big data according to claim 1 is characterized in that: The intensity label generation module includes: The spatial construction submodule constructs a two-dimensional feature combination that characterizes the differences between mining areas based on the mean value of the mining area load change and the mean value of the residence time in the workload feature set, and establishes a mining area load feature distribution structure based on the feature combination of all mining areas; The clustering submodule, based on the distribution structure of the mining area load characteristics, regards each mining area as a node in the feature space, uses the Euclidean distance between mining area feature pairs as the basis for similarity between nodes, constructs a similarity connection graph, groups the mining area nodes through the spectral clustering algorithm, and generates mining area load clustering results; The label generation submodule sets the operation intensity level range according to the mining area load clustering results and the combined level of the normalized mean value of cargo load change and the mean value of residence time corresponding to each mining area, and generates an operation intensity label set.

5. The mining operation scheduling system based on big data according to claim 4 is characterized in that: To construct the similarity connection graph, the formula is used: ； The Euclidean distances between the normalized feature coordinates of all mining areas are converted into similarity weights using the Gaussian kernel function. Weighted edges are constructed based on the weights, with nodes representing mining areas and edges representing similarities, to obtain a similarity connection graph. in, Indicates the The mining area and The similarity weight between mining areas, the value range ; Indicates the With the The Euclidean distance between mining areas, is the scale control parameter of the Gaussian kernel function, which determines the attenuation degree of distance to similarity. of the median.

6. The mining operation scheduling system based on big data according to claim 1 is characterized in that: The density mutation recognition module includes: The trajectory acquisition submodule obtains the transportation equipment trajectory data from the mining area scheduling records and the distribution records of the transportation equipment in the mining area, classifies and organizes the appearance of the transportation equipment in each mining area in chronological order, and generates a mining area equipment distribution trajectory set; The sequence construction submodule calls the mining area equipment distribution trajectory set, calculates the number of transportation equipment in each mining area in each time period based on the time axis, arranges all equipment quantity records in chronological order, and generates a mining area density time series; The mutation identification submodule calls the mining area density time series, calculates the slope of the density value change direction according to adjacent time periods, makes a joint judgment based on the local density standard deviation ratio, identifies the trend mutation position in the sequence based on the judgment result, uses the CUSUM cumulative sum control chart algorithm to determine the mutation node, and generates an operation density mutation point sequence.

7. The mining operation scheduling system based on big data according to claim 6 is characterized in that: To determine the mutation node, the formula is used: ； in, Indicates the The cumulative sum of CUSUM at the moment, For the previous moment The CUSUM cumulative sum is initially set to 0. Indicates the The number of devices in a time period, represents the average number of devices in the early stable period of the time series, is the sensitivity coefficient, take half, Used to reset the accumulated value to zero; By calculating the cumulative sum of the number of devices and CUSUM at each moment, when the cumulative value continuously increases and exceeds the fluctuation threshold, it can be determined that the current node is a density mutation node.

8. The mining operation scheduling system based on big data according to claim 1 is characterized in that: The rhythm type determination module includes: The label matching submodule calls the operation intensity label set, matches each operation intensity label to the corresponding mining area according to the mining area number, identifies the intensity information for rhythm identification in each mining area, and generates a mining area intensity label mapping result; The time division submodule calls the mining area intensity label mapping result and the operation density mutation point sequence, identifies adjacent mutation time nodes in each mining area, divides the time period between adjacent time nodes into continuous operation time periods, organizes the start and end time of each time period and the mining area number to which it belongs, and generates a mining area operation time period structure; The type determination submodule calls the mining area operation time period structure, combines the intensity label of the corresponding mining area with the density change direction in each time period, determines the rhythm type of each time period according to the rhythm classification rules, and generates the operation rhythm division result.

9. The mining operation scheduling system based on big data according to claim 1, characterized in that: The mining area operation scheduling result includes a time sequence number, a scheduling coverage range, and a sorting priority level.

10. The mining operation scheduling system based on big data according to claim 1, characterized in that: The job timing sorting module includes: The weight construction submodule calls the operation rhythm division result, extracts the rhythm type information of each mining area operation time segment, obtains the time priority weight parameter in the scheduling plan, sets the scheduling scoring standard corresponding to the rhythm level, and generates the operation rhythm weight parameter set; The priority calculation submodule calls the operation rhythm weight parameter set, performs score calculation on each mining area operation time segment according to the rhythm type and time priority weight, and generates mining area time segment priority information; The sorting output submodule calls the priority information of the mining area time segment, performs scheduling order sorting according to the priority value, uses the path selection algorithm to process the sorting logic, combines all mining area operation time segments in the sorting order, and generates the mining area operation scheduling result.

Images