Dynamic material loading scheduling method and system based on smart mine service

By using dynamic material loading scheduling methods in smart mining services, and using decision tree network for timeliness analysis and feasibility estimation, the problems of smart mining server linkage and strategy accuracy are solved, and the mining production efficiency and decision-making efficiency are improved.

CN118036841BActive Publication Date: 2024-11-19BEIJING UNITED MEDIA TECH CO LTD
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202410059338.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-01-15
Publication Date
2024-11-19
Estimated Expiration
2044-01-15

AI Technical Summary

Technical Problem

The linkage problems between smart mine servers, the accuracy of timeliness evaluation, and the feasibility estimation accuracy and rationality of mine material loading scheduling strategies lead to low information transmission and processing efficiency, affecting the effectiveness and production efficiency of scheduling.

Method used

Using a dynamic material loading and scheduling method based on smart mining services, by obtaining material loading and scheduling simulation data, using the first scheduling decision tree network for timeliness analysis, determining the linkage feasibility estimation results, and using the second scheduling decision tree network for transportation The service demand vector set and the policy attribute vector set are optimized to optimize the scheduling strategy and improve the feasibility estimation accuracy and rationality.

Benefits of technology

It improves the timeliness and targetedness of the mining material loading and scheduling strategy, ensures the effective implementation of the strategy in actual operations, improves the mining production efficiency and decision-making efficiency, and reduces production risks.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118036841B_ABST
    Figure CN118036841B_ABST
Patent Text Reader

Abstract

The dynamic material loading and scheduling method and system based on the smart mine service provided in the embodiment of the present application analyzes the loading and scheduling simulation data including the mine material category data, the mine material storage data and the mine production business data, and uses the decision tree network to evaluate the timeliness and feasibility of the transportation route planning information of the material loading and scheduling strategy of the smart mine server in different loading and scheduling cycles. This method can not only ensure the accuracy of the timeliness evaluation, but also mine the analytical data of the material loading and scheduling strategy of the smart mine server, thereby improving the feasibility estimation accuracy and rationality of the mine material loading and scheduling strategy. This strategy helps to improve the timeliness and pertinence of mine material loading and scheduling, and further improve the efficiency of mine production.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of data processing technology, and in particular to a dynamic material loading scheduling method and system based on smart mining services. Background Art

[0002] With the continuous development of mine automation and intelligence, how to improve the efficiency of material loading and scheduling has become an important issue. Traditional scheduling methods mainly rely on manual experience to make decisions, which may not achieve the optimal solution in complex and changeable actual operations, and it is also difficult to cope with various emergencies.

[0003] At present, many mines have begun to use smart mine servers for operation and management. However, there are still some challenges due to the linkage between the servers, the accuracy of timeliness evaluation, and the feasibility estimation accuracy and rationality of mine material loading scheduling strategies.

[0004] First, the linkage problem between smart mine servers may lead to inefficient information transmission and processing, affecting the operating efficiency of the entire system. Secondly, if the accuracy of timeliness evaluation is not high, it may cause a large deviation between the predicted results and the actual situation, thus affecting the effectiveness of scheduling. Furthermore, if the accuracy and rationality of feasibility estimation are not high, it may lead to the inability to effectively implement the formulated strategy in actual operation, thus affecting production efficiency. Summary of the invention

[0005] In order to improve the above problems, the present application provides a dynamic material loading scheduling method and system based on smart mining services.

[0006] In a first aspect, a dynamic material loading scheduling method based on smart mining services is provided, which is applied to a material loading scheduling decision system, and the method comprises:

[0007] Acquire material loading scheduling simulation data, wherein the material loading scheduling simulation data includes transportation route planning information of the target smart mine server for the mine material loading scheduling strategy in the first loading scheduling cycle;

[0008] Performing timeliness analysis on the transport route planning information in the material loading scheduling simulation data through a first scheduling decision tree network to obtain timeliness analysis data corresponding to the material loading scheduling simulation data;

[0009] Based on the timeliness analysis data corresponding to the material loading scheduling simulation data, determining a first linkage feasibility estimation result of the target smart mine server on the mine material loading scheduling strategy in the first loading scheduling cycle;

[0010] A second scheduling decision tree network is used to determine a second linkage feasibility estimation result of the mine material loading scheduling strategy by the target smart mine server in a second loading scheduling period based on the first linkage feasibility estimation result, the scheduling strategy attribute vector set of the mine material loading scheduling strategy, and the transportation service demand vector set of the target smart mine server. The second loading scheduling period is later than the first loading scheduling period.

[0011] In some possible design schemes, the second scheduling decision tree network includes an image description mining branch, an image description cross branch, and a feasibility estimation branch; the second scheduling decision tree network determines the second linkage feasibility estimation result of the target smart mine server for the mine material loading scheduling strategy in the second loading scheduling period based on the first linkage feasibility estimation result, the scheduling strategy attribute vector set of the mine material loading scheduling strategy, and the transportation service demand vector set of the target smart mine server, including:

[0012] Aggregating the first linkage feasibility estimation result, the scheduling strategy attribute vector set of the mine material loading scheduling strategy, and the transportation service demand vector set of the target smart mine server to obtain a target linkage image feature set;

[0013] Performing feature selection on the target linkage image feature set to obtain an attention image description vector sequence;

[0014] Calling the image description mining branch to perform reversible processing on the attention image description vector sequence to obtain a reversible description vector sequence;

[0015] Calling the image description cross branch to cross the attention image description vector sequence with the reversible description vector sequence to obtain a cross description vector sequence;

[0016] The feasibility estimation branch is called to perform feasibility mapping on the cross description vector sequence, and a second linkage feasibility estimation result of the target smart mine server on the mine material loading scheduling strategy in the second loading scheduling cycle is output.

[0017] In some possible designs, the debugging process of the second scheduling decision tree network includes:

[0018] Obtain a past learning sample set, the past learning sample set at least including a past transportation service demand vector set of a past smart mine server in a first past loading scheduling cycle, a past scheduling strategy attribute vector set of a past mine material loading scheduling strategy in the first past loading scheduling cycle, and a third linkage feasibility estimation result of the past smart mine server for the past mine material loading scheduling strategy in the first past loading scheduling cycle and a fourth linkage feasibility estimation result in the second past loading scheduling cycle;

[0019] The third linkage feasibility estimation result is determined based on the first material loading scheduling simulation data of the past smart mine server for the past mine material loading scheduling strategy in the first past loading scheduling cycle, and the fourth linkage feasibility estimation result is determined based on the second material loading scheduling simulation data of the past smart mine server for the past mine material loading scheduling strategy in the second past loading scheduling cycle, and the second past loading scheduling cycle is located after the first past loading scheduling cycle;

[0020] Generate a target linkage image feature set sample based on the past transport service demand vector set, the past scheduling strategy attribute vector set and the third linkage feasibility estimation result;

[0021] Inputting the target linkage image feature set sample into the second basic neural network for feasibility estimation processing, and obtaining the second linkage feasibility estimation result of the past smart mine server on the past mine material loading scheduling strategy in the second loading scheduling period;

[0022] Determining a second target training cost according to the second linkage feasibility estimation result and the fourth linkage feasibility estimation result;

[0023] Based on the second target training cost, the network parameters of the second basic neural network are improved, and loop debugging is performed until the debugging requirements are met to obtain a second scheduling decision tree network.

[0024] In some possible designs, the method further includes:

[0025] Acquire second material loading scheduling simulation data of the past intelligent mine server for the past mine material loading scheduling strategy in the second past loading scheduling period;

[0026] Performing timeliness analysis on the second material loading scheduling simulation data through the first scheduling decision tree network, and outputting timeliness analysis data corresponding to the second material loading scheduling simulation data;

[0027] Based on the timeliness analysis data corresponding to the second material loading scheduling simulation data, determine the fourth linkage feasibility estimation result of the past smart mine server on the past mine material loading scheduling strategy in the second past loading scheduling period.

[0028] In some possible design solutions, before performing timeliness analysis on the transport route planning information in the material loading scheduling simulation data through the first scheduling decision tree network and obtaining timeliness analysis data corresponding to the material loading scheduling simulation data, the method further includes:

[0029] Acquire past material loading and scheduling simulation data and corresponding past timeliness analysis data, wherein the past material loading and scheduling simulation data at least includes past transportation route planning information of past smart mine servers for past mine material loading and scheduling strategies;

[0030] Entering the past material loading scheduling simulation data into the pre-debugged first basic neural network, so as to perform timeliness analysis based on the past transportation route planning information in the past material loading scheduling simulation data to obtain a timeliness prediction result;

[0031] Determining a first target training cost according to the past timeliness analysis data and the timeliness prediction result;

[0032] Based on the first target training cost, the network parameters of the pre-debugging first basic neural network are improved, and cyclic debugging is performed until the debugging requirements are met to obtain a first scheduling decision tree network.

[0033] In some possible design schemes, the timeliness analysis data includes the confidence weight of the material loading scheduling simulation data belonging to the target timeliness level annotation; the timeliness analysis data corresponding to the material loading scheduling simulation data is determined based on the target smart mine server. The feasibility estimation result of the first linkage of the mine material loading scheduling strategy in the first loading scheduling period includes:

[0034] Determining the timeliness level value corresponding to the material loading scheduling simulation data based on the confidence weight that the material loading scheduling simulation data belongs to the target timeliness level annotation;

[0035] Based on the timeliness level value corresponding to the material loading scheduling simulation data, a first linkage feasibility estimation result of the target smart mine server on the mine material loading scheduling strategy in the first loading scheduling cycle is determined.

[0036] In some possible design schemes, the determining of the first linkage feasibility estimation result of the target smart mine server on the mine material loading scheduling strategy in the first loading scheduling cycle based on the timeliness level value corresponding to the material loading scheduling simulation data includes:

[0037] Obtaining quantitative weighting factors and linkage involvement factors;

[0038] The sum of the timeliness level value corresponding to the material loading scheduling simulation data and the quantitative weighting factor is used as the target weighting result;

[0039] By combining the target weighted result with the linkage involvement factor, a first linkage feasibility estimation result of the target smart mine server on the mine material loading scheduling strategy in the first loading scheduling cycle is obtained.

[0040] In some possible design schemes, after determining the second linkage feasibility estimation result of the target smart mine server for the mine material loading scheduling strategy in the second loading scheduling period based on the first linkage feasibility estimation result, the scheduling strategy attribute vector set of the mine material loading scheduling strategy, and the transportation service demand vector set of the target smart mine server through the second scheduling decision tree network, the method further includes:

[0041] Determine a smart mine server pool, wherein the smart mine server pool includes a plurality of target smart mine servers;

[0042] Generate a strategy matching tag for the mine material loading scheduling strategy according to the first linkage feasibility estimation result and the second linkage feasibility estimation result corresponding to each of the target smart mine servers;

[0043] The strategy matching tag is sent to a strategy generation system of the mine material loading scheduling strategy.

[0044] In some possible design schemes, the material loading scheduling simulation data also includes associated transport route planning information of an associated smart mine server for the mine material loading scheduling strategy in a first loading scheduling cycle, and the associated smart mine server belongs to the same server queue as the target smart mine server. The material loading scheduling simulation data is obtained, including:

[0045] Acquire vehicle transport route planning information of the target smart mine server for the mine material loading scheduling strategy in the first loading scheduling period;

[0046] Determine the server queue to which the target smart mine server belongs, and obtain the initial transport route planning information of each associated smart mine server in the server queue for the mine material loading scheduling strategy in the first loading scheduling period;

[0047] According to the initial transport route planning information of each associated smart mine server in the server queue for the mine material loading scheduling strategy in the first loading scheduling cycle, the associated transport route planning information of the associated smart mine server for the mine material loading scheduling strategy in the first loading scheduling cycle is screened.

[0048] In some possible design schemes, the initial transport route planning information of each associated smart mine server in the server queue for the mine material loading scheduling strategy in the first loading scheduling period, screening the associated transport route planning information of the associated smart mine server for the mine material loading scheduling strategy in the first loading scheduling period, includes:

[0049] Determine the involvement weight between each associated smart mine server in the server queue and the target smart mine server;

[0050] Based on the initial transportation route planning information of each associated smart mine server for the mine material loading scheduling strategy, the initial transportation route planning information of the associated smart mine server with the highest weight is selected as the associated transportation route planning information of the associated smart mine server for the mine material loading scheduling strategy in the first loading scheduling cycle.

[0051] In a second aspect, a material loading scheduling decision system is provided, comprising a processor and a memory communicating with each other, wherein the processor is used to retrieve a computer program from the memory and implement the method described in the first aspect by running the computer program.

[0052] According to a third aspect, a computer-readable storage medium is provided, on which a computer program is stored, and the computer program implements the method described in the first aspect when running.

[0053] The dynamic material loading and scheduling method and system based on smart mine services provided in the embodiment of the present application can perform timeliness evaluation through a first scheduling decision tree network for a linkage image feature set including the transport route planning information of the mine material loading scheduling strategy of the target smart mine server in the first loading scheduling cycle, so as to mine out the timeliness analysis data of the mine material loading scheduling strategy of the target smart mine server in the first loading scheduling cycle, ensure the accuracy of the timeliness evaluation, and then determine the first linkage feasibility estimation result of the mine material loading scheduling strategy of the target smart mine server in the first loading scheduling cycle based on the timeliness analysis data; finally, determine the second linkage feasibility estimation result of the mine material loading scheduling strategy of the target smart mine server in the second loading scheduling cycle based on the first linkage feasibility estimation result of the first loading scheduling cycle, the scheduling strategy attribute vector set of the mine material loading scheduling strategy, and the transport service demand vector set of the target smart mine server through a second scheduling decision tree network, so as to improve the feasibility estimation accuracy and rationality of the mine material loading scheduling strategy. In this way, by jointly analyzing the material loading and scheduling simulation data including at least mine material category data, mine material storage volume data and mine production business data, the linkage feasibility estimation results corresponding to the dynamic loading and scheduling strategy of the material are determined, which can ensure the timeliness and pertinence of the scheduling during the subsequent execution of the strategy, thereby ensuring the efficiency of mine production. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0055] Figure 1 A flow chart of a dynamic material loading scheduling method based on smart mining services provided in an embodiment of the present application. DETAILED DESCRIPTION

[0056] In order to better understand the above technical scheme, the technical scheme of the present application is described in detail below through the accompanying drawings and specific embodiments. It should be understood that the embodiments of the present application and the specific features in the embodiments are detailed descriptions of the technical scheme of the present application, rather than limitations on the technical scheme of the present application. In the absence of conflict, the embodiments of the present application and the technical features in the embodiments can be combined with each other.

[0057] Figure 1A dynamic material loading scheduling method based on smart mine service is shown, which is applied to a material loading scheduling decision system. The method includes the following steps 110 to 140.

[0058] Step 110: Acquire material loading scheduling simulation data, wherein the material loading scheduling simulation data includes transportation route planning information of the target smart mine server for the mine material loading scheduling strategy in the first loading scheduling cycle.

[0059] Among them, material loading and scheduling simulation data is data generated based on prediction and simulation, including mine material category data, mine material storage data and mine production business data. For example, material category data may include different types of materials such as coal and metal ore; material storage data covers the current inventory of each material; and mine production business data may include the use of materials in the mine production process.

[0060] The target smart mine server is a centralized computing device or system responsible for processing and managing all mine-related data, including but not limited to material loading scheduling simulation data.

[0061] The first loading scheduling period refers to the time period during which material loading scheduling is performed for the first time, for example, it may be a working day, a shift, or any other predefined time unit.

[0062] The mine material loading scheduling strategy is about how to arrange and optimize the material loading process. It is formulated based on factors such as material category, storage volume and production demand to improve the operating efficiency of the mine.

[0063] The transport route planning information is information about how to arrange the transportation routes for material loading, including but not limited to which roads to choose, when to depart, the order of loading and unloading, etc. For example, if the storage volume of a material is expected to decrease in the future, the transport route planning may give priority to the loading and transportation of this material.

[0064] Step 120: Perform timeliness analysis on the transport route planning information in the material loading scheduling simulation data through a first scheduling decision tree network to obtain timeliness analysis data corresponding to the material loading scheduling simulation data.

[0065] The first scheduling decision tree network is a machine learning model that generates a decision tree through training data for prediction or classification. In this scenario, the first scheduling decision tree network is used to perform timeliness analysis on material loading scheduling simulation data.

[0066] Timeliness analysis is an analytical method used to determine the effectiveness and adaptability of material loading scheduling strategies within a specific time period. It can consider a variety of factors, such as changes in material demand, changes in storage volume, actual conditions of transportation routes, etc., to evaluate whether the existing material loading scheduling strategy can be effectively implemented within a given time period.

[0067] Timeliness analysis data is data generated by timeliness analysis, which contains information about the effectiveness and adaptability of material loading scheduling strategies within a specific time period. For example, if the original transportation route cannot be carried out smoothly due to poor road conditions during a certain period of time, then the timeliness analysis data will reflect this problem.

[0068] Combined with the above definition, step 120 mainly uses the first scheduling decision tree network to perform timeliness analysis on the transport route planning information in the material loading scheduling simulation data to generate corresponding timeliness analysis data. These data will be used to evaluate and optimize the material loading scheduling strategy to ensure its effectiveness and adaptability in actual operation.

[0069] Step 130: Based on the timeliness analysis data corresponding to the material loading scheduling simulation data, determine the first linkage feasibility estimation result of the target smart mine server on the mine material loading scheduling strategy in the first loading scheduling cycle.

[0070] Among them, the first linkage feasibility estimation result is an estimate of the possibility of actual execution of the mine material loading scheduling strategy within the first loading scheduling cycle. This estimation result is determined based on the timeliness analysis data corresponding to the material loading scheduling simulation data. For example, if the timeliness analysis data shows that the original transportation route cannot be carried out smoothly due to poor road conditions during a certain period of time, then the first linkage feasibility estimation result may be negative, indicating that the scheduling strategy is not feasible during this period of time.

[0071] It can be seen that step 130 is mainly based on the timeliness analysis data to evaluate the feasibility of the target smart mine server to implement the mine material loading scheduling strategy in the first loading scheduling cycle, that is, to generate the first linkage feasibility estimation result.

[0072] Step 140, based on the first linkage feasibility estimation result, the scheduling strategy attribute vector set of the mine material loading scheduling strategy, and the transportation service demand vector set of the target smart mine server, determine the second linkage feasibility estimation result of the mine material loading scheduling strategy by the target smart mine server in the second loading scheduling cycle through a second scheduling decision tree network, and the second loading scheduling cycle is later than the first loading scheduling cycle.

[0073] The scheduling strategy attribute vector set is a set of data describing the characteristics of the mine material loading scheduling strategy. Each characteristic is converted into a numerical value or vector form for calculation and analysis. For example, the attributes of the scheduling strategy may include priority, transport line length, expected load, etc.

[0074] The transportation service demand vector set is a set of data describing the transportation service demand of the target smart mine server, which is also converted into numerical or vector form. For example, these demands may include the loading demand, transportation distance, transportation time window, etc. of each material.

[0075] The second loading scheduling cycle refers to the next loading scheduling period after the first loading scheduling cycle, which can be the next working day, shift or other predefined time units.

[0076] The second linkage feasibility estimation result is an estimation of the possibility of actual execution of the mine material loading scheduling strategy in the second loading scheduling cycle. This estimation result is determined based on the first linkage feasibility estimation result, the scheduling strategy attribute vector set and the transportation service demand vector set.

[0077] Step 140 mainly uses the second scheduling decision tree network to predict the possibility of the target smart mine server implementing the mine material loading scheduling strategy in the second loading scheduling cycle based on the first linkage feasibility estimation result, the scheduling strategy attribute vector set and the transportation service demand vector set, that is, to generate the second linkage feasibility estimation result.

[0078] Furthermore, the above two vector sets can be described with specific examples.

[0079] Scheduling strategy attribute vector set: For example, there is a scheduling strategy, in which the attributes may include priority (5), transport line length (10km), expected load (200 tons), etc. These attributes can be converted into numerical values ​​to form a feature vector. For example, in this example, the scheduling strategy attribute vector may be [5, 10, 200].

[0080] Transport service demand vector set: For example, the target smart mine server's demand for transport services includes the loading demand, transportation distance, transportation time window, etc. of each material. For example, at a certain moment, the server's demand may be to transport 100 tons of iron ore, with a transportation distance of 15 kilometers and a transportation time window of 8:00-12:00. These demands can also be converted into numerical values ​​to form a feature vector. In this example, the transport service demand vector may be [100, 15, (8, 12)].

[0081] The above two vector sets can effectively reflect the characteristics of the scheduling strategy and the demand for transportation services, which is very important for predicting the feasibility of material loading scheduling strategies.

[0082] In the above example, the scheduling decision tree network is a machine learning method that uses a decision tree model for prediction. This network is mainly used for classification and regression problems, and its goal is to learn simple decision rules from data features to solve the problem.

[0083] Structure: The decision tree network consists of many nodes and edges. Each internal node represents a feature or attribute, each edge represents a decision rule, and each leaf node represents a decision result.

[0084] Working principle: When a new instance needs to be predicted, the decision tree network starts from the root node and moves down along the edges that satisfy the corresponding decision rules according to the values ​​of each feature of the instance until it reaches a leaf node. The decision result of the leaf node is the prediction result.

[0085] Training process: The training process of decision tree networks usually involves three steps: feature selection, decision tree generation, and decision tree pruning. Feature selection is to determine which features are more important for decision making; decision tree generation is to create a decision tree that can predict as accurately as possible; decision tree pruning is to prevent overfitting, that is, the model learns too complex training data, resulting in inaccurate predictions on new unknown data.

[0086] In the above embodiment, the second scheduling decision tree network predicts the possibility of implementing the mine material loading scheduling strategy by the target smart mine server in the second loading scheduling cycle based on the first linkage feasibility estimation result, the scheduling strategy attribute vector set and the transportation service demand vector set.

[0087] The technical solution described in the above steps 110 to 140 is introduced below through a complete example.

[0088] First, obtain the material loading scheduling simulation data. This data contains the transportation route planning information of the target smart mine server for the mine material loading scheduling strategy in the first loading scheduling cycle. For example, the data may show that it is planned to transport 200 tons of iron ore from mine A to port B in the first loading scheduling cycle, with a transportation distance of 50 kilometers.

[0089] Then, the first scheduling decision tree network is used to analyze the timeliness of these transportation route planning information. For example, the decision tree may analyze the expected carrying volume, transportation distance and other relevant factors to determine whether the transportation can be completed within the scheduled time.

[0090] Next, based on the obtained timeliness analysis data, the first linkage feasibility estimation result of the target smart mine server for the mine material loading scheduling strategy in the first loading scheduling period is determined. For example, if the timeliness analysis indicates that the transportation can be completed within the scheduled time, then the first linkage feasibility estimation result is "feasible".

[0091] Finally, the second scheduling decision tree network is used to determine the second linkage feasibility estimation result of the target smart mine server for the mine material loading scheduling strategy in the second loading scheduling cycle. This step is based on the first linkage feasibility estimation result, the scheduling strategy attribute vector set (for example, [5, 50, 200], representing the priority, transportation line length and expected load, respectively), and the transportation service demand vector set (for example, [200, 50, (8, 12)], representing the material loading demand, transportation distance and transportation time window, respectively). If the decision tree network predicts that the scheduling strategy in the second loading scheduling cycle is also "feasible", then the cycle can be carried out according to the predetermined scheduling strategy.

[0092] The above example demonstrates how to use the scheduling decision tree network and related data to predict the feasibility of the smart mine server's mine material loading scheduling strategy in different loading scheduling cycles.

[0093] By applying the above embodiment, a timeliness evaluation can be performed through a first scheduling decision tree network for a linkage image feature set including the transport route planning information of the mine material loading scheduling strategy of the target smart mine server in the first loading scheduling cycle, so as to mine out the timeliness analysis data of the mine material loading scheduling strategy of the target smart mine server in the first loading scheduling cycle, thereby ensuring the accuracy of the timeliness evaluation, and then determining the first linkage feasibility estimation result of the mine material loading scheduling strategy of the target smart mine server in the first loading scheduling cycle based on the timeliness analysis data. Finally, a second scheduling decision tree network is used to determine the second linkage feasibility estimation result of the mine material loading scheduling strategy of the target smart mine server in the second loading scheduling cycle based on the first linkage feasibility estimation result of the first loading scheduling cycle, the scheduling strategy attribute vector set of the mine material loading scheduling strategy, and the transport service demand vector set of the target smart mine server, thereby improving the accuracy and rationality of the feasibility estimation of the mine material loading scheduling strategy. In this way, by jointly analyzing the material loading and scheduling simulation data including at least mine material category data, mine material storage volume data and mine production business data, the linkage feasibility estimation results corresponding to the dynamic loading and scheduling strategy of the material are determined, which can ensure the timeliness and pertinence of the scheduling during the subsequent execution of the strategy, thereby ensuring the efficiency of mine production.

[0094] In detail, by performing timeliness analysis and linkage feasibility estimation on the material loading scheduling simulation data, the feasibility of the mine material loading scheduling strategy of the smart mine server in different loading scheduling cycles can be more accurately predicted. This will help decision makers formulate more reasonable and accurate scheduling strategies to avoid waste of resources or reduced production efficiency due to wrong predictions. According to the timeliness analysis data and the linkage feasibility estimation results, the scheduling timeliness and pertinence when executing the strategy can be guaranteed, thereby effectively improving the efficiency of mine production. This is important for achieving production goals and reducing production costs. Through the application of the first and second scheduling decision tree networks, the decision-making process is greatly simplified. The system can automatically perform complex calculations and predictions, thereby saving a lot of time and energy for manual decision-making and improving decision-making efficiency. This technical solution can dynamically determine the linkage feasibility estimation results corresponding to the dynamic loading scheduling strategy of the material based on the material loading scheduling simulation data obtained in real time, which is conducive to the dynamic adjustment of the scheduling strategy in the mine production process and improves the flexibility and response speed of the system. Through accurate predictions and reasonable decision-making, various risks caused by incorrect scheduling, such as equipment damage, production accidents, etc., can be effectively reduced, thereby ensuring the smooth progress of production.

[0095] In some examples, the second scheduling decision tree network includes an image description mining branch, an image description intersection branch, and a feasibility estimation branch. Based on this, step 140, based on the first linkage feasibility estimation result, the scheduling strategy attribute vector set of the mine material loading scheduling strategy, and the transport service demand vector set of the target smart mine server, determines the second linkage feasibility estimation result of the target smart mine server for the mine material loading scheduling strategy in the second loading scheduling period through the second scheduling decision tree network, including steps 141-145.

[0096] Step 141: Aggregate the first linkage feasibility estimation result, the scheduling strategy attribute vector set of the mine material loading scheduling strategy, and the transportation service demand vector set of the target smart mine server to obtain a target linkage image feature set.

[0097] Step 142: perform feature selection on the target linkage image feature set to obtain an attention image description vector sequence.

[0098] Step 143: Call the image description mining branch to perform reversible processing on the attention image description vector sequence to obtain a reversible description vector sequence.

[0099] Step 144: Call the image description cross branch to cross the attention image description vector sequence with the reversible description vector sequence to obtain a cross description vector sequence.

[0100] Step 145, calling the feasibility estimation branch to perform feasibility mapping on the cross description vector sequence, and outputting the second linkage feasibility estimation result of the target smart mine server on the mine material loading scheduling strategy in the second loading scheduling cycle.

[0101] In the above embodiment, the target linkage image feature set is a data set aggregated by the first linkage feasibility estimation result, the scheduling strategy attribute vector set, and the transport service demand vector set. For example, the first linkage feasibility estimation result is "feasible", the scheduling strategy attribute vector set is [5, 50, 200] (representing priority, transport line length and expected load), and the transport service demand vector set is [200, 50, (8, 12)] (representing material loading demand, transportation distance and transportation time window). Aggregating these three elements together, the target linkage image feature set is obtained.

[0102] The attention image description vector sequence is a data sequence obtained by feature selection of the target linkage image feature set. The attention mechanism can help the model focus on the most important features, thereby improving the accuracy of the prediction.

[0103] The reversible description vector sequence is a data sequence obtained by reversibly processing the attention image description vector sequence through the image description mining branch. Reversible processing means that the processing process can be reversed, that is, the reversible description vector sequence can be restored to the original attention image description vector sequence.

[0104] The cross description vector sequence is a data sequence obtained by crossing the attention image description vector sequence with the reversible description vector sequence through the image description cross branch. The cross process can help the model understand and parse the data from two different perspectives, thereby improving the robustness of the prediction.

[0105] The second linkage feasibility estimation result is the prediction result obtained by feasibility mapping the cross description vector sequence through the feasibility estimation branch. If the mapping result shows that it is feasible to execute the current material loading scheduling strategy in the second loading scheduling cycle, then the second linkage feasibility estimation result is "feasible".

[0106] For example, in the first loading scheduling cycle, the target smart mine server has planned a material loading scheduling strategy, and the goal is to predict whether this strategy is still feasible in the second loading scheduling cycle.

[0107] First, we need to obtain the feasibility estimation result of the first linkage, the scheduling strategy attribute vector set, and the transportation service demand vector set. Assume that the feasibility estimation result of the first linkage is "feasible", the scheduling strategy attribute vector set is [5, 50, 200] (representing priority, transportation line length, and expected load), and the transportation service demand vector set is [200, 50, (8, 12)] (representing material loading demand, transportation distance, and transportation time window). Aggregating these three elements together, we get the target linkage image feature set.

[0108] Next, feature selection is performed on the target linkage image feature set to obtain an attention image description vector sequence. This process may involve some deep learning algorithms, such as the attention mechanism, which is used to select the most important features from the original feature set.

[0109] Then, the image description mining branch is called to perform reversible processing on the attention image description vector sequence to obtain a reversible description vector sequence. This process may involve some deep learning algorithms, such as autoencoders, to reduce the dimension of the data while retaining the main information.

[0110] Next, the image description cross branch is called to cross the attention image description vector sequence with the reversible description vector sequence to obtain a cross description vector sequence. This process may involve some deep learning algorithms, such as hybrid models, to understand and parse data from different perspectives.

[0111] Finally, the feasibility estimation branch is called to perform feasibility mapping on the cross description vector sequence, and the second linkage feasibility estimation result of the target smart mine server for the mine material loading scheduling strategy in the second loading scheduling cycle is output. If the mapping result shows that it is feasible to execute the current material loading scheduling strategy in the second loading scheduling cycle, then the second linkage feasibility estimation result is "feasible".

[0112] The above is a complete process example, showing how to use the second scheduling decision tree network and related data to predict the feasibility of the mine material loading scheduling strategy of the smart mine server in the second loading scheduling cycle. In summary, this technical solution further improves the estimation accuracy and rationality of the feasibility of the mine material loading scheduling strategy by introducing a series of methods such as target linkage image feature set, attention mechanism, reversible processing, cross description and feasibility mapping, optimizes production efficiency, improves decision-making efficiency, and reduces production risks.

[0113] In some possible embodiments, the debugging process of the second scheduling decision tree network includes steps 210 to 250.

[0114] Step 210, obtaining a past learning sample set, wherein the past learning sample set includes at least a past transportation service demand vector set of a past smart mine server in a first past loading scheduling cycle, a past scheduling strategy attribute vector set of a past mine material loading scheduling strategy in the first past loading scheduling cycle, and a third linkage feasibility estimation result of the past smart mine server for the past mine material loading scheduling strategy in the first past loading scheduling cycle and a fourth linkage feasibility estimation result in the second past loading scheduling cycle.

[0115] Among them, the third linkage feasibility estimation result is determined based on the first material loading scheduling simulation data of the past smart mine server for the past mine material loading scheduling strategy in the first past loading scheduling cycle, and the fourth linkage feasibility estimation result is determined based on the second material loading scheduling simulation data of the past smart mine server for the past mine material loading scheduling strategy in the second past loading scheduling cycle, and the second past loading scheduling cycle is located after the first past loading scheduling cycle.

[0116] Step 220: Generate a target linkage image feature set sample based on the past transport service demand vector set, the past scheduling strategy attribute vector set and the third linkage feasibility estimation result.

[0117] Step 230: Input the target linkage image feature set sample into the second basic neural network for feasibility estimation processing to obtain the second linkage feasibility estimation result of the past smart mine server on the past mine material loading scheduling strategy in the second loading scheduling period.

[0118] Step 240: Determine a second target training cost according to the second linkage feasibility estimation result and the fourth linkage feasibility estimation result.

[0119] Step 250: improve the network parameters of the second basic neural network based on the second target training cost, and perform cyclic debugging until the debugging requirements are met to obtain a second scheduling decision tree network.

[0120] In the above embodiment, the past learning sample set is a set of past data, which is used to train and debug the second scheduling decision tree network. It at least includes the transportation service demand vector set of the past smart mine server in different loading scheduling cycles, the scheduling strategy attribute vector set of the mine material loading scheduling strategy, and the linkage feasibility estimation result of the smart mine server for the mine material loading scheduling strategy.

[0121] The terms third linkage feasibility estimation result and fourth linkage feasibility estimation result refer to the linkage feasibility estimation results of the smart mine server for the mine material loading scheduling strategy in the first and second past loading scheduling cycles. These results are determined based on the material loading scheduling simulation data of the smart mine server for the mine material loading scheduling strategy in the corresponding cycle.

[0122] The target linkage image feature set samples are sample data generated based on the past transport service demand vector set, the past scheduling strategy attribute vector set and the third linkage feasibility estimation results. These sample data will be used to train and debug the second basic neural network.

[0123] The second basic neural network is a deep learning model for performing feasibility estimation processing. It receives the target linkage image feature set sample as input and outputs the second linkage feasibility estimation result of the smart mine server for the mine material loading scheduling strategy in the second loading scheduling cycle.

[0124] The second objective training cost is an indicator to measure the prediction effect of the second basic neural network. It is calculated based on the second linkage feasibility estimation results and the fourth linkage feasibility estimation results. This cost value will be used to optimize the network parameters of the second basic neural network.

[0125] For example, a past learning sample set has been obtained, which includes the transportation service demand vector set of the smart mine server in each loading scheduling cycle in the past period of time, the scheduling strategy attribute vector set of the mine material loading scheduling strategy, and the linkage feasibility estimation result of the smart mine server for the mine material loading scheduling strategy. Then, based on the past transportation service demand vector set, the past scheduling strategy attribute vector set and the first linkage feasibility estimation result, a target linkage image feature set sample is generated. Next, the target linkage image feature set sample is input into the second basic neural network for feasibility estimation processing, and the second linkage feasibility estimation result of the mine material loading scheduling strategy of the smart mine server in the second loading scheduling cycle is obtained. Then, the second target training cost is determined based on the second linkage feasibility estimation result and the second first linkage feasibility estimation result. Finally, the network parameters of the second basic neural network are improved based on the second target training cost, and cyclic debugging is performed until the debugging compliance requirements are met to obtain the second scheduling decision tree network.

[0126] The key part of the above technical solution is to determine the second target training cost, that is, the training loss function. In deep learning, the loss function is used to measure the gap between the model prediction result and the actual result. For this technical solution, the following loss function can be defined: Assuming that the model's linkage feasibility estimation result for the mine material loading scheduling strategy in the second loading scheduling cycle is y_hat, and the actual second and first linkage feasibility estimation result is y, then the loss function L can be defined as the mean square error (Mean Squared Error) between them, that is: L = (y-y_hat)^2;

[0127] In step 240, the second target training cost, that is, the loss function L, is calculated based on the second linkage feasibility estimation result y_hat and the second first linkage feasibility estimation result y.

[0128] Then in step 250, the parameters of the neural network need to be optimized based on the loss function L. Common optimization methods include gradient descent and stochastic gradient descent. The specific optimization process is to iteratively update the parameter value through the gradient of the network parameters to minimize the loss function.

[0129] For example, if the neural network parameter is W, then in each iteration, the value of W is updated to minimize the loss function L:

[0130] Where P is the learning rate, is the gradient of the loss function L with respect to W.

[0131] In this way, the network parameters of the second basic neural network can be continuously improved and debugged in a loop until the debugging requirements are met (for example, the loss function L drops to a certain threshold or reaches a preset number of training rounds), thereby obtaining the second scheduling decision tree network. In summary, this technical solution can improve the prediction accuracy of the feasibility of mine material loading scheduling strategies by using past data and deep learning models for training and debugging, which helps to optimize production efficiency and decision-making efficiency and reduce production risks. Specifically, by introducing a series of methods such as past learning sample sets, target linkage image feature set samples, second basic neural networks, and second target training costs, effective debugging of the second scheduling decision tree network is achieved. This helps to improve prediction accuracy, reduce prediction errors, optimize production efficiency, improve decision-making efficiency, and reduce production risks.

[0132] In some other possible embodiments, the method further includes steps 310 to 330.

[0133] Step 310: Obtain second material loading scheduling simulation data of the past smart mine server for the past mine material loading scheduling strategy in the second past loading scheduling period.

[0134] Step 320: Perform timeliness analysis on the second material loading scheduling simulation data through the first scheduling decision tree network, and output timeliness analysis data corresponding to the second material loading scheduling simulation data.

[0135] Step 330: Based on the timeliness analysis data corresponding to the second material loading scheduling simulation data, determine the fourth linkage feasibility estimation result of the past smart mine server on the past mine material loading scheduling strategy in the second past loading scheduling period.

[0136] In the above embodiment, the second material loading scheduling simulation data refers to the simulation data of the mine material loading scheduling strategy of the smart mine server in the second past loading scheduling cycle. These data can help understand the effect of the material loading scheduling strategy under specific conditions. The timeliness analysis data is the data obtained by performing timeliness analysis on the second material loading scheduling simulation data through the first scheduling decision tree network. These data reflect the trend and characteristics of the simulation data over time. The fourth linkage feasibility estimation result is the linkage feasibility estimation result of the mine material loading scheduling strategy of the smart mine server in the second past loading scheduling cycle determined based on the timeliness analysis data corresponding to the second material loading scheduling simulation data.

[0137] For example, the second material loading scheduling simulation data of the intelligent mine server for the mine material loading scheduling strategy in the second past loading scheduling period has been obtained. These data may include information such as the material loading amount, loading time, and loading location.

[0138] Then, the first scheduling decision tree network is used to perform time-efficiency analysis on these simulated data to obtain time-efficiency analysis data. This process may involve some deep learning algorithms, such as recurrent neural networks (RNN) or long short-term memory networks (LSTM), which are used to analyze the changing trends and characteristics of data over time.

[0139] Then, based on the timeliness analysis data, the fourth linkage feasibility estimation result of the intelligent mine server for the mine material loading scheduling strategy in the second past loading scheduling cycle is determined. This process may involve some machine learning algorithms, such as logistic regression or support vector machine (SVM), which are used to predict whether the material loading scheduling strategy is feasible under given conditions.

[0140] In this way, by introducing a series of methods such as the second material loading scheduling simulation data, timeliness analysis data and the fourth linkage feasibility estimation results, the timeliness analysis and prediction of the mine material loading scheduling strategy are realized. This helps to improve the prediction accuracy, optimize production efficiency, improve decision-making efficiency, and reduce production risks.

[0141] In some preferred embodiments, the method further includes steps 410 to 440 before performing timeliness analysis on the transport route planning information in the material loading scheduling simulation data through the first scheduling decision tree network and obtaining timeliness analysis data corresponding to the material loading scheduling simulation data.

[0142] Step 410: Obtain past material loading and scheduling simulation data and corresponding past timeliness analysis data, wherein the past material loading and scheduling simulation data at least includes past transportation route planning information of past smart mine servers for past mine material loading and scheduling strategies.

[0143] Step 420: input the past material loading and scheduling simulation data into the pre-debugged first basic neural network to perform timeliness analysis based on the past transportation route planning information in the past material loading and scheduling simulation data to obtain a timeliness prediction result.

[0144] Step 430: Determine a first target training cost based on the past timeliness analysis data and the timeliness prediction result.

[0145] Step 440: Improve the network parameters of the pre-debugging first basic neural network based on the first target training cost, and perform cyclic debugging until the debugging requirements are met to obtain a first scheduling decision tree network.

[0146] In the above embodiment, the past material loading and scheduling simulation data is the simulation data collected by the smart mine server for the past mine material loading and scheduling strategy, which at least includes the past transportation route planning information. Past timeliness analysis data These data come from the timeliness analysis of past material loading and scheduling simulation data, reflecting the changing trends and characteristics of transportation route planning information in the past period of time. The timeliness prediction result is the prediction result obtained after the timeliness analysis of the past material loading and scheduling simulation data by the pre-debugging first basic neural network. The first target training cost is a measurement indicator determined based on the past timeliness analysis data and the timeliness prediction results, which is used to evaluate the prediction accuracy of the pre-debugging first basic neural network. The pre-debugging first basic neural network is a neural network model that has not been fully trained, and is used for timeliness analysis and prediction.

[0147] This technical solution is explained below through a specific example.

[0148] First, the past material loading scheduling simulation data collected by the smart mine server for past mine material loading scheduling strategies were obtained. These data at least include transportation route planning information. At the same time, the past timeliness analysis data corresponding to these simulation data were also obtained.

[0149] Then, these past material loading and scheduling simulation data were entered into the pre-debugged first basic neural network, and timeliness analysis was performed based on past transportation route planning information to obtain timeliness prediction results.

[0150] Next, the first target training cost is determined based on the past timeliness analysis data and timeliness prediction results. Assuming that the real past timeliness analysis data is y and the timeliness prediction result is y_hat, the mean square error can be used as the first target training cost, that is, L = (y-y_hat)^2.

[0151] Finally, the network parameters of the pre-debugged first basic neural network are improved according to the first target training cost, and cyclic debugging is performed until the debugging requirements are met, such as the loss function drops to a certain threshold or reaches a preset number of training rounds, thereby obtaining the first scheduling decision tree network.

[0152] In this way, by introducing a series of methods such as past material loading scheduling simulation data, past timeliness analysis data, timeliness prediction results, first target training cost and pre-debugging first basic neural network, effective training and debugging of the first scheduling decision tree network is achieved. This helps to improve prediction accuracy, optimize production efficiency, improve decision-making efficiency, and reduce production risks.

[0153] In some examples, the timeliness analysis data includes a confidence weight that the material loading scheduling simulation data belongs to a target timeliness level annotation. Based on this, step 130 determines the first linkage feasibility estimation result of the target smart mine server for the mine material loading scheduling strategy in the first loading scheduling cycle based on the timeliness analysis data corresponding to the material loading scheduling simulation data, including steps 131-132.

[0154] Step 131: Determine the timeliness level value corresponding to the material loading scheduling simulation data based on the confidence weight of the material loading scheduling simulation data belonging to the target timeliness level annotation.

[0155] Step 132: Based on the timeliness level value corresponding to the material loading scheduling simulation data, determine the first linkage feasibility estimation result of the target smart mine server on the mine material loading scheduling strategy in the first loading scheduling cycle.

[0156] In some embodiments, the material loading scheduling simulation data is a data set generated by simulation operation, which describes the loading scheduling of mine materials under specific conditions. These data may include various parameters, such as work efficiency, equipment utilization rate, material flow speed, etc. The confidence weight of the target timeliness level annotation is an indicator to measure the accuracy and reliability of the simulation data. It represents the confidence level of the simulation data corresponding to the expected results in actual operation. The first linkage feasibility estimation result is a prediction or estimate of the possibility of executing the mine material loading scheduling strategy within a specific period. This result is determined based on the simulation data and the timeliness level value.

[0157] For example, a smart mine system is considering changing its material loading scheduling strategy to improve efficiency. In order to evaluate the possible impact of the new strategy, the system generates a set of material loading scheduling simulation data. The system then assigns a confidence weight to these simulation data, indicating the reliability of these data in actual operations.

[0158] In step 131, the system determines a timeliness rating value based on the confidence weight of the simulation data. For example, if the confidence weight is high, the timeliness rating value will be increased accordingly. In step 132, the system uses this timeliness rating value to determine the estimated linkage feasibility of implementing the new strategy within the first loading scheduling cycle. If the timeliness rating value is high, the predicted linkage feasibility will also be high.

[0159] In this way, the material loading scheduling strategy can be accurately simulated and evaluated, thereby helping the smart mine server to formulate more effective and efficient material loading scheduling strategies. At the same time, by introducing confidence weights and timeliness level values, the accuracy and reliability of simulation data can be improved, further improving the accuracy of decision-making and optimization effects.

[0160] In some preferred embodiments, step 132 describes determining the first linkage feasibility estimation result of the target smart mine server on the mine material loading scheduling strategy in the first loading scheduling cycle based on the timeliness level value corresponding to the material loading scheduling simulation data, including steps 1321-1322.

[0161] Step 1321: Obtain quantization weighting factors and linkage involvement factors.

[0162] Step 1322: The sum of the timeliness grade value corresponding to the material loading scheduling simulation data and the quantitative weighting factor is used as the target weighted result.

[0163] Step 1323, combining the target weighted result with the linkage involvement factor, to obtain a first linkage feasibility estimation result of the target smart mine server for the mine material loading scheduling strategy in the first loading scheduling cycle.

[0164] In this embodiment, the quantitative weighting factor is a numerical value used to measure the impact of each variable on the result. It can be derived based on actual experience or statistical data, and can provide a standardized method to compare and understand the importance of different variables. The linkage involvement factor is a parameter that describes the degree of interaction between different parts of the system. For example, in a material loading scheduling strategy, the linkage involvement factor may describe the dependency between various loading equipment.

[0165] In step 1321, firstly, a quantitative weighting factor and a linkage involvement factor are obtained. The quantitative weighting factor may come from the statistics of historical data, while the linkage involvement factor may be determined by the physical model of the system.

[0166] In step 1322, the timeliness level value is added to the quantitative weighting factor to obtain a target weighted result. For example, if the timeliness level value is 5 and the quantitative weighting factor is 0.8, the target weighted result is 5.8.

[0167] In step 1323, the target weighted result is combined with the linkage involvement factor to obtain a final first linkage feasibility estimation result. For example, if the linkage involvement factor is 0.6, then the final estimation result is 5.8*0.6=3.48.

[0168] In this way, by introducing quantitative weighting factors and linkage involvement factors, the feasibility assessment of material loading scheduling strategies is more accurate and can better reflect the impact of different factors on the possibility of strategy implementation. At the same time, this method also provides a structured and quantitative way to evaluate strategy changes in complex systems, which helps decision makers better understand and control the behavior of the system.

[0169] In some alternative embodiments, after the second linkage feasibility estimation result of the mine material loading scheduling strategy by the target smart mine server in the second loading scheduling cycle is determined by the second scheduling decision tree network based on the first linkage feasibility estimation result, the scheduling strategy attribute vector set of the mine material loading scheduling strategy, and the transportation service demand vector set of the target smart mine server, the method also includes steps 510-530.

[0170] Step 510: determine a smart mine server pool, where the smart mine server pool includes a plurality of the target smart mine servers.

[0171] Step 520: Generate a strategy matching tag for the mine material loading scheduling strategy based on the first linkage feasibility estimation result and the second linkage feasibility estimation result corresponding to each of the target smart mine servers.

[0172] Step 530: Send the strategy matching tag to the strategy generation system of the mine material loading scheduling strategy.

[0173] In the above embodiment, the strategy matching tag is an identifier indicating whether a material loading scheduling strategy matches a specific smart mine server. It is generated based on the first and second linkage feasibility estimation results of each target smart mine server.

[0174] In step 510, the material loading scheduling decision system first determines a smart mine server pool, which includes multiple target smart mine servers, which may have different hardware configurations, workloads, and scheduling requirements.

[0175] In step 520, the material loading scheduling decision system generates a strategy matching label for the mine material loading scheduling strategy based on the first and second linkage feasibility estimation results of each target smart mine server. For example, if the linkage feasibility estimation result of a server is high, it may get a "match" label; otherwise, it may get a "mismatch" label.

[0176] In step 530, the material loading scheduling decision system sends these policy matching tags to the policy generation system. After receiving these tags, the policy generation system can know which policies match which servers, so as to make better scheduling decisions.

[0177] In this way, by introducing the smart mine server pool, the second scheduling decision tree network and the strategy matching label, the material loading scheduling decision is made more accurate and efficient. It can generate the corresponding strategy matching label according to the actual situation and prediction results of each server, thus providing a flexible and adaptive method to optimize the material loading scheduling strategy, and ultimately improve the operating efficiency of the entire smart mine system.

[0178] In some examples, the material loading scheduling simulation data also includes associated transportation route planning information of the associated smart mine server for the mine material loading scheduling strategy in the first loading scheduling cycle, and the associated smart mine server belongs to the same server queue as the target smart mine server. Then, obtaining the material loading scheduling simulation data in step 110 includes steps 111-113.

[0179] Step 111: Obtain vehicle transport route planning information of the target smart mine server for the mine material loading scheduling strategy in the first loading scheduling cycle.

[0180] Step 112: determine the server queue to which the target smart mine server belongs, and obtain the initial transport route planning information of each associated smart mine server in the server queue for the mine material loading scheduling strategy in the first loading scheduling cycle.

[0181] Step 113: based on the initial transport route planning information of each associated smart mine server in the server queue for the mine material loading scheduling strategy in the first loading scheduling period, screen the associated transport route planning information of the associated smart mine server for the mine material loading scheduling strategy in the first loading scheduling period.

[0182] In the above embodiment, the associated transport route planning information is data about the transportation route plan of the smart mine server for the material loading scheduling strategy within a specific loading scheduling period. This may include information such as route selection, equipment allocation, and time scheduling. A server queue is a group of smart mine servers arranged in a specific order. In a server queue, each server may affect or be affected by other servers.

[0183] In step 111, the system first obtains the vehicle transportation route planning information of the target smart mine server for the material loading scheduling strategy in the first loading scheduling period. For example, this may include which vehicles to use for transportation in the period, what routes to choose, how to arrange the time, etc.

[0184] In step 112, the system determines the server queue to which the target smart mine server belongs, and obtains the initial transport route planning information of each associated smart mine server in the queue. This information may come from the historical data or preset parameters of each server.

[0185] In step 113, the system selects the associated transport route planning information of the target smart mine server in the first loading scheduling cycle based on the initial transport route planning information of each associated smart mine server. This information may include equipment shared with the target server, intersecting routes, and time schedules that affect each other.

[0186] In this way, by acquiring and processing the associated transport route planning information, the simulation and prediction of material loading scheduling strategies are more accurate. It takes into account the impact of each server in the server queue, thus better reflecting the complex situation in actual operation. This approach can help improve the operating efficiency of the entire system, reduce conflicts and delays, and thus improve the productivity of the entire smart mine.

[0187] In some alternative embodiments, step 113 screens the associated transportation route planning information of the associated smart mine server for the mine material loading scheduling strategy in the first loading scheduling cycle based on the initial transportation route planning information of each associated smart mine server in the server queue for the mine material loading scheduling strategy in the first loading scheduling cycle, including steps 1131-1132.

[0188] Step 1131, determining the involvement weight between each associated smart mine server in the server queue and the target smart mine server.

[0189] Step 1132, based on the initial transportation route planning information of each associated smart mine server for the mine material loading scheduling strategy, select the initial transportation route planning information of the associated smart mine server with the highest involvement weight as the associated transportation route planning information of the associated smart mine server for the mine material loading scheduling strategy in the first loading scheduling cycle.

[0190] In the above embodiment, the involvement weight is an indicator of the degree of mutual influence between the target smart mine server and the associated smart mine server. It may be calculated based on various factors, such as the physical distance between the two, the number of shared resources, historical collaboration efficiency, etc.

[0191] In step 1131, the system first determines the involvement weight between each associated smart mine server in the server queue and the target smart mine server. For example, if an associated server is close to the target server, has many shared devices, and has a high historical collaboration efficiency, its involvement weight may be high.

[0192] In step 1132, the system selects the initial transport route planning information of the associated smart mine server with the highest weight according to the initial transport route planning information of each associated smart mine server. This means that the system will give priority to the information of those associated servers that have the greatest mutual influence with the target server.

[0193] In this way, by introducing the weight of involvement and screening the associated transport route planning information, the simulation and prediction are more accurate and effective. It gives priority to the associated servers that have the greatest mutual influence with the target server, thus better reflecting the complex situation in actual operation. This method can improve the operating efficiency of the system, reduce conflicts and delays, and further improve the productivity of the entire smart mine.

[0194] Combined with the above, the mine material loading scheduling strategy mainly involves how to effectively schedule various materials in the mine (such as ore, rock, etc.) between different equipment and locations. The following is a specific example: For example, a mine has two mining areas A and B, which produce iron ore and copper ore respectively. The mine also has two loaders L1 and L2, and three transport vehicles T1, T2 and T3.

[0195] At the beginning of the day, the system first generates a set of material loading scheduling simulation data based on historical data, equipment status, expected demand, etc. For example, it may predict the output of area A and area B in the next period of time, as well as the working efficiency of each loader and each transport vehicle.

[0196] The system then develops a preliminary material loading scheduling strategy based on these simulation data. For example, it may decide to have L1 work in the morning to load iron ore for T1 and T2 first, while L2 works in the afternoon to load copper ore for T3.

[0197] During the execution process, the system will continuously collect actual operation data and compare it with the simulation data. If there are large deviations, such as the efficiency of a loader is lower than expected, or the output of a mining area is higher than expected, the system will update the material loading scheduling strategy based on the new data.

[0198] For example, if it is found that L1 is less efficient, the system may decide to adjust the strategy and let L2 start working earlier; if it is found that the copper ore output in area B has increased significantly, the system may decide to let T1 and T2 also transport some of the copper ore.

[0199] In this way, the mine material loading scheduling strategy can be dynamically adjusted to adapt to the actual operating conditions of the mine, thereby maximizing resource utilization and improving production efficiency.

[0200] In some independent embodiments, the mine material loading scheduling strategy is obtained by jointly analyzing the mine material category, the mine material storage volume and the mine production business data.

[0201] Determining the dynamic loading scheduling strategy for materials requires considering multiple factors such as the type of mine materials, storage volume, and production operations, so as to ensure the timeliness and pertinence of scheduling. The following is a specific process:

[0202] Analyze material categories: First, you need to understand what types of materials are in the mine, such as iron ore, copper ore, gold ore, etc. Each material may require different processing equipment and methods, so you need to plan the corresponding loading and transportation tasks according to the material category.

[0203] Evaluate material storage volume: Next, you need to evaluate the storage volume of each material. If the storage volume of a certain material is large, you may need to prioritize more loading and transportation tasks to handle this material; conversely, if the storage volume of a certain material is small, you may need to reduce the loading and transportation tasks for this material to avoid wasting resources.

[0204] Investigate production operations: In addition, it is also necessary to investigate the production operations of the mine. For example, if the demand for a certain material suddenly increases, then it may be necessary to adjust the strategy and increase the loading and transportation tasks for this material; conversely, if the demand for a certain material suddenly decreases, then it may be necessary to reduce the loading and transportation tasks for this material.

[0205] Determine the dynamic loading scheduling strategy: Based on the above analysis, the dynamic loading scheduling strategy of the material can be determined. This strategy should be able to flexibly respond to various situations to ensure the efficiency of mine production.

[0206] For example, if the iron ore stockpile is large and the demand is high, more loaders and haulers may be deployed to handle the iron ore; conversely, if the gold ore stockpile is small and the demand is low, the loading and hauling of gold ore may be reduced.

[0207] In general, determining the dynamic loading scheduling strategy for materials is a process that takes multiple factors into consideration and requires continuous adjustment and optimization.

[0208] On the basis of the above, a material loading scheduling decision system is provided, comprising a processor and a memory communicating with each other, wherein the processor is used to retrieve a computer program from the memory and implement the above method by running the computer program.

[0209] Based on the above, a computer-readable storage medium is provided, on which a computer program is stored, and the computer program implements the above method when running.

[0210] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A dynamic material loading scheduling method based on smart mine service, characterized in that: Applied to a material loading scheduling decision system, the method comprises: Acquire material loading scheduling simulation data, wherein the material loading scheduling simulation data includes transportation route planning information of the target smart mine server for the mine material loading scheduling strategy in the first loading scheduling cycle; Performing timeliness analysis on the transport route planning information in the material loading scheduling simulation data through a first scheduling decision tree network to obtain timeliness analysis data corresponding to the material loading scheduling simulation data; Based on the timeliness analysis data corresponding to the material loading scheduling simulation data, determining a first linkage feasibility estimation result of the target smart mine server on the mine material loading scheduling strategy in the first loading scheduling cycle; Determine, by a second scheduling decision tree network, a second linkage feasibility estimation result of the target smart mine server for the mine material loading scheduling strategy in a second loading scheduling period based on the first linkage feasibility estimation result, the scheduling strategy attribute vector set of the mine material loading scheduling strategy, and the transport service demand vector set of the target smart mine server, wherein the second loading scheduling period is later than the first loading scheduling period; The second scheduling decision tree network includes an image description mining branch, an image description cross branch, and a feasibility estimation branch; the second scheduling decision tree network determines the second linkage feasibility estimation result of the target smart mine server for the mine material loading scheduling strategy in the second loading scheduling period based on the first linkage feasibility estimation result, the scheduling strategy attribute vector set of the mine material loading scheduling strategy, and the transportation service demand vector set of the target smart mine server, including: Aggregating the first linkage feasibility estimation result, the scheduling strategy attribute vector set of the mine material loading scheduling strategy, and the transportation service demand vector set of the target smart mine server to obtain a target linkage image feature set; Performing feature selection on the target linkage image feature set to obtain an attention image description vector sequence; Calling the image description mining branch to perform reversible processing on the attention image description vector sequence to obtain a reversible description vector sequence; Calling the image description cross branch to cross the attention image description vector sequence with the reversible description vector sequence to obtain a cross description vector sequence; Calling the feasibility estimation branch to perform feasibility mapping on the cross description vector sequence, and outputting a second linkage feasibility estimation result of the target smart mine server on the mine material loading scheduling strategy in the second loading scheduling period; The debugging process of the second scheduling decision tree network includes: Obtain a past learning sample set, the past learning sample set at least including a past transportation service demand vector set of a past smart mine server in a first past loading scheduling cycle, a past scheduling strategy attribute vector set of a past mine material loading scheduling strategy in the first past loading scheduling cycle, and a third linkage feasibility estimation result of the past smart mine server for the past mine material loading scheduling strategy in the first past loading scheduling cycle and a fourth linkage feasibility estimation result in the second past loading scheduling cycle; The third linkage feasibility estimation result is determined based on the first material loading scheduling simulation data of the past smart mine server for the past mine material loading scheduling strategy in the first past loading scheduling cycle, and the fourth linkage feasibility estimation result is determined based on the second material loading scheduling simulation data of the past smart mine server for the past mine material loading scheduling strategy in the second past loading scheduling cycle, and the second past loading scheduling cycle is located after the first past loading scheduling cycle; Generate a target linkage image feature set sample based on the past transport service demand vector set, the past scheduling strategy attribute vector set and the third linkage feasibility estimation result; Inputting the target linkage image feature set sample into the second basic neural network for feasibility estimation processing, and obtaining the second linkage feasibility estimation result of the past smart mine server on the past mine material loading scheduling strategy in the second loading scheduling period; Determining a second target training cost according to the second linkage feasibility estimation result and the fourth linkage feasibility estimation result; Improving the network parameters of the second basic neural network based on the second target training cost, and performing cyclic debugging until the debugging standard requirement is met, thereby obtaining a second scheduling decision tree network; The method further comprises: Acquire second material loading scheduling simulation data of the past intelligent mine server for the past mine material loading scheduling strategy in the second past loading scheduling period; Performing timeliness analysis on the second material loading scheduling simulation data through the first scheduling decision tree network, and outputting timeliness analysis data corresponding to the second material loading scheduling simulation data; Based on the timeliness analysis data corresponding to the second material loading scheduling simulation data, determining a fourth linkage feasibility estimation result of the past smart mine server on the past mine material loading scheduling strategy in the second past loading scheduling period; Before performing timeliness analysis on the transport route planning information in the material loading scheduling simulation data through the first scheduling decision tree network and obtaining timeliness analysis data corresponding to the material loading scheduling simulation data, the method further includes: Acquire past material loading and scheduling simulation data and corresponding past timeliness analysis data, wherein the past material loading and scheduling simulation data at least includes past transportation route planning information of past smart mine servers for past mine material loading and scheduling strategies; Entering the past material loading scheduling simulation data into the pre-debugged first basic neural network, so as to perform timeliness analysis based on the past transportation route planning information in the past material loading scheduling simulation data to obtain a timeliness prediction result; Determining a first target training cost according to the past timeliness analysis data and the timeliness prediction result; Improving the network parameters of the pre-debugging first basic neural network based on the first target training cost, and performing cyclic debugging until the debugging standard requirement is met, thereby obtaining a first scheduling decision tree network; The timeliness analysis data includes the confidence weight of the material loading scheduling simulation data belonging to the target timeliness level annotation; the timeliness analysis data corresponding to the material loading scheduling simulation data is determined based on the target smart mine server. The feasibility estimation result of the first linkage of the mine material loading scheduling strategy in the first loading scheduling period includes: Determining the timeliness level value corresponding to the material loading scheduling simulation data based on the confidence weight that the material loading scheduling simulation data belongs to the target timeliness level annotation; Based on the timeliness level value corresponding to the material loading scheduling simulation data, a first linkage feasibility estimation result of the target smart mine server on the mine material loading scheduling strategy in the first loading scheduling cycle is determined.

2. The method according to claim 1, characterized in that The determining of the first linkage feasibility estimation result of the target smart mine server on the mine material loading scheduling strategy in the first loading scheduling period based on the timeliness level value corresponding to the material loading scheduling simulation data includes: Obtaining quantitative weighting factors and linkage involvement factors; The sum of the timeliness level value corresponding to the material loading scheduling simulation data and the quantitative weighting factor is used as the target weighting result; By combining the target weighted result with the linkage involvement factor, a first linkage feasibility estimation result of the target smart mine server on the mine material loading scheduling strategy in the first loading scheduling cycle is obtained.

3. The method according to claim 1, characterized in that: After determining the second linkage feasibility estimation result of the mine material loading scheduling strategy by the target smart mine server in the second loading scheduling period based on the first linkage feasibility estimation result, the scheduling strategy attribute vector set of the mine material loading scheduling strategy, and the transportation service demand vector set of the target smart mine server through the second scheduling decision tree network, the method further includes: Determine a smart mine server pool, wherein the smart mine server pool includes a plurality of target smart mine servers; Generate a strategy matching tag for the mine material loading scheduling strategy according to the first linkage feasibility estimation result and the second linkage feasibility estimation result corresponding to each of the target smart mine servers; The strategy matching tag is sent to a strategy generation system of the mine material loading scheduling strategy.

4. The method according to claim 1, characterized in that: The material loading scheduling simulation data also includes associated transport route planning information of the associated smart mine server for the mine material loading scheduling strategy in the first loading scheduling cycle, and the associated smart mine server belongs to the same server queue as the target smart mine server. The material loading scheduling simulation data is obtained, including: Acquire vehicle transport route planning information of the target smart mine server for the mine material loading scheduling strategy in the first loading scheduling period; Determine the server queue to which the target smart mine server belongs, and obtain the initial transport route planning information of each associated smart mine server in the server queue for the mine material loading scheduling strategy in the first loading scheduling period; According to the initial transport route planning information of each associated smart mine server in the server queue for the mine material loading scheduling strategy in the first loading scheduling period, the associated transport route planning information of the associated smart mine server for the mine material loading scheduling strategy in the first loading scheduling period is screened; Among them, the initial transport route planning information of each associated smart mine server in the server queue for the mine material loading scheduling strategy in the first loading scheduling period, screening the associated transport route planning information of the associated smart mine server for the mine material loading scheduling strategy in the first loading scheduling period, includes: Determine the involvement weight between each associated smart mine server in the server queue and the target smart mine server; Based on the initial transportation route planning information of each associated smart mine server for the mine material loading scheduling strategy, the initial transportation route planning information of the associated smart mine server with the highest weight is selected as the associated transportation route planning information of the associated smart mine server for the mine material loading scheduling strategy in the first loading scheduling cycle.

5. A material loading scheduling decision system, characterized in that: The material loading scheduling decision system includes a processor and a memory that communicate with each other, and the processor is used to call a computer program from the memory and implement the method described in any one of claims 1 to 4 by running the computer program.

Citation Information

Cited By

  • Redundant path design system and method for improving stability of coal conveying system

    CN119847025A

  • Redundant path design system and method for improving the stability of coal transmission system

    CN119847025B