Ore body or block economic value evaluation method based on quantum environment perception

Abstract

The invention relates to an ore body or block economic value evaluation method based on quantum environment perception. The method comprises the following steps: S1, selecting an ore body or block economic evaluation data set and importing data; s2, dividing the price data into a training set and a test set according to a ratio of 7: 3; s3, carrying out encoding operation on the data; s4, constructing a graph structure; s5, quantizing the graph structure; s6, setting hyper-parameters of an economic value assessment model (QEM) based on quantum environment perception; s7, training the economic value evaluation model QEM by using the training set until the model loss function is converged; s8, evaluating the economic value of the ore body or the block section by using an economic value evaluation model QEM; and S9, performing test evaluation and verification on the economic value evaluation model QEM through the test set. According to the method, the mining value of the ore body or the blocks is estimated in real time, and the result is displayed in an explicit mode through Web visualization; compared with other existing systems, the method utilizes the compatibility of quantum state modeling to multi-dimensional feature data to promote the estimation of the economic value of the ore body or the block.

Description

Technical Field

[0001] The present invention belongs to the field of mine big data and artificial intelligence technology, and specifically relates to a method for evaluating the economic value of ore bodies or blocks based on quantum environmental perception. Background Art

[0002] An ore body is a collection of ore of a certain size, grade, and mineability within a mineral deposit, typically characterized by relatively well-defined morphology, occurrence, and spatial distribution. Blocks are defined by dividing an ore body into several relatively independent, easily mineable units based on multiple factors, including its geological characteristics, mining conditions, resource distribution, and economic viability. Each block typically exhibits relatively uniform ore grade, thickness, and lithologic distribution, making it easy to mine and facilitating the development of targeted mining plans and technical solutions.

[0003] The division of ore bodies and blocks has a significant impact on the economic and technical feasibility of mining. Proper ore body division can effectively assess the ore body's resource reserves, avoid over-exploitation and resource waste, and accurately reflect the ore body's mining potential, providing a scientific basis for subsequent development planning and environmental protection. Proper block division can optimize mining methods and processes, improve recovery rates, reduce costs, and minimize resource losses during the mining process. For example, through proper block division, over-exploitation of depleted ore bodies can be avoided, ensuring the sustainable use of resources. Furthermore, block division can help mines design the most appropriate mining methods, ensure maximum ore recovery during mining, and enhance the mine's overall economic value. Summary of the Invention

[0004] The purpose of the present invention is to overcome the shortcomings of the existing technology and provide a method for economic value assessment of ore bodies or blocks based on quantum environmental perception. It deeply studies and analyzes factors such as ore grade, metal content, average uranium content, etc. of the ore body or block, and uses the multi-dimensional construction capability of quantum superposition state to extract the potential correlation of different characteristics, improve the accuracy of economic value assessment, further reduce costs and optimize the mining supply chain system.

[0005] The present invention solves the technical problem by the following technical solutions:

[0006] A method for evaluating the economic value of an ore body or block based on quantum environmental perception, the method comprising the following steps:

[0007] S1. Select the ore body or block economic evaluation data set and import the data;

[0008] S2: Divide the price data imported in S1 into a training set and a test set in a ratio of 7:3. The training set is used for model training, and the test set is used to evaluate the final performance of the system.

[0009] S3, encode the data in S2;

[0010] S4, build graph structure;

[0011] S5, graph structure quantization;

[0012] S6. Setting hyperparameters of the economic value assessment model QEM based on quantum environment perception;

[0013] S7. Use the training set to train the economic value evaluation model QEM until the model loss function converges;

[0014] S8. Use the economic value assessment model QEM trained in S7 to assess the economic value of the ore body or block;

[0015] S9. Conducting experimental evaluation and verification on the economic value assessment model QEM through a test set.

[0016] Moreover, the S4 is specifically:

[0017] (1) Node feature design: Each ore body or block is a node in the graph. Each node contains characteristic data of ore grade, average uranium content per square meter, average uranium concentration and metal content. The node features are numerical or discrete data, depending on the different characteristics of the ore body.

[0018] (2) Edge design and weight assignment: Edge weights are assigned based on geographical location, ore body similarity, and geological characteristics, and are calculated based on the relative position, similarity, and mining difficulty factors between ore bodies or blocks.

[0019] Moreover, the S5 is specifically:

[0020] (1) Quantum state encoding of nodes: embedding node features into quantum states through Givens rotation;

[0021] (2) Quantum representation of the adjacency matrix: realized through the tensor product of quantum bits.

[0022] Furthermore, the S7 is specifically:

[0023] (1) Using the quantum graph attention network to comprehensively learn and perceive the ore body modeled by the graph structure and the mining environment in which it is located, and integrating it with economic evaluation indicators to predict the economic value of each ore body or block;

[0024] (2) Evaluate the model on the test set and adjust the hyperparameters based on the results to ensure the model's generalization ability;

[0025] (3) Traverse different hyperparameter combinations through random search and select the optimal hyperparameter configuration.

[0026] Furthermore, the S8 is specifically:

[0027] (1) Use the trained QEM model to evaluate the test set data;

[0028] (2) Calculate the error between the evaluation category and the true category to evaluate the model capability;

[0029] (3) Evaluate model performance through multi-dimensional indicators and compare it with existing technologies.

[0030] The advantages and beneficial effects of the present invention are:

[0031] The present invention can perform real-time estimation of the mining value of ore bodies or blocks, and explicitly display the results through Web visualization; compared with other existing systems, the present invention uses the compatibility of quantum state modeling with multi-dimensional feature data to promote the economic value estimation of ore bodies or blocks; the present invention introduces quantum computing into the mine's self-developed system and embeds it into the system in a black box manner, with a simple overall interface that is operator-friendly. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 This is a diagram of the economic value assessment system for ore bodies or blocks based on quantum environmental perception of the present invention;

[0033] Figure 2 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0034] The present invention will be further described in detail below through specific examples. The following examples are only illustrative and not restrictive, and the scope of protection of the present invention cannot be limited thereto.

[0035] A method for economic value assessment of ore bodies or blocks based on quantum environmental perception, the innovation of which lies in the following steps:

[0036] Step 1: Select the ore body-block economic evaluation data and display the data. The specific steps are as follows:

[0037] Select the button "Ore body-block economic evaluation" on the page and click "Show data" to ensure that the data selection is correct.

[0038] Step 2: Based on step 1, the ore body-block economic evaluation data set is divided into a training set T and a test set S in a ratio of 7:3.

[0039] Step 3: Encode the structured data based on steps 1 and 2. The specific steps are as follows:

[0040] For structured data sets, one-hot encoding is used to encode the data. Specifically, each entity Q in the data set is encoded. his represented by a f-dimensional one-hot encoded binary vector, Q h The hth element of is set to 1, and the other elements are set to 0, and the encoding process ends.

[0041] Step 4 and Step 5: Based on Step 1, Step 2, and Step 3, construct a quantum graph dataset. The specific steps are as follows:

[0042] First, we construct graph data of orebody-block information using traditional methods. Then, we quantize the data. The quantum state of the node feature matrix X is represented as follows:

[0043]

[0044] Where N is the number of nodes and C is the dimensional feature. Here we have ore bodies or blocks, metal content, average square meter uranium content, average uranium concentration, average grade, evaluation results, etc.

[0045] Adjacency matrix A ij The quantum state of is represented as follows:

[0046]

[0047] where |i>,|j> correspond to v i , v j quantum state.

[0048] Step 6, Step 7: Based on Step 1, Step 2, Step 3, Step 4, and Step 5, establish and train the quantum walk-driven model. The specific steps are as follows:

[0049] Start setting the model's initial learning rate, batch size, ent_vec_di, and maximum number of iterations.

[0050]

[0051] Step 7. Based on steps 1, 2, 3, 4, 5, and 6, use the training set to train QEM until the model loss function converges. The specific steps are as follows:

[0052] The parameters are quantized and the similarity between adjacent nodes is calculated through operations such as controlled-Hadamard gate and controlled rotation gate to simulate the attention score in the classic algorithm. The specific formula is as follows:

[0053] α ij =<ψ i |U * WU|ψ i >

[0054] Where U and W are parameterized quantum gates.

[0055] After measuring the state of the quantum gate, the parameters are updated using the classical deep learning gradient descent method. Finally, the softmax function is used to classify the economic value of each node into three categories: economic, sub-marginal economic, and uneconomic.

[0056] Here is the probability vector predicted by the model and is the label vector, which is set to 1 for true multi-tuples and 0 for false multi-tuples, and L is the loss function.

[0057] Step 8. Based on steps 1, 2, 3, 4, 5, 6, and 7, use the trained QEM to evaluate the test set S. The specific steps are as follows:

[0058] For each set of data in a given test set S, use the trained QEM to evaluate, obtain the traversal test set, and complete the test set inference.

[0059] Step 9: Based on Step 1, Step 2, Step 3, Step 4, Step 5, Step 6, and Step 7, the effect of the QEM model proposed in the present invention is evaluated and verified through experiments. The specific steps are as follows:

[0060] The accuracy of the experiment and the overall precision (mean of the precision) are calculated to evaluate the model effect. In order to better balance these three indicators, comparative experiments are used to evaluate and verify the performance of the technology, as shown in Table 1.

[0061] Table 1

[0062] Accuracy Precision(Average) LSTM 0.75632 0.75331 GRU 0.83421 0.80526 GNN 0.83433 0.81245 GCN 0.85221 0.79631 GRU+GCN 0.85452 0.83421 OUR 0.90371 0.88932

[0063] The QEM model achieved an accuracy of 0.90371 and a precision of 0.88932, respectively, which are 0.04919 and 0.05511 higher than the previously used top-performing method, GRU+GCN. It also achieved improvements of 0.05150 and 0.09301, respectively, compared to the traditional GCN method. This demonstrates QEM's ability to efficiently process the associations of multiple nodes and multidimensional features in mining data, significantly improving the efficiency of complex pattern recognition and enabling better evaluation of the mining value of ore bodies or blocks.

[0064] Although the embodiments and drawings of the present invention are disclosed for illustrative purposes, those skilled in the art will understand that various substitutions, changes and modifications are possible without departing from the spirit and scope of the present invention and the appended claims. Therefore, the scope of the present invention is not limited to the contents disclosed in the embodiments and drawings.

Claims

1. A method for evaluating the economic value of an ore body or block based on quantum environmental perception, characterized by: The steps of the method are: S1. Select the ore body or block economic evaluation data set and import the data; S2: Divide the price data imported in S1 into a training set and a test set in a ratio of 7:

3. The training set is used for model training, and the test set is used to evaluate the final performance of the system. S3, encode the data in S2; S4, build graph structure; S5, graph structure quantization; S6. Setting hyperparameters of the economic value assessment model QEM based on quantum environment perception; S7. Use the training set to train the economic value evaluation model QEM until the model loss function converges; S8. Use the economic value assessment model QEM trained in S7 to assess the economic value of the ore body or block; S9. Conducting experimental evaluation and verification on the economic value assessment model QEM through a test set.

2. The method for economic value assessment of an ore body or block based on quantum environmental perception according to claim 1, characterized in that: The S4 is specifically: (1) Node feature design: Each ore body or block is a node in the graph. Each node contains characteristic data of ore grade, average uranium content per square meter, average uranium concentration and metal content. The node features are numerical or discrete data, depending on the different characteristics of the ore body. (2) Edge design and weight assignment: Edge weights are assigned based on geographical location, ore body similarity, and geological characteristics, and are calculated based on the relative position, similarity, and mining difficulty factors between ore bodies or blocks.

3. The method for economic value assessment of an ore body or block based on quantum environmental perception according to claim 1, characterized in that: The S5 is specifically: (1) Quantum state encoding of nodes: embedding node features into quantum states through Givens rotation; Quantum representation of adjacency matrices: implemented via tensor products of qubits.

4. The method for economic value assessment of an ore body or block based on quantum environmental perception according to claim 1, characterized in that: The S7 is specifically: (1) Using the quantum graph attention network to comprehensively learn and perceive the ore body modeled by the graph structure and the mining environment in which it is located, and integrating it with economic evaluation indicators to predict the economic value of each ore body or block; (2) Evaluate the model on the test set and adjust the hyperparameters based on the results to ensure the model's generalization ability; (3) Traverse different hyperparameter combinations through random search and select the optimal hyperparameter configuration.

5. The method for economic value assessment of an ore body or block based on quantum environmental perception according to claim 1, characterized in that: The S8 is specifically: (1) Use the trained QEM model to evaluate the test set data; (2) Calculate the error between the evaluation category and the true category to evaluate the model capability; (3) Evaluate model performance through multi-dimensional indicators and compare it with existing technologies.

Images