Material reserve priority evaluation method and device

Abstract

The invention provides a material reserve priority evaluation method and device. The method comprises the following steps: acquiring a reserve material information set; processing the reserve material information set to obtain a material characteristic relation network; processing the material characteristic relation network to obtain a characteristic attribute matrix; and processing the characteristic attribute matrix to obtain a priority evaluation result. Therefore, the method provided by the invention can comprehensively consider the substitutability, the profit period, the introduction rate, the storability and the economical efficiency, and carries out quantitative evaluation on the priority of the material reserve.

Description

Technical Field

[0001] The present invention belongs to the technical field of material reserve priority, and specifically relates to a material reserve priority evaluation method and device. Background Art

[0002] The material reserve system is the core of the national reserve system, an important part of ensuring national security, and a solid guarantee for improving the country's ability to respond to emergencies. Ensuring adequate material supply is the key to effectively responding to various emergencies. However, due to objective factors such as the shelf life of reserve materials, limited procurement funds, and uncertainty in the occurrence of emergencies, it is not appropriate to reserve a large amount of materials. In practice, excessive reserves leading to material expiration or insufficient reserves leading to material shortages have become huge challenges faced by governments around the world. There are a large number of evaluation and decision-making behaviors in material reserve activities. How to carry out material reserve activities under the influence of objective factors and the priority evaluation of material reserves are particularly critical.

[0003] The priority assessment task of material reserves has the characteristics of involving many types of materials, strong professionalism, and large amount of calculation. From the current research on the priority assessment method of material reserves, there are still some problems: First, the assessment method is not objective enough. It is mainly based on the actual experience of the evaluator and is mostly assessed manually. Not only is the calculation and processing cycle long, but it also has a considerable degree of subjective arbitrariness. The level and quality of the assessment mainly depend on the personal quality and ability of the manager; second, the priority assessment of material reserves considers fewer factors, lacks the correlation analysis between materials and characteristics, and between materials, which affects the reliability of the assessment results. In view of the above shortcomings of the current priority assessment of material reserves, a more scientific and reasonable priority assessment method for material reserves is urgently needed. From the dimensions of material security demand, fundraising and supply, storage management, etc., a priority assessment process for material reserves is formulated, and the priority of material reserves is evaluated using an algorithm model to provide scientific and reasonable auxiliary decision-making for material reserve managers. Summary of the invention

[0004] The technical problem to be solved by the present invention is to provide a material reserve priority assessment method and device, which is conducive to quantitatively assessing the priority of material reserves by integrating multiple characteristics.

[0005] In order to solve the above technical problems, the first aspect of the embodiment of the present invention discloses a material reserve priority evaluation method, the method comprising:

[0006] S1, obtaining a set of reserve material information; the set of reserve material information includes N pieces of reserve material information, where N is a positive integer;

[0007] The reserve material information includes replacement information, fundraising period, number of foreign purchases, total number of purchases, storage time, material unit price, and set information set;

[0008] The set information set includes set information; the set information includes serial numbers of the reserve materials belonging to the same category;

[0009] S2, processing the reserve material information set to obtain a material property relationship network; the material property relationship network includes N material nodes, a first attribute node, a second attribute node, a third attribute node, a fourth attribute node, a fifth attribute node and a plurality of attribute edges;

[0010] S3, processing the material property relationship network to obtain a property attribute matrix;

[0011] S4, processing the characteristic attribute matrix to obtain a priority evaluation result.

[0012] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the processing of the reserve material information set to obtain a material property relationship network includes:

[0013] S21, quantizing the replacement information to obtain a first associated information set; the first associated information set includes N first associated information;

[0014] S22, processing the fundraising period to obtain a second associated information set; the second associated information set includes N second associated information;

[0015] S23, processing the foreign purchase quantity and the total purchase quantity to obtain a third associated information set; the third associated information set includes N third associated information;

[0016] S24, processing the storage duration to obtain a fourth association information set; the fourth association information set includes N fourth association information;

[0017] S25, processing the material unit price to obtain a fifth associated information set; the fifth associated information set includes N fifth associated information;

[0018] S26, processing the first association information set, the second association information set, the third association information set, the fourth association information set, the fifth association information set and the set information set to obtain a material property relationship network.

[0019] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the processing of the fundraising period to obtain a second set of associated information includes:

[0020] S221, obtaining the maximum value of the N fundraising periods to obtain the longest fundraising period;

[0021] S222, obtaining the minimum value of the N fundraising periods to obtain the shortest fundraising period;

[0022] S223, using a second association information calculation model, processing the longest fundraising period, the shortest fundraising period and each fundraising period to obtain corresponding second association information;

[0023]

[0024] In the formula, R i,2 is the second associated information corresponding to the i-th fundraising period, t 2,max is the longest fundraising period, t 2,min is the shortest fundraising period, t 2,i The i-th fundraising period, i being an integer from 1 to N;

[0025] S224: Combine N pieces of the second associated information to obtain a set of the second associated information.

[0026] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the processing of the storage duration to obtain a fourth association information set includes:

[0027] S241, obtaining the maximum value of the N storage durations to obtain the longest storage duration;

[0028] S242, obtaining the minimum value of the N storage durations to obtain the shortest storage duration;

[0029] S243, using the fourth association information calculation model, processing the longest storage time, the shortest storage time and each storage time to obtain the corresponding fourth association information;

[0030]

[0031] In the formula, R j,4 is the fourth association information corresponding to the jth storage duration, t 4,max is the maximum storage time, t 4,min is the shortest storage time, t 4,j is the j-th storage duration, where j is an integer from 1 to N;

[0032] S244: Combine the N fourth association information to obtain the fourth association information set.

[0033] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the material unit price is processed to obtain a fifth associated information set, including:

[0034] S251, obtaining the maximum value of the N material unit prices to obtain the highest material unit price;

[0035] S252, using the fifth association information calculation model, processing the highest material unit price and each material unit price to obtain the corresponding fifth association information;

[0036]

[0037] In the formula, R k,5 is the fifth associated information corresponding to the kth material unit price, p max is the highest unit price of the materials, p k is the kth unit price of the material, where k is an integer from 1 to N;

[0038] S253: Combine the N pieces of the fifth association information to obtain the fifth association information set.

[0039] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the first association information set, the second association information set, the third association information set, the fourth association information set, the fifth association information set and the set information set are processed to obtain a material property relationship network, including:

[0040] S261, initializing the first relationship network to include only the first attribute node, the second attribute node, the third attribute node, the fourth attribute node, the fifth attribute node and N material nodes;

[0041] S262, performing a first connection process on the first relationship network based on the first association information set, the second association information set, the third association information set, the fourth association information set, and the fifth association information set to obtain a second relationship network;

[0042] S263: Based on the set information set, perform a second connection process on the second relationship network to obtain the material property relationship network.

[0043] As an optional implementation, in the first aspect of the embodiment of the present invention, the processing of the material property relationship network to obtain a property attribute matrix includes:

[0044] S31, extracting features from the material property relationship network to obtain a property extraction matrix; the number of rows of the property extraction matrix is ​​N+5;

[0045] S32, splitting the feature extraction matrix to obtain a first feature vector sequence and a second feature vector sequence; the first feature vector sequence includes N feature vectors; the second feature vector sequence includes 5 feature vectors;

[0046] S33, using a characteristic attribute calculation model, processing the first characteristic vector sequence and the second characteristic vector sequence to obtain the characteristic attribute matrix;

[0047] The characteristic attribute calculation model is:

[0048]

[0049] Where X is the characteristic attribute matrix, x pq is the element in the pth row and qth column of the characteristic attribute matrix, where p is an integer from 1 to N, and q is an integer from 1 to 5; u p is the pth first eigenvector in the first eigenvector sequence, v q is the kth second eigenvector in the second eigenvector sequence.

[0050] As an optional implementation, in the first aspect of the embodiment of the present invention, the feature extraction of the material property relationship network to obtain a feature extraction matrix includes:

[0051] S311, the number of training times and the optimal value are initialized to 0 and the preset initial optimal value, respectively, and the feature matrix H is randomly initialized (0) ; Initialize the initial feature extraction matrix to H (0) ;

[0052] The feature matrix and the initial feature extraction matrix are both matrices of N+5 rows and K columns, where K is a preset feature length;

[0053] S312, obtaining the adjacency matrix A and degree matrix D of the material property relationship network;

[0054] S313, determining the number of training times;

[0055] When the number of training times is greater than the preset total number of training times, executing S319; the total number of training times is greater than 1;

[0056] When the number of training times is less than or equal to the total number of training times, executing S314;

[0057] S314, using a preset feature extraction model, extracting the adjacency matrix A, the degree matrix D and the feature matrix H (0) Processing is performed to obtain the output feature matrix;

[0058] S315, using the target value calculation model, processing the material property relationship network and the output feature matrix to obtain a target value L;

[0059] The target value calculation model is:

[0060]

[0061] Where M is the material property relationship network, m i is the i-th node of M, i is an integer from 1 to N+5; Ne(m i ) represents node m i The set of neighbor nodes in M, Pe(m i ) represents m i The set of non-neighbor nodes in M, m j and m k are the j-th and k-th nodes of M respectively, and m j and m k They belong to Ne(m i ) and Pe(m i );σ * is the sigmoid function, <·> represents the inner product of the calculated vector; δ is the preset negative sample adoption rate; v i 、v j and v k are the i-th, j-th and k-th row vectors in the output feature matrix respectively;

[0062] S316, adding 1 to the value of the number of training times;

[0063] S317, judging the target value;

[0064] When the target value is greater than or equal to the preset optimal value, updating the parameters of the feature extraction model to random values; executing S313;

[0065] When the target value is less than the optimal value, the optimal value is updated to the target value; the initial feature extraction matrix is ​​updated to the output feature matrix; and S318 is executed;

[0066] S318, determining the optimal value;

[0067] If the optimal value is greater than or equal to the preset optimal threshold, the parameters of the feature extraction model are updated to random values; and S313 is executed;

[0068] If the optimal value is less than the optimal threshold, determining the feature extraction matrix as the initial feature extraction matrix, and executing S32;

[0069] S319: Determine the feature extraction matrix as the initial feature extraction matrix.

[0070] As an optional implementation manner, in the first aspect of the embodiment of the present invention, the processing of the characteristic attribute matrix to obtain a priority evaluation result includes:

[0071] S41, using a competition matrix calculation model, processing the characteristic attribute matrix to obtain a competition matrix;

[0072] The competition matrix calculation model is:

[0073]

[0074] In the formula, E uv is the element in the uth row and vth column of the competition matrix C, e is the intermediate calculation result, x uk is the element in the uth row and kth column of the characteristic attribute matrix, x vk is the element in the vth row and kth column of the characteristic attribute matrix, where u, v and k are all integers from 1 to N;

[0075] S42, processing the competition matrix to obtain an evaluation vector;

[0076] S43, processing the evaluation vector to obtain the priority evaluation result.

[0077] The second aspect of the embodiment of the present invention discloses a material reserve priority evaluation device, the device comprising a reserve material information acquisition module, a material characteristic relationship network generation module, a characteristic attribute matrix calculation module and a priority evaluation module;

[0078] The reserve material information acquisition module is used to acquire a reserve material information set.

[0079] The material property relationship network generation module is used to process the reserve material information set to obtain a material property relationship network.

[0080] The characteristic attribute matrix calculation module is used to process the material characteristic relationship network to obtain a characteristic attribute matrix.

[0081] The priority evaluation module is used to process the characteristic attribute matrix to obtain a priority evaluation result.

[0082] A third aspect of an embodiment of the present invention discloses another material reserve priority evaluation device, the device comprising:

[0083] A memory storing executable program code;

[0084] a processor coupled to the memory;

[0085] The processor calls the executable program code stored in the memory to execute part or all of the steps in the material reserve priority evaluation method disclosed in the first aspect of the embodiment of the present invention.

[0086] The fourth aspect of the present invention discloses a computer storage medium, which stores computer instructions. When the computer instructions are called, they are used to execute some or all of the steps in the material reserve priority evaluation method disclosed in the first aspect of an embodiment of the present invention.

[0087] Compared with the prior art, the embodiments of the present invention have the following beneficial effects:

[0088] This method processes the acquired reserve material information set to obtain the material characteristic relationship network, characteristic attribute matrix and priority evaluation results in turn. It can comprehensively consider the substitutability, fundraising period, introduction rate, storability and economy, and conduct quantitative evaluation on the priority of material reserves. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0090] Figure 1 This is a flow chart of a material reserve priority assessment method proposed in the present invention.

[0091] Figure 2 This is a structural schematic diagram of a material reserve priority assessment device proposed by the present invention.

[0092] Figure 3 This is a schematic diagram of the structure of another material reserve priority assessment device proposed by the present invention. DETAILED DESCRIPTION

[0093] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0094] The terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish different objects, rather than to describe a specific order. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, device, product or equipment that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units that are not listed, or may optionally include other steps or units that are inherent to these processes, methods, products or equipment.

[0095] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present invention. The appearance of the phrase in various places in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0096] The present invention discloses a material reserve priority assessment method and device, which are described in detail below.

[0097] Embodiment 1

[0098] See also Figure 1 , Figure 1 : is a flow chart of a material reserve priority evaluation method disclosed in an embodiment of the present invention. Figure 1 The described material reserve priority assessment method is applied to the field of material reserve priority assessment, such as reserve priority assessment of large-scale disaster relief materials, and the embodiments of the present invention are not limited thereto.

[0099] like Figure 1 As shown, the material reserve priority assessment method may include the following operations:

[0100] S1, obtaining a set of reserve material information; the set of reserve material information includes N pieces of reserve material information, where N is a positive integer;

[0101] The reserve material information includes replacement information, fundraising period, foreign purchase quantity, total purchase quantity, storage time, material unit price, and set information set.

[0102] It should be noted that the above N represents the type of reserve materials, and each reserve material corresponds to a reserve material information in the reserve material information set; the value of the above replacement information is existence or non-existence, which is used to indicate whether there is a replacement material for the reserve material corresponding to the reserve material information; the above fundraising period is used to indicate the purchase time of the corresponding reserve material; the above total purchase amount and foreign purchase amount respectively indicate the total purchase quantity of the corresponding reserve material and the quantity purchased from abroad; the above storage time indicates the total storage time of the corresponding reserve material in the warehouse;

[0103] It should be noted that the above-mentioned set information set includes set information; the above-mentioned set information includes the serial numbers of the reserve materials belonging to the same category. Take five kinds of reserve materials (compressed dry food g1, emergency rescue whistle g2, safety rope g3, cold-proof clothes g4 and rain boots g5) as an example, where compressed dry food and emergency rescue whistle are both earthquake first aid kit items, the set information set includes set information {1,2}; in addition, when there is no association relationship between the reserve materials, the set information set is empty;

[0104] S2, processing the reserve material information set to obtain a material property relationship network; the material property relationship network includes N material nodes, a first attribute node, a second attribute node, a third attribute node, a fourth attribute node, a fifth attribute node and a plurality of attribute edges.

[0105] It should be noted that the above-mentioned N material nodes are used to represent N types of reserve materials respectively; the above-mentioned attribute edges are authorized edges, which are used to connect any two nodes of the material characteristic relationship network; the above-mentioned nodes are material nodes, first attribute nodes, second attribute nodes, third attribute nodes, fourth attribute nodes or fifth attribute nodes.

[0106] S3, processing the material property relationship network to obtain a property attribute matrix.

[0107] S4, processing the characteristic attribute matrix to obtain a priority evaluation result.

[0108] In an optional embodiment, the above-mentioned processing of the reserve material information set to obtain a material property relationship network includes:

[0109] S21, quantizing the replacement information to obtain a first associated information set; the first associated information set includes N first associated information;

[0110] Optionally, the quantization process is to judge the N replacement information respectively to obtain the corresponding first associated information; when the value of the replacement information is present, the corresponding first associated information is set to 1; when the value of the replacement information is not present, the corresponding first associated information is set to 0;

[0111] S22, processing the fundraising period to obtain a second associated information set; the second associated information set includes N second associated information;

[0112] S23, processing the foreign purchase quantity and the total purchase quantity to obtain a third associated information set; the third associated information set includes N third associated information;

[0113] It should be noted that the above processing of the foreign purchase number and the total purchase amount is to calculate the ratio of each foreign purchase number to the corresponding total purchase amount to obtain the corresponding third related information; and the N third related information are combined to obtain a third related information set;

[0114] S24, processing the storage duration to obtain a fourth association information set; the fourth association information set includes N fourth association information;

[0115] S25, processing the material unit price to obtain a fifth associated information set; the fifth associated information set includes N fifth associated information;

[0116] S26, using the first association information set, the second association information set, the third association information set, the fourth association information set, the fifth association information set and the set information set to perform processing to obtain a material property relationship network.

[0117] In another optional embodiment, the above processing of the fundraising period to obtain a second set of associated information includes:

[0118] S221, obtaining the maximum value of the N fundraising periods to obtain the longest fundraising period;

[0119] S222, obtaining the minimum value of the N fundraising periods to obtain the shortest fundraising period;

[0120] S223, using a second association information calculation model, processing the longest fundraising period, the shortest fundraising period and each fundraising period to obtain corresponding second association information;

[0121]

[0122] In the formula, R i,2 is the second associated information corresponding to the i-th fundraising period, t 2,max is the longest fundraising period, t 2,min is the shortest fundraising period, t 2,i The i-th fundraising period, i being an integer from 1 to N;

[0123] S224: Combine N pieces of the second associated information to obtain a set of the second associated information.

[0124] In yet another optional embodiment, the storage duration is processed to obtain a fourth association information set, including:

[0125] S241, obtaining the maximum value of the N storage durations to obtain the longest storage duration;

[0126] S242, obtaining the minimum value of the N storage durations to obtain the shortest storage duration;

[0127] S243, using the fourth association information calculation model, processing the longest storage time, the shortest storage time and each storage time to obtain the corresponding fourth association information;

[0128]

[0129] In the formula, R j,4 is the fourth association information corresponding to the jth storage duration, t 4,max is the maximum storage time, t 4,min is the shortest storage time, t 4,j is the j-th storage duration, where j is an integer from 1 to N;

[0130] S244: Combine the N fourth association information to obtain the fourth association information set.

[0131] In yet another optional embodiment, the material unit price is processed as described above to obtain a fifth associated information set, including:

[0132] S251, obtaining the maximum value of the N material unit prices to obtain the highest material unit price;

[0133] S252, using the fifth association information calculation model, processing the highest material unit price and each material unit price to obtain the corresponding fifth association information;

[0134]

[0135] In the formula, R k,5 is the fifth associated information corresponding to the kth material unit price, p max is the highest unit price of the materials, p k is the kth unit price of the material, where k is an integer from 1 to N;

[0136] S253: Combine the N pieces of the fifth association information to obtain the fifth association information set.

[0137] In yet another optional embodiment, the material property relationship network is obtained by processing the first association information set, the second association information set, the third association information set, the fourth association information set, the fifth association information set, and the set information set, including:

[0138] S261, initializing the first relationship network to include only the first attribute node, the second attribute node, the third attribute node, the fourth attribute node, the fifth attribute node and N material nodes;

[0139] S262, performing a first connection process on the first relationship network based on the first association information set, the second association information set, the third association information set, the fourth association information set, and the fifth association information set to obtain a second relationship network;

[0140] S263: Based on the set information set, perform a second connection process on the second relationship network to obtain the material property relationship network.

[0141] In yet another optional embodiment, the first connection processing is performed on the first relationship network based on the first association information set, the second association information set, the third association information set, the fourth association information set, and the fifth association information set to obtain the second relationship network, including:

[0142] S1621, initializing the first cycle number ii to 1;

[0143] S1622, respectively obtain the i-th first association information, the second association information, the third association information, the fourth association information, and the fifth association information in the first association information set, the second association information set, the third association information set, the fourth association information set, and the fifth association information set, to obtain current first association information, current second association information, current third association information, current fourth association information, and current fifth association information;

[0144] S1623, judging the current first association information;

[0145] When the first association information is 1, in the first relationship network, the attribute edge with a weight of 1 connecting the ii-th material node and the first characteristic node is established; when the first association information is 0, the first relationship network remains unchanged;

[0146] S1624, judging the current third association information;

[0147] When the third association information is 1, in the first relationship network, the attribute edge with a weight of 1 connecting the ii-th material node and the third characteristic node is established; when the third association information is 0, the first relationship network remains unchanged;

[0148] S1625, in the first relationship network, establishing the attribute edges connecting the ii-th material node to the second characteristic node, the fourth characteristic node, and the fifth characteristic node, respectively, and setting the weight values ​​of the corresponding attribute edges to the current second association information, the current fourth association information, and the current fifth association information, respectively;

[0149] S1626, adding 1 to the value of the first cycle number ii;

[0150] S1627, repeat S1622 to S1626 until ii equals N+1; determine that the second relationship network is the first relationship network.

[0151] In yet another optional embodiment, the second connection processing is performed on the second relationship network based on the set information set to obtain the material property relationship network, including:

[0152] S1631, initialize the second loop number jj to 1;

[0153] S1632, obtaining the jjth set of set information in the set information set to obtain the current set of set information;

[0154] S1633, obtaining any number in the current set information to obtain the current number;

[0155] S1634, deleting the current number from the current set information;

[0156] S1635, in the second relationship network, establishing the material node corresponding to the current number, and the attribute edge with a weight of 1 connected to the material node corresponding to each number in the current set information;

[0157] S1636, repeat S1633 to S1635 until the current set information contains only one number;

[0158] S1637, adding 1 to the value of the second cycle number jj;

[0159] S1638, repeat S1632 to S1637 until jj is equal to the number of set information in the set information set plus 1; determine that the material characteristic relationship network is the second relationship network.

[0160] In yet another optional embodiment, the material property relationship network is processed to obtain a property attribute matrix, including:

[0161] S31, extracting features from the material property relationship network to obtain a property extraction matrix; the number of rows of the property extraction matrix is ​​N+5;

[0162] It should be noted that the number of rows of the above-mentioned feature extraction matrix is ​​equal to the number of nodes of the material feature relationship network, and each row vector is used to describe the embedded features of a node. For example, the row vectors of the 1st to Nth rows of the feature extraction matrix correspond to the embedded features of the material nodes numbered 1 to N, respectively, and the row vectors of the N+1st to N+5th rows correspond to the embedded features of the first attribute node, the second attribute node, the third attribute node, the fourth attribute node, and the fifth attribute node, respectively, which is not limited in the embodiment of the present invention.

[0163] S32, splitting the feature extraction matrix to obtain a first feature vector sequence and a second feature vector sequence; the first feature vector sequence includes N feature vectors; the second feature vector sequence includes 5 feature vectors;

[0164] S33, using a characteristic attribute calculation model, processing the first characteristic vector sequence and the second characteristic vector sequence to obtain the characteristic attribute matrix;

[0165] The characteristic attribute calculation model is:

[0166]

[0167] Where X is the characteristic attribute matrix, x pq is the element in the pth row and qth column of the characteristic attribute matrix, where p is an integer from 1 to N, and q is an integer from 1 to 5; u p is the pth first eigenvector in the first eigenvector sequence, v q is the kth second eigenvector in the second eigenvector sequence.

[0168] In yet another optional embodiment, the above-mentioned splitting process of the feature extraction matrix to obtain the first feature vector sequence and the second feature vector sequence includes:

[0169] Initialize the first eigenvector sequence and the second eigenvector sequence to be empty;

[0170] Sequentially obtain the row vectors corresponding to the material nodes numbered 1 to N, and insert them into the end of the first feature vector sequence in the order of acquisition to obtain the first feature vector sequence;

[0171] The row vectors corresponding to the first attribute node, the second attribute node, the third attribute node, the fourth attribute node and the fifth attribute node are obtained in sequence, and are sequentially inserted into the end of the second feature vector sequence to obtain the second feature vector sequence.

[0172] In yet another optional embodiment, the feature extraction of the material property relationship network is performed to obtain a feature extraction matrix, including:

[0173] S311, the number of training times and the optimal value are initialized to 0 and the preset initial optimal value, respectively, and the feature matrix H is randomly initialized (0) ; Initialize the initial feature extraction matrix to H (0) ;

[0174] The feature matrix and the initial feature extraction matrix are both matrices with N+5 rows and K columns, where K is a preset feature length.

[0175] It should be noted that the elements in the feature matrix are initialized to random values.

[0176] Preferably, the above initial optimal value is 10.

[0177] S312, obtaining the adjacency matrix A and degree matrix D of the material property relationship network.

[0178] S313, determining the number of training times;

[0179] When the number of training times is greater than the preset total number of training times, executing S319; the total number of training times is greater than 1;

[0180] When the number of training times is less than or equal to the total number of training times, S314 is executed.

[0181] Preferably, the total number of training times is 1000.

[0182] S314, using a preset feature extraction model, extracting the adjacency matrix A, the degree matrix D and the feature matrix H (0) After processing, the output feature matrix is ​​obtained.

[0183] It should be noted that the above feature extraction model is a GCN (graph convolutional neural network), which includes an input layer and m hidden layers, where m is an integer greater than 1. The above input layer is used to receive the adjacency matrix A, degree matrix D and feature matrix H of the material property relationship network. (0) The above hidden layers are used to extract node features through graph convolution calculations. The elements of the weight matrix of each hidden layer are initialized to random values. Taking mm equal to 5 as an example, the 1st to 5th hidden layers perform the following graph convolution calculations respectively:

[0184]

[0185] Where W (0) , W (1) , W (2) , W (3) and W (4) are the weight matrices of the 1st to 5th hidden layers, with sizes of (N+5)×(N+5); H (0) is the feature matrix; H (1) , H (2) , H (3) , H (4) and H (5) are the first feature matrix, second feature matrix, third feature matrix, fourth feature matrix and fifth feature matrix of the output of the 1st to 5th hidden layers, respectively, and all of them are N+5 rows and K columns, where the fifth feature matrix is ​​the above output feature matrix; σ is the nonlinear activation function ReLU; matrix is the sum of the adjacency matrix A and the identity matrix I; the matrix is the sum of the degree matrix and the identity matrix I.

[0186] It should be noted that, in the above output feature matrix, the 1st to Nth row vectors respectively correspond to the node feature vectors of the material nodes numbered 1 to N in the material property relationship network, and the row vectors of the N+1th to N+5th rows respectively correspond to the node feature vectors of the first attribute node, the second attribute node, the third attribute node, the fourth attribute node and the fifth attribute node.

[0187] S315, using the target value calculation model, processing the material property relationship network and the output feature matrix to obtain a target value L;

[0188] The target value calculation model is:

[0189]

[0190] Where M is the material property relationship network, m i is the i-th node of M, i is an integer from 1 to N+5; Ne(m i ) represents node m i The set of neighbor nodes in M, Pe(m i ) represents m i The set of non-neighbor nodes in M, m j and m k are the j-th and k-th nodes of M respectively, and m j and m k They belong to Ne(m i ) and Pe(m i );σ *is the sigmoid function, <·> represents the inner product of the calculated vector; δ is the preset negative sample adoption rate; v i 、v j and v k are the i-th, j-th and k-th row vectors in the output feature matrix respectively.

[0191] It should be noted that the above M=Ne(m i )∪Pe(m i )∪m i .

[0192] Preferably, the above δ is 0.5.

[0193] S316, adding 1 to the value of the number of training times.

[0194] S317, judging the target value;

[0195] When the target value is greater than or equal to the preset optimal value, updating the parameters of the feature extraction model to random values; executing S313;

[0196] When the target value is less than the optimal value, the optimal value is updated to the target value; the initial feature extraction matrix is ​​updated to the output feature matrix; and S318 is executed.

[0197] S318, determining the optimal value;

[0198] If the optimal value is greater than or equal to the preset optimal threshold, the parameters of the feature extraction model are updated to random values; and S313 is executed;

[0199] If the optimal value is less than the optimal threshold, the feature extraction matrix is ​​determined to be the initial feature extraction matrix, and S32 is executed.

[0200] Preferably, the above optimal threshold is 0.3.

[0201] S319: Determine the feature extraction matrix as the initial feature extraction matrix.

[0202] It should be noted that the above-mentioned updating of the parameters of the feature extraction model to random values ​​is to update the elements of the weight matrix of each hidden layer in the feature extraction model to random values.

[0203] In yet another optional embodiment, the above processing of the characteristic attribute matrix to obtain a priority evaluation result includes:

[0204] S41, using a competition matrix calculation model, processing the characteristic attribute matrix to obtain a competition matrix;

[0205] The competition matrix calculation model is:

[0206]

[0207] In the formula, E uv is the element in the uth row and vth column of the competition matrix C, e is the intermediate calculation result, x uk is the element in the uth row and kth column of the characteristic attribute matrix, x vk is the element in the vth row and kth column of the characteristic attribute matrix, where u, v and k are all integers from 1 to N;

[0208] S42, processing the competition matrix to obtain an evaluation vector;

[0209] It should be noted that the above processing of the competition matrix is ​​to calculate the eigenvector corresponding to the maximum eigenvalue of the competition matrix as an evaluation vector; the length of the above evaluation vector is N, and the lth element in the evaluation vector corresponds to the original evaluation value of the reserve material numbered l, and l is an integer from 1 to N.

[0210] S43, processing the evaluation vector to obtain the priority evaluation result.

[0211] In yet another optional embodiment, the above processing of the evaluation vector to obtain the priority evaluation result includes:

[0212] S431, normalizing the evaluation vector to obtain a normalized evaluation vector;

[0213] S432, establishing a tag sequence of length N; the tag sequence sequentially includes integers from 1 to N;

[0214] It should be noted that the numerical value of the element in the label sequence is used to represent the serial number of the corresponding reserve material.

[0215] S433, combining the normalized evaluation vector with the label sequence to obtain an initial evaluation result sequence;

[0216] It should be noted that the normalized evaluation vector is combined with the elements with the same number in the label sequence to obtain N evaluation result pairs. In each evaluation result pair, the element from the normalized evaluation vector is used as the evaluation value, and the element from the label sequence is used as the number of the reserve material corresponding to the evaluation value; the N evaluation result pairs together constitute the initial evaluation result sequence.

[0217] S434, sorting the initial evaluation result sequence to obtain a priority evaluation result.

[0218] It should be noted that the above sorting is to arrange the evaluation result pairs in order from large to small evaluation values ​​in the evaluation result pairs.

[0219] Take the case of only five types of reserve materials (compressed dry food g1, emergency rescue whistle g2, safety rope g3, cold-proof clothing g4, and rain boots g5) as an example. Assuming that the calculated normalized vector is [0.194, 0.236, 0.255, 0.136, 0.179], the label sequence is (1, 2, 3, 4, 5). The initial evaluation result sequence after combination is ((0.194, 1), (0.236, 2), (0.255, 3), (0.136, 4), (0.179,1)), and the priority evaluation results after sorting are ((0.255,3),(0.236,2),(0.194,1),(0.179,1),(0.136,4)), indicating that the priorities of the reserve materials from high to low are safety rope g3, emergency rescue post g2, compressed dry food g1, rain boots g5, cold-proof clothes g4, and the corresponding evaluation values ​​are 0.255, 0.236, 0.194, 0.179, 0.136 respectively.

[0220] It can be seen that the implementation of the material reserve priority evaluation method described in the embodiment of the present invention can comprehensively consider the substitutability, fundraising period, introduction rate, storability and economy, and conduct a quantitative evaluation of the priority of material reserves, which effectively makes up for the shortcomings of traditional material reserve priority evaluation methods and provides powerful decision-making support for material reserve managers.

[0221] Embodiment 2

[0222] See also Figure 2 .in, Figure 2 The described material reserve priority assessment device can be applied to the field of material reserve priority assessment, such as the reserve priority assessment of large-scale disaster relief materials, and the embodiments of the present invention are not limited thereto. Figure 2 As shown, the device may include a reserve material information acquisition module 201, a material characteristic relationship network generation module 202, a characteristic attribute matrix calculation module 203 and a priority evaluation module 204;

[0223] The reserve material information acquisition module 201 is used to acquire a reserve material information set.

[0224] The material property relationship network generation module 202 is used to process the reserve material information set to obtain a material property relationship network.

[0225] The characteristic attribute matrix calculation module 203 is used to process the material characteristic relationship network to obtain a characteristic attribute matrix.

[0226] The priority evaluation module 204 is used to process the characteristic attribute matrix to obtain a priority evaluation result.

[0227] Embodiment 3

[0228] See also Figure 3 , Figure 3 is a schematic diagram of the structure of another material reserve priority evaluation device disclosed in an embodiment of the present invention. Figure 3 The described material reserve priority assessment device can be applied to the field of material reserve priority assessment, such as the reserve priority assessment of large-scale disaster relief materials, and the embodiments of the present invention are not limited thereto. Figure 3 As shown, the material reserve priority evaluation device may include the following parts:

[0229] A memory 301 storing executable program codes;

[0230] a processor 302 coupled to the memory 301;

[0231] The processor 302 calls the executable program code stored in the memory 301 to execute the steps in the material reserve priority evaluation method described in the first embodiment.

[0232] Embodiment 4

[0233] An embodiment of the present invention discloses a computer-readable storage medium storing a computer program for electronic data exchange, wherein the computer program enables a computer to execute the steps of the material reserve priority evaluation method described in the first embodiment.

[0234] The device embodiments described above are only illustrative, wherein the modules described as separate components may or may not be physically separated, and the components displayed as modules may or may not be physical modules, i.e., they may be located in one place, or they may be distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative work.

[0235] Through the specific description of the above embodiments, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on such an understanding, the above technical solution can be essentially or partly contributed to the prior art in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, and the storage medium includes a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically erasable rewritable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, a magnetic disk storage, a magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0236] Finally, it should be noted that the material reserve priority assessment method and device disclosed in the embodiment of the present invention discloses only the preferred embodiments of the present invention, which are only used to illustrate the technical scheme of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical schemes described in the aforementioned embodiments can still be modified, or some of the technical features therein can be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical schemes from the spirit and scope of the technical schemes of the various embodiments of the present invention.

Claims

1. A material reserve priority assessment method, characterized in that: include: S1, obtaining a set of reserve material information; the set of reserve material information includes N pieces of reserve material information, where N is a positive integer; The reserve material information includes replacement information, fundraising period, number of foreign purchases, total number of purchases, storage time, material unit price, and set information set; The set information set includes set information; the set information includes serial numbers of the reserve materials belonging to the same category; S2, processing the reserve material information set to obtain a material property relationship network; the material property relationship network includes N material nodes, a first attribute node, a second attribute node, a third attribute node, a fourth attribute node, a fifth attribute node and a plurality of attribute edges; S3, processing the material property relationship network to obtain a property attribute matrix; S4, processing the characteristic attribute matrix to obtain a priority evaluation result.

2. The material reserve priority evaluation method according to claim 1 is characterized in that: The step of processing the reserve material information set to obtain a material property relationship network includes: S21, quantizing the replacement information to obtain a first associated information set; the first associated information set includes N first associated information; S22, processing the fundraising period to obtain a second associated information set; the second associated information set includes N second associated information; S23, processing the foreign purchase quantity and the total purchase quantity to obtain a third associated information set; the third associated information set includes N third associated information; S24, processing the storage duration to obtain a fourth association information set; the fourth association information set includes N fourth association information; S25, processing the material unit price to obtain a fifth associated information set; the fifth associated information set includes N fifth associated information; S26, processing the first association information set, the second association information set, the third association information set, the fourth association information set, the fifth association information set and the set information set to obtain a material property relationship network.

3. The material reserve priority evaluation method according to claim 2 is characterized in that: The processing of the fundraising period to obtain a second set of associated information includes: S221, obtaining the maximum value of the N fundraising periods to obtain the longest fundraising period; S222, obtaining the minimum value of the N fundraising periods to obtain the shortest fundraising period; S223, using a second association information calculation model, processing the longest fundraising period, the shortest fundraising period and each fundraising period to obtain corresponding second association information; In the formula, R i,2 is the second associated information corresponding to the i-th fundraising period, t 2,max is the longest fundraising period, t 2,min is the shortest fundraising period, t 2,i The i-th fundraising period, i being an integer from 1 to N; S224: Combine N pieces of the second associated information to obtain a set of the second associated information.

4. The material reserve priority evaluation method according to claim 2 is characterized in that: The processing of the storage duration to obtain a fourth set of associated information includes: S241, obtaining the maximum value of the N storage durations to obtain the longest storage duration; S242, obtaining the minimum value of the N storage durations to obtain the shortest storage duration; S243, using the fourth association information calculation model, processing the longest storage time, the shortest storage time and each storage time to obtain the corresponding fourth association information; In the formula, R j,4 is the fourth association information corresponding to the jth storage duration, t 4,max is the maximum storage time, t 4,min is the shortest storage time, t 4,j is the j-th storage duration, where j is an integer from 1 to N; S244: Combine the N fourth association information to obtain the fourth association information set.

5. The material reserve priority evaluation method according to claim 2 is characterized in that: The material unit price is processed to obtain a fifth associated information set, including: S251, obtaining the maximum value of the N material unit prices to obtain the highest material unit price; S252, using the fifth association information calculation model, processing the highest material unit price and each material unit price to obtain the corresponding fifth association information; In the formula, R k,5 is the fifth associated information corresponding to the kth material unit price, p max is the highest unit price of the materials, p k is the kth unit price of the material, where k is an integer from 1 to N; S253: Combine the N pieces of the fifth association information to obtain the fifth association information set.

6. The material reserve priority evaluation method according to claim 2 is characterized in that: The processing of the first association information set, the second association information set, the third association information set, the fourth association information set, the fifth association information set and the set information set to obtain a material property relationship network includes: S261, initializing the first relationship network to include only the first attribute node, the second attribute node, the third attribute node, the fourth attribute node, the fifth attribute node and N material nodes; S262, performing a first connection process on the first relationship network based on the first association information set, the second association information set, the third association information set, the fourth association information set, and the fifth association information set to obtain a second relationship network; S263: Based on the set information set, perform a second connection process on the second relationship network to obtain the material property relationship network.

7. The material reserve priority evaluation method according to claim 1 is characterized in that: The material property relationship network is processed to obtain a property attribute matrix, including: S31, extracting features from the material property relationship network to obtain a property extraction matrix; the number of rows of the property extraction matrix is ​​N+5; S32, splitting the feature extraction matrix to obtain a first feature vector sequence and a second feature vector sequence; the first feature vector sequence includes N feature vectors; the second feature vector sequence includes 5 feature vectors; S33, using a characteristic attribute calculation model, processing the first characteristic vector sequence and the second characteristic vector sequence to obtain the characteristic attribute matrix; The characteristic attribute calculation model is: Where X is the characteristic attribute matrix, x pq is the element in the pth row and qth column of the characteristic attribute matrix, where p is an integer from 1 to N, and q is an integer from 1 to 5; u p is the pth first eigenvector in the first eigenvector sequence, v q is the kth second eigenvector in the second eigenvector sequence.

8. The material reserve priority evaluation method according to claim 1 is characterized in that: The feature extraction of the material property relationship network to obtain a feature extraction matrix includes: S311, the number of training times and the optimal value are initialized to 0 and the preset initial optimal value, respectively, and the feature matrix H is randomly initialized (0) ; Initialize the initial feature extraction matrix to H (0) ; The feature matrix and the initial feature extraction matrix are both matrices of N+5 rows and K columns, where K is a preset feature length; S312, obtaining the adjacency matrix A and degree matrix D of the material property relationship network; S313, determining the number of training times; When the number of training times is greater than the preset total number of training times, executing S319; the total number of training times is greater than 1; When the number of training times is less than or equal to the total number of training times, executing S314; S314, using a preset feature extraction model, extracting the adjacency matrix A, the degree matrix D and the feature matrix H (0) Processing is performed to obtain the output feature matrix; S315, using the target value calculation model, processing the material property relationship network and the output feature matrix to obtain a target value L; The target value calculation model is: Where M is the material property relationship network, m i is the i-th node of M, i is an integer from 1 to N+5; Ne(m i ) represents node m i The set of neighbor nodes in M, Pe(m i ) represents m i The set of non-neighbor nodes in M, m j and m k are the j-th and k-th nodes of M respectively, and m j and m k They belong to Ne(m i ) and Pe(m i );σ * is the sigmoid function, <·> represents the inner product of the calculated vector; δ is the preset negative sample adoption rate; v i 、v j and v k are the i-th, j-th and k-th row vectors in the output feature matrix respectively; S316, adding 1 to the value of the number of training times; S317, judging the target value; When the target value is greater than or equal to the preset optimal value, updating the parameters of the feature extraction model to random values; executing S313; When the target value is less than the optimal value, the optimal value is updated to the target value; the initial feature extraction matrix is ​​updated to the output feature matrix; and S318 is executed; S318, determining the optimal value; If the optimal value is greater than or equal to the preset optimal threshold, the parameters of the feature extraction model are updated to random values; and S313 is executed; If the optimal value is less than the optimal threshold, determining the feature extraction matrix as the initial feature extraction matrix, and executing S32; S319: Determine the feature extraction matrix as the initial feature extraction matrix.

9. The material reserve priority evaluation method according to claim 1, characterized in that: The processing of the characteristic attribute matrix to obtain a priority evaluation result includes: S41, using a competition matrix calculation model, processing the characteristic attribute matrix to obtain a competition matrix; The competition matrix calculation model is: In the formula, E uv is the element in the uth row and vth column of the competition matrix C, e is the intermediate calculation result, x uk is the element in the uth row and kth column of the characteristic attribute matrix, x vk is the element in the vth row and kth column of the characteristic attribute matrix, where u, v and k are all integers from 1 to N; S42, processing the competition matrix to obtain an evaluation vector; S43, processing the evaluation vector to obtain the priority evaluation result.

10. A computer storage medium, characterized in that: The computer storage medium stores computer instructions, which, when called, are used to execute the material reserve priority assessment method as described in any one of claims 1-9.

Images