Path planning method, electronic equipment, storage medium and program product

Abstract

The embodiment of the invention provides a path planning method, electronic equipment, a storage medium and a program product. The method relates to the field of guiding path planning, and comprises the steps that a plurality of sub-networks are established, the plurality of sub-networks are generated based on information of candidate objects provided by corresponding cooperation mechanisms, and the plurality of sub-networks are stored in different network storage nodes in a block chain; under the condition that the single network path from the known object to the current target object does not exist in the plurality of sub-networks, respectively acquiring the plurality of sub-networks by each cross-network node set in the block chain; searching according to the plurality of sub-networks by adopting each cross-network node, and determining a target path from the known object to the current target object; and determining a final path by adopting a majority consensus mode based on the target paths obtained by the plurality of cross-network nodes respectively. According to the method and the device, the technical problems that the contradiction between the data security and the data sharing is difficult to balance in the related technology and the accuracy of the guiding path planning is not ideal are solved.

Description

Technical Field

[0001] This application relates to the technical field of data processing, and in particular, to a path planning method, an electronic device, a storage medium, and a program product. Background Art

[0002] In today's highly competitive social environment, talents are the core driving force for promoting scientific and technological progress and industrial development. How to accurately identify potential talent teams and design reasonable talent attraction paths for them is crucial. With the rapid development of technology, talent recruitment and talent attraction path planning have shifted from talent discovery methods based on experience and connections to data-driven and intelligent solutions.

[0003] In the related art, on the one hand, the inter-institutional trust cost is high, and complex technical interfaces need to be established to ensure data authenticity. The sub-network data of each institution cannot be safely interconnected, and cross-network path planning relies on cumbersome data synchronization protocols, which are inefficient and error-prone. Therefore, there is a problem that the accuracy of talent attraction path planning in the related art is not ideal.

[0004] In response to the above problems, no solutions have been proposed in the current related art. Summary of the Invention

[0005] Embodiments of this application provide a path planning method, an electronic device, a storage medium, and a program product to alleviate or solve the contradiction between data security and data sharing in the related art, and thus the technical problem that the accuracy of talent attraction path planning is not ideal.

[0006] In a first aspect, embodiments of this application provide a path planning method, including: Establish multiple sub-networks, where the multiple sub-networks are respectively generated based on information of candidate objects provided by corresponding cooperation institutions, and the multiple sub-networks are respectively stored in different network storage nodes in the blockchain; In the case where there is no single-network path from a known object to the current target object in the multiple sub-networks, each cross-network node set in the blockchain respectively obtains the multiple sub-networks, the known object is a candidate object with a contact path among the multiple candidate objects, and the current target object is an object that needs to establish a connection among the multiple candidate objects; Use each cross-network node to search according to the multiple sub-networks to determine a target path from the known object to the current target object; Based on the target paths respectively obtained by the multiple cross-network nodes, use the majority consensus method to determine the final path.

[0007] In a second aspect, embodiments of this application provide an electronic device, including a memory, a processor, and a computer program stored on the memory, where the processor implements the method according to any one of the embodiments of this application when executing the computer program.

[0008] In a third aspect, an embodiment of the present application provides a computer-readable storage medium, in which a computer program is stored, and when the computer program is executed by a processor, the method according to any one of the embodiments of the present application is implemented.

[0009] In a fourth aspect, an embodiment of the present application provides a computer program product, including a computer program, and when the computer program is executed by a processor, the method according to any one of the embodiments of the present application is implemented.

[0010] Based on the path planning method of the above first aspect, the present application has at least the following beneficial effects or advantages: The sub-network data (such as cooperation records, educational backgrounds) maintained by each cooperation institution is stored on the chain, and all modifications need to pass consensus verification. This prevents a single institution from forging or tampering with association relationships and ensures the data credibility of the talent introduction path dependence. The permission control of the blockchain allows cross-network nodes to securely access the sub-network data of other institutions. When cross-network nodes authorize access to heterogeneous sub-networks through smart contracts, zero-knowledge proof technology is used to verify query permissions without directly sharing the original database. Each cross-network node can determine the target path from a known object to the current target object based on multiple sub-networks. Each cross-network node independently calculates the target path based on the local topology and filters the final path through majority consensus. The redundant calculation mechanism can identify and eliminate the problem of tampered guidance of individual cross-network nodes, improving the robustness of the talent introduction path. By adding the blockchain mechanism, all sub-network information queries and cross-network node consensus votes are recorded on the chain, supporting the traceability of the path decision-making process and avoiding the drawbacks of closed talent introduction such as acquaintance recommendation.

[0011] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the following specific embodiments of the present application are specifically given. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] In the drawings, unless otherwise specified, the same reference numerals throughout the drawings denote the same or similar components or elements. These drawings are not necessarily drawn to scale. It should be understood that these drawings only depict some embodiments according to the present application and should not be regarded as limiting the scope of the present application.

[0013] Figure 1 The flowchart of the path planning method according to the embodiment of the present application is shown; Figure 2 The schematic diagram of the path planning method according to the embodiment of the present application is shown; Figure 3 The schematic block diagram of the path planning device according to the embodiment of the present application is shown; Figure 4 The block diagram of the electronic device provided by the embodiment of the present application is shown. Detailed implementation manners

[0014] In the following, only some exemplary embodiments are simply described. As those skilled in the art can recognize, the described embodiments can be modified in various different ways without departing from the concept or scope of the present application. Therefore, the drawings and the description are considered to be exemplary in nature and not restrictive.

[0015] To facilitate the understanding of the technical solutions of the embodiments of the present application, the related technologies of the embodiments of the present application are described below. The following related technologies can be arbitrarily combined with the technical solutions of the embodiments of the present application as optional solutions, and all of them fall within the protection scope of the embodiments of the present application.

[0016] It should be noted that the application scenarios or application examples provided in the embodiments of the present application are for the convenience of understanding, and the embodiments of the present application do not specifically limit the application of the technical solutions. In addition, the user information (including but not limited to user device information, user personal information, relevant information of multiple candidate objects, etc.) and data (including but not limited to data for analysis, stored data, displayed data, historical path records, data stored in the blockchain, etc.) involved in the present application are all information and data authorized by the user or fully authorized by all parties, and the collection, use, and processing of the relevant data need to comply with the relevant laws, regulations, and standards of the relevant countries and regions, and corresponding operation entrances are provided for the user to select authorization or rejection.

[0017] The technical solutions of the present application and how the technical solutions of the present application solve the foregoing technical problems are described in detail below with specific embodiments. The several specific embodiments listed can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of the present application will be described in detail below with reference to the drawings.

[0018] Figure 1 The flowchart of the path planning method according to the embodiment of the present application is shown. As Figure 1 shown, the method may include steps S101 to S104.

[0019] Step S101: Establish a plurality of sub-networks, the plurality of sub-networks are respectively generated based on the information of candidate objects provided by corresponding cooperation institutions, and the plurality of sub-networks are respectively stored in different network storage nodes in the blockchain; Step S102: When there is no single-network path from a known object to the current target object among multiple sub-networks, each cross-network node set in the blockchain respectively obtains multiple sub-networks. The known object is a candidate object with a contact path among multiple candidate objects, and the current target object is an object that needs to establish a connection among multiple candidate objects; Step S103: Use each cross-network node to search based on multiple sub-networks to determine the target path from the known object to the current target object; Step S104: Based on the target paths respectively obtained by multiple cross-network nodes, use the majority consensus method to determine the final path.

[0020] In the embodiment of the present application, the above execution entity can be a control node in the blockchain. The blockchain ensures the reliability and immutability of path information through its consensus mechanism. The decentralized feature of the blockchain makes the path planning process more transparent and trustworthy. At the same time, through smart contracts, the dynamic update of multiple sub-networks and the automatic path planning operation can be executed automatically and with high credibility. The permission of each above cross-network node is set to allow access to any network storage node in the blockchain.

[0021] In the embodiment of the present application, multiple sub-networks are respectively generated according to the candidate object information provided by different cooperation institutions. These sub-networks are stored on different network storage nodes of the blockchain. The information provided by different cooperation institutions, such as one institution provides academic cooperation data and another provides career association data, respectively generates corresponding cooperation relationship sub-networks and career association sub-networks, ensuring the diversity and independence of data sources, and using blockchain storage to ensure data security and immutability. When there is no path from the known object to the current target object within a single network, multiple cross-network nodes set in the blockchain start to work. These cross-network nodes have the right to access any network storage node in the blockchain, can obtain the information of each sub-network, and prepare for cross-network path search.

[0022] In the path search phase, each cross-network node searches based on multiple obtained sub-networks, starting from a known object to find multiple candidate paths leading to the current target object. A known object refers to a candidate object that has a connection path among multiple candidate objects. If talent information is understood as a kind of database, then a known object can be understood as in-stock talent. To evaluate the quality of these candidate paths, a path scoring mechanism is introduced. The calculation of the path score comprehensively considers the association strength between candidate objects and a predetermined attenuation factor. The association strength is dynamically adjusted based on historical path records. The historical path records include the historical paths used to successfully contact historical target objects, which can serve as prior knowledge of feasible paths. The above dynamic adjustment mechanism enables the system to continuously optimize the weight of the association relationship according to the actual situation. The attenuation factor is used to represent the way in which the path score decreases as the path length increases. As the path length increases, the number of intermediate candidate objects also increases, which in turn leads to a decrease in the reliability of the target object. By balancing the relationship between path length and path quality, problems such as a decrease in reliability caused by too many intermediate links are avoided. Path pruning is performed on multiple candidate paths according to the path score to select the optimal target path. After multiple cross-network nodes respectively obtain the target paths, the majority consensus method is used to determine the final path. If the target paths determined by most cross-network nodes are the same, or a certain path appears the most times, it is used as the final path, integrating the results of multiple nodes to improve the reliability and accuracy of the path.

[0023] Exemplarily, the above-mentioned cooperation institutions can be schools, enterprises, academic organizations, etc. In the application scenario of talent introduction, the above-mentioned multiple candidate objects can be represented as talents or experts. It should be noted that the sub-network data maintained by different cooperation institutions often involves information in different fields or types. Each cooperation institution can formulate personalized confidentiality measures for the edge server according to its own data characteristics and security requirements. The institution responsible for maintaining the educational background sub-network can adopt a more strict access control strategy for sensitive data such as student information. Only authorized personnel can access the talent data of specific educational levels or institutions, reducing the risk of sensitive data leakage and protecting the privacy of cooperation institutions and relevant personnel.

[0024] When a new cooperation institution is added, only a new sub-network smart contract needs to be deployed and the cross-node query interface needs to be updated, without reconstructing the global network. The modular design makes each sub-network relatively independent. The data of each cooperation institution is encapsulated in its own sub-network, and data acquisition and query are carried out through the edge server deployed on the side of the cooperation institution. The above encapsulation mechanism ensures that the newly added sub-network will not affect the existing sub-networks, and each sub-network can operate and be maintained independently. For example, the addition of an enterprise cooperation institution will not interfere with the sub-network of an existing academic cooperation institution. The two interact through cross-network nodes, but the internal data structure and management logic remain independent.

[0025] In the embodiments provided by the present application, in step S101: Establish multiple sub-networks, including the specific steps: For the association relationship of the cooperation type, based on the academic achievement information respectively corresponding to multiple candidate objects, determine the cooperation object corresponding to each candidate object among the multiple candidate objects; Determine the cooperation intensity between each candidate object and its corresponding cooperation object according to the cooperation times between each candidate object and its corresponding cooperation object, and the number of times the achievement is cited; Take multiple candidate objects as nodes, the connection line between each candidate object and its corresponding cooperation object as an edge, and the cooperation intensity as the edge weight of the edge, and establish a sub-network of the cooperation type.

[0026] In the embodiments provided by the present application, traverse multiple candidate objects. For each candidate object, analyze its academic achievement information (such as papers, projects, etc.), and extract other objects that jointly completed the academic achievement with this candidate object. These objects are the cooperation objects of this candidate object. For each candidate object and its corresponding cooperation object, count the cooperation times between the two. The cooperation times can be measured by the number of papers jointly published, the number of projects jointly participated in, etc. Take multiple candidate objects as nodes. For each candidate object and its corresponding cooperation object, take the connection line between them as an edge. Take the calculated cooperation intensity as the edge weight, and finally form a sub-network with candidate objects as nodes, the connection lines between cooperation objects as edges, and the cooperation intensity as the edge weight.

[0027] Exemplarily, in the process of constructing the sub-network of the cooperation relationship, extract the author cooperation information from the paper data, which includes analyzing the author list of each academic achievement to determine the author and his co-authors, and accordingly establish the relationship pairs between the author and the co-authors. The above-mentioned academic achievements include at least papers, patents, academic works (including software works), research reports, technical standards and specifications formulated by experts, and so on. When calculating the cooperation intensity, comprehensively consider the frequency of cooperation and the quality of cooperation results for network modeling. In the sub-network of the cooperation relationship (i.e., in the form of a graph), each candidate object is defined as a node. If two nodes have cooperated in at least one paper or patent, establish an edge between these two nodes, and the weight of the edge is used to represent the index of the cooperation intensity between the two, providing data support for the subsequent talent introduction path planning.

[0028] Exemplarily, the above calculation method of the cooperation intensity can be based on factors such as the frequency of cooperation, the quality of cooperation results (such as the number of paper citations, patent influence, etc.), so as to more accurately evaluate the relationship intensity between co-authors. For the method of allocating contribution weights according to the ranking of candidate objects in academic achievements, an example is given, and the specific calculation formula is as follows:

[0029] Among them, represents the weight of candidate object i, represents the sorting position of the candidate object (counting from 1), and n represents the total number of authors of the academic achievement. It emphasizes that the contributions of authors with higher rankings are greater.

[0030] The formula for calculating the average contribution degree of two authors in a single paper is as follows:

[0031] Among them, and are respectively used to represent different candidate objects, and represent the average contribution degree.

[0032] The above calculation method of the cooperation intensity can be expressed as: Cooperation intensity = number of co-authored papers × average total cited papers × average contribution degree + number of co-authored patents × technical influence; The above calculates the cooperation intensity between two collaborators on a specific cooperation project (paper or patent). Among them, the number of co-authored papers and the number of co-authored patents respectively represent the number of cooperation times on papers and patents, the average total cited papers and the average contribution degree respectively represent the average number of citations of papers and the average contribution degree of each author, and the technical influence represents the technical influence of patents.

[0033] Exemplarily, the cooperation intensity can also be corrected by using the citation decay factor. The above citation decay factor is used to represent the decay of the citation times of academic achievements over time, and can more accurately reflect the timeliness and influence of academic achievements. Over time, some early studies may gradually be replaced by new research results, and their citation times will also decrease accordingly. By introducing the decay factor, the actual influence of academic achievements in different time periods can be evaluated more reasonably. There are differences in citation behaviors and citation half-lives in different disciplines. The citation decay factor can be adjusted according to the characteristics of disciplines to better adapt to the citation rules of different disciplines. It is preferably set that the cooperation weight in the recent 3 years is 1 and the weight from 3 to 5 years is 0.7.

[0034] According to the embodiments provided by the present application, in step S101: Establish a plurality of sub-networks, including the specific steps: For the association relationship of the education background matching relationship type, based on the education background information respectively corresponding to a plurality of candidate objects, determine the background matching objects associated with the education background of each candidate object among the plurality of candidate objects; Determine the association strength between each candidate object and its corresponding background matching object based on the degree of overlap in educational backgrounds. Using multiple candidate objects as nodes, the connection line between each candidate object and its corresponding background matching object as an edge, and the association strength as the edge weight of the edge, establish a sub-network of the type of educational background matching relationship.

[0035] In the embodiments provided in this application, based on the extracted educational background information, other candidate objects with similar or identical educational backgrounds to each candidate object are identified among multiple candidate objects, and these objects are the background matching objects. For each candidate object and its corresponding background matching object, calculate the association strength according to the degree of overlap in educational backgrounds. The degree of overlap may include the same graduating school, the same major, and the existence of the same courses across majors, etc. The association strength can be quantified through a set algorithm. For example, different weights can be given based on the amount and importance of the common educational background, and then the numerical value of the association strength can be calculated. Using multiple candidate objects as nodes in the network, if there is an educational background matching relationship between two nodes (i.e., two candidate objects), establish an edge between these two nodes. Use the calculated association strength as the edge weight to represent the degree of educational background matching between the two candidate objects, and obtain a sub-network of educational background matching.

[0036] Exemplarily, extract direct classmate relationships from the educational background information. If two candidate objects study in the same school, the same major, the same grade, or the same class, there is a direct classmate relationship between them. Indirect classmate relationships consider a broader educational association. For example, talents in different majors or different grades may also have an indirect classmate relationship if they participate in the same courses, projects, or activities.

[0037] Preferably, different weights can be set according to the classmate relationship. For example, classmates in the same class are set to 1.0, students with the same supervisor but different graduation years are set to 0.7, and students participating in the same project at school are set to 0.5.

[0038] In the embodiments provided by this application, expert team discovery can also be performed based on the above-mentioned multiple sub-networks. Analyze the collected data to discover potential talent teams. By constructing an analysis model and using scientometric methods, such as paper contribution, H-index, G-index, etc., and leveraging the graph structure characteristics of the knowledge graph, perform operations such as path analysis and community detection to discover expert talent teams. The H-index and G-index are two indicators used to measure the influence of academic researchers, obtained based on the number of papers published by the researcher and the frequency of citation of these papers. Among them, the H-index represents a hybrid quantitative indicator that takes into account both the number of papers published by the researcher and the number of citations of these papers. The G-index is an improved version of the H-index, which also considers the number of papers and the number of citations, but focuses more on the influence of highly cited papers. Talents with a contribution greater than a certain threshold and more than 30 papers / patents published are regarded as expert talents. By using graph traversal algorithms such as depth-first search (DFS) and breadth-first search (BFS) to identify the three-degree relationship network of the expert, those with a contribution greater than a certain threshold (the contribution threshold is dynamically calculated: the 30th percentile in the field) are retained as the core talents of the team.

[0039] Exemplarily, the team contribution index can be expressed in the following way: Contribution = 0.4×H-index + 0.3×Patent authorization rate + 0.3×Proportion of international cooperation papers; The work number statistics include the number of papers published in the past 5 years, the number of patent authorizations, and the number of major project participations. The KeyBERT algorithm can be used to extract keywords from the paper abstracts and patent claims to generate team technology tags. KeyBERT (Keyword Extraction with BERT) is a keyword extraction algorithm based on the BERT (Bidirectional Encoder Representations from Transformers) model. The keyword correlation score is calculated by the following formula:

[0040] where, represents the candidate keyword, represents the document content. The higher the semantic similarity score between the word t and the document d calculated by the BERT model, the stronger the topic relevance between the word t and the document d. With the real-time processing ability of big data, continuously monitor and update the collaborator information of talents to ensure the timeliness and accuracy of the collaborator network. As new data is added, continuously expand and optimize the knowledge graph so that it can dynamically reflect the changes and development trends of the talent cooperation network.

[0041] Figure 2The figure shows a schematic diagram of the path planning method according to an embodiment of the present application. As Figure 2 shown, it exemplifies the relationship among teams with a contribution degree of 91.2, 32 papers, and 23 patents (including 5 PCT patents). The technical labels of this team include gene editing and immune cells. Figure 2 The size of the circles in [Figure] represents the importance of different objects. Object 7 represents the core figure of the team.

[0042] Exemplarily, expert certification criteria are set through a smart contract (such as H-index ≥ 30, number of patents ≥ 20, proportion of international collaborative papers ≥ 40%). Those who meet the conditions are automatically marked as expert nodes. Certification data (such as papers and patents) need to be verified by blockchain consensus and then uploaded to the chain to ensure authenticity. The identity information, certification records, and technical labels of expert nodes are all stored on the chain. Any modification requires node verification by a majority consensus method.

[0043] In the comprehensive network, the edge weights between expert nodes are automatically increased (such as the ordinary cooperation intensity × 1.5), and high-quality paths are preferentially constructed through the expert network. When searching for cross-network nodes, if there are expert nodes in the target path (such as A → expert → B), a fast consensus channel is directly triggered to shorten the voting time. The above fast consensus channel is implemented through a smart contract and is only activated when the path contains expert nodes. If the final path contains expert nodes and the talent recruitment is successful, the smart contract automatically increases the association strength of the expert node. If the path associated with the expert node fails to execute a predetermined number of times (such as 3 times without response), automatic weight reduction (weight × 0.8) is triggered, and it can also be decided by voting whether to cancel the expert certification.

[0044] In the embodiment provided by the present application, multiple sub-networks respectively correspond to edge servers, and the permissions of the edge servers are set to allow access to the network storage nodes in the blockchain corresponding to the sub-networks. After performing step S101: establishing multiple sub-networks, the method further includes the following specific steps: Use multiple edge servers to search in the corresponding sub-networks respectively to determine whether there are single-network paths; In the case of the existence of a unique single-network path, determine the single-network path as the final path; In the case of the existence of multiple single-network paths, determine the path scores corresponding to the multiple single-network paths respectively; Determine the one with the highest path score among the multiple single-network paths as the final path.

[0045] In the embodiments provided in this application, multiple sub-networks respectively correspond to edge servers, and the edge servers are granted permissions to access the network storage nodes of the corresponding sub-networks in the blockchain. Each edge server conducts a search in the corresponding sub-network to determine whether there is a single-network path from a known object to the current target object. In the academic cooperation sub-network, check whether there is a direct or indirect cooperation path from a known academic staff to the target academic staff. If only one single-network path is found during the search process, then this path will be directly determined as the final path. When there are multiple single-network paths, further evaluation of these paths is required to determine the corresponding path scores for each single-network path.

[0046] Through the above processing, the edge server directly conducts a search in the corresponding sub-network, avoiding the concentration of all search tasks on a central node for processing.

[0047] Exemplarily, the calculation of the above path score can comprehensively consider multiple factors, such as the credibility of the association relationship, the association strength, the path length, etc. between candidate objects. The priority of the known object serving as the starting point of the path can also be incorporated into the calculation of the path score. For example, for a known object with stable cooperation experience, there is a greater tendency to assist in the talent introduction task. Therefore, a candidate path starting from it can have a higher path score compared to a candidate path starting from a known object without a historical cooperation record and other conditions being equal. After obtaining the scores of each path, select the path with the highest score as the final path.

[0048] Exemplarily, each edge server can only access the corresponding blockchain sub-network node through a dedicated channel (for example, the education network edge server is only connected to the education sub-network node). The isolation at the physical network level ensures that even if a certain server is invaded, the attacker cannot obtain data from other sub-networks through lateral movement within the intranet. Compared with the solution provided by the related technology, in a centralized server cluster, once the boundary protection is breached, all data is at risk of leakage. Further, the edge servers of different sub-networks use independent communication certificates and VPN (Virtual Private Network) tunnels. Even if the same cooperation institution manages multiple servers, cryptographic isolation can still prevent crosstalk. The above method of setting edge servers can effectively solve the practical requirements of sharing minutes while isolating in cross-institution data collaboration.

[0049] Exemplarily, the search priorities of multiple sub-networks can also be determined according to the object type of the target object. Taking the sub-network of the cooperation relationship and the sub-network matching the educational background as an example, when the object type of the target object is the student type, the sub-network matching the educational background is determined as the highest priority among the multiple sub-networks. When the object type of the target object is a non-student type, the sub-network of the cooperation relationship is determined as the highest priority among the multiple sub-networks. Since the association types represented by the multiple sub-networks are different, staged single-network searches can be performed according to the above object types, reducing the number of edge servers for starting calculations and effectively reducing the duration of waiting for calculation feedback.

[0050] For cross-network search requirements, such as combining the collaborator network and the classmate network, multiple sub-networks can be integrated.

[0051] According to the embodiment provided by the present application, in step S103: each cross-network node is used to search based on multiple sub-networks to determine the target path from the known object to the current target object. Each cross-network node can similarly execute the following steps: Fuse multiple sub-networks to obtain a comprehensive network; the nodes included in the comprehensive network are multiple candidate objects, and the edges included in the comprehensive network represent different types of association relationships among the multiple candidate objects; Adopt a path search strategy to search the comprehensive network and record the paths passed during the search as multiple candidate paths. The path search strategy is determined based on the network scale of the comprehensive network; Use each cross-network node to determine the path scores corresponding to the multiple candidate paths according to the association strength between the multiple candidate objects and a predetermined attenuation factor; the attenuation factor is used to represent the change mode in which the path score of the corresponding candidate path decreases as the path length increases. The path length represents the number of candidate objects passed through, and the association strength is adjusted based on the historical path record. The historical path record includes the historical paths used to successfully contact the historical target object; Based on the path scores corresponding to the multiple candidate paths, perform path pruning on the multiple candidate paths to determine the target path obtained by each cross-network node.

[0052] In the embodiments provided by this application, the nodes in the integrated comprehensive network represent multiple candidate objects, while the edges represent different types of association relationships between these candidate objects. The comprehensive network can comprehensively reflect various association relationships between candidate objects, such as co-operator relationships, educational background matching relationships, and so on. The scale of the network involves factors such as the number of nodes and the density of edges, and these factors will affect the choice of search strategies. For example, for a large-scale network, a heuristic search algorithm may be adopted to improve efficiency; while for a smaller-scale network, a more comprehensive search algorithm can be used to ensure accuracy. The comprehensive network is searched using a path search strategy, and the paths passed through during the search process are recorded, and these paths are used as multiple candidate paths. By fusing multiple sub-networks to obtain a comprehensive network, different types of association relationships can be comprehensively integrated, the advantages and disadvantages of candidate paths can be evaluated more accurately, and a richer information basis is provided for path search. Dynamically selecting a path search strategy according to the scale of the comprehensive network can flexibly adjust the search method in networks of different scales. The above dynamic adaptation mechanism ensures the search efficiency and is conducive to improving the accuracy of path search.

[0053] By considering the association strength and the attenuation factor to calculate the path score, the advantages and disadvantages of the path can be evaluated more accurately. The association strength is adjusted based on historical path records, making the path selection more well-founded and enabling the preferential selection of reliable paths that have been verified. At the same time, the introduction of the attenuation factor avoids the selection of overly long paths, improving the practicality and accuracy of the path. The above historical path records provide practical path examples. Even if it is not for the same target object, the intermediate objects are proven to be reliable, providing timeliness information about the relationships between the candidate objects included in the path. Some intermediate objects played a key role in past successful cases, indicating that they may still have high reliability at the current time point. By analyzing the path characteristics of successful cases, it is possible to extract which path structures, combinations of intermediate objects, or association relationships were effective in the past. Using path pruning techniques can effectively reduce the search space, avoid unnecessary path exploration, and thus improve the efficiency of path search. By pre-judging the potential of nodes, a large amount of computational overhead for analysis and expansion is avoided.

[0054] Exemplarily, for historical path records, it can be regarded as using backpropagation based on successful cases. By analyzing the path characteristics of successful talent recruitment cases, the weight parameters are adjusted backward to optimize the path scoring model. If an intermediate object played a key role in multiple successful cases, the weight of that object can be increased; conversely, if an intermediate object frequently appears in failure cases, its weight can be decreased. The update period for the weights can be set as needed. For example, according to real-time data feedback, the weight parameters are dynamically adjusted to ensure the accuracy and timeliness of the path scoring model.

[0055] Exemplarily, the above-mentioned predetermined attenuation factor can be set in various ways, such as exponential attenuation, linear attenuation, etc., and the attenuation function can also be custom-set according to requirements. The linear attenuation formula can be expressed as: attenuation factor = 1 - α × path length; where α is the attenuation coefficient, which is a positive number less than 1. The path score of linear attenuation decreases linearly with the increase of the path length. The exponential function formula can be expressed as , where is the base of the natural logarithm,[[]]END]] represents a predetermined coefficient,[[]]END]] represents the path length. The above-mentioned predetermined coefficient is preferably set to 0.15, and can be obtained by fitting data through historical path records, which can better balance the relationship between the path length and the path quality.[[]]END]]

[0056] Exemplarily, the comprehensive network is represented as , where is the set of nodes (representing talents),[[]]END]] is the set of edges (representing cooperation relationships or classmate relationships). The goal of the secondary path search is to find a path from the starting node to the target node . The path can be represented as a node sequence , where , and there is an edge connection between adjacent nodes. The length of the path can be measured by the number of edges or the sum of the weights of the edges, and the weight can represent the tightness of the cooperation relationship, etc.[[]]END]]

[0057] For the network scale less than the predetermined threshold, it is regarded as being applied to a small-scale comprehensive network. In the optional embodiment provided by the present application, a path search strategy is adopted to search the comprehensive network, including the following methods:[[]]END]] When the network scale is less than the predetermined threshold, it is determined that the path search strategy is: starting from the starting node representing the known object in the comprehensive network, searching whether there is a target node representing the target object within the predetermined range;[[]]END]] When there is no target node within the predetermined range, search based on the adjacent nodes of the nodes within the predetermined range, and determine whether there is a target node among the adjacent nodes;[[]]END]] When there is no target node among the adjacent nodes, start with the adjacent nodes as the new starting node and search the nodes within the predetermined range;[[]]END]] Repeat the above process until a target node exists within the predetermined range or among the adjacent nodes, and stop the search process.[[]]END]]

[0058] In the embodiments provided by the present application, the search process starts from the starting node representing the known object, and first looks for the target node representing the target object within a predetermined range. If the target node is not found within the predetermined range, the search continues based on the adjacent nodes of the nodes within the current range. It can be understood as a breadth-first search to determine whether there is a target node among these adjacent nodes. The set of nodes that can be reached through the breadth-first search ensures that a certain range of nodes is quickly covered in the initial stage, that is, the nodes within a predetermined number of hops from the starting node are explored first.

[0059] If the target node is not found within the predetermined range, the search strategy changes to continue the search based on the adjacent nodes of the nodes within the current range. Based on the nodes at the current level, further explore their adjacent nodes, which belongs to a depth-first search, that is, preferentially explore a certain branch path in depth, rather than evenly distributing the search resources among multiple branches. The search process can explore the network structure more deeply to find the target node.

[0060] If the target node is still not found among these adjacent nodes, then these adjacent nodes are used as the new starting nodes, and the nodes within the predetermined range are searched again. By continuously expanding the search boundary, it combines the comprehensiveness of the breadth-first search and the in-depthness of the depth-first search. By using the adjacent nodes as the new starting nodes, the search process can gradually expand the search breadth while exploring the depth, ensuring that no possible path is missed. The above process will be repeated continuously until the target node is found within the predetermined range or among the adjacent nodes, at which point the search process stops.

[0061] Exemplarily, searching for nodes within the predetermined range belongs to the breadth-first search (BFS), and searching for the branch where the adjacent nodes are located belongs to the depth-first search (DFS). DFS and BFS are two graph search algorithms.

[0062] The breadth-first search (BFS) starts from the starting node and expands the nodes layer by layer to ensure finding the shortest path. The search process first initializes a queue and enqueues the starting node s. When the queue is not empty, dequeue a node v in turn and check whether this node is the target node t. If it is the target node, return the found path; if not, enqueue all the unvisited adjacent nodes of this node for further exploration in subsequent steps. The layer-by-layer expansion method of BFS can ensure that when the target node is found, the path found is the shortest path, and it is suitable for use when the path is short or the network structure is shallow.

[0063] Depth-First Search (DFS) starts from the starting node. DFS will explore as deeply as possible along a path until the target node is found or a node where no further progress can be made is reached. During the search, if the current node v is the target node t, the found path is returned. Otherwise, DFS traverses all adjacent nodes u of the current node v and recursively performs the DFS operation on each adjacent node. The depth-first search method can quickly explore paths deeply and is suitable for finding possible paths in complex networks, especially when the paths are long or the network structure is deep.

[0064] In the embodiments provided in this application, DFS and BFS can be combined to make full use of their respective advantages. First, use BFS to expand nodes layer by layer to quickly cover the local area around the starting node and ensure that no possible paths are missed. In each step of BFS, for each dequeued node, DFS can be used to deeply explore the path of this node until the target node is found or no further progress can be made. If the target node is found during the DFS process, the found path is returned; if the target node is not found after all nodes in the current layer have been explored, continue the layer-by-layer expansion of BFS until the target node is found.

[0065] Exemplarily, during the search, according to certain heuristic rules, the Breadth-First Search BFS and the Depth-First Search DFS can be alternately used, and the switching strategy can also be determined according to the current search state. For example, the condition for switching from BFS to DFS can be further set to meet any of the following: when a predetermined number (such as 3) of high-weight nodes (for example, the cooperation strength is greater than 80) continuously appear in the path, or the target is not found when BFS reaches a predetermined range (such as the node range within 4 hops).

[0066] The condition for switching from DFS to BFS can be further set to meet any of the following: the DFS recursive depth is greater than a predetermined number of hops, or the path score growth is less than a predetermined amplitude (regarded as the path growth stagnating).

[0067] For the sake of easy understanding, the following example is given. For the switch from BFS to DFS, when BFS discovers that the cooperation strength of 3 consecutive nodes > 80, it is regarded as discovering a strong relationship chain, that is, three consecutive high-weight connections of "cooperation strength 85 → 92 → 88" appear on the search path, and immediately switch from breadth expansion to depth-first, and deeply explore the subsequent nodes along this high-quality path. Or when BFS expands to 4 hops and still does not find the target, such as after four-layer node expansion from the starting node (that is, after passing through four-hop nodes), the target talent is still not reached, then switch to the DFS mode and preferentially explore the branch with the highest weight in the existing path.

[0068] To avoid ineffective in-depth search, when the DFS recursive depth > 10 hops, it is regarded as having traced three levels of relationships along a certain path (such as 1→2→3), and the target node has not been found yet, then it is determined as over-in-depth search, and the BFS mode is automatically switched back to find a new direction.

[0069] Or when the growth of the path score stagnates, such as the cumulative score growth in the recent 3 hops < 5%, it is regarded that during the extension of a certain path, the continuously newly added "cooperation intensity 62→58→65" only increases the total score from 210 to 215 (a growth of 2.3%), and the system determines that the quality of this path has declined, terminates the DFS and returns to the BFS.

[0070] For a network size greater than or equal to a predetermined threshold, it is regarded as being applied in a large-scale integrated network. In the optional embodiments provided in this application, a path search strategy is adopted to search the integrated network, including the following methods: When the network size is greater than or equal to a predetermined threshold, the path search strategy is determined as: starting the search from the starting node representing the known object in the integrated network; Based on the actual path cost from the starting node to the current node and the remaining path cost, determine the next node that minimizes the cumulative path cost from the starting node to the target node; the target node is used to represent the target object in the integrated network, and the remaining path cost is used to estimate the cost required to reach the target node from the current node; Process in the way of determining the next node until the new next node is the target node, and stop executing the search process.

[0071] In the embodiments provided in this application, each cross-network node guides the search direction by dynamically evaluating the path cost to ensure finding the optimal talent introduction path in the most efficient way. The search process starts from the starting node representing the known object, and will continuously track the actual path cost from the starting point to the current node. The actual path cost comprehensively considers key factors such as the strength of the association relationship and the timeliness of the relationship. By combining the actual path cost with the remaining estimated cost, the selection priority of each next node is further judged. In each step of the search, the node that minimizes the cumulative path cost will be selected as the next search target. The above process is continuously iterated until finally reaching the target node representing the target talent. Through the above method, paths can be effectively searched and optimized in a complex network, providing strong support for talent discovery and talent introduction path planning.

[0072] Exemplarily, the above remaining path cost is calculated by using a heuristic search algorithm, and there can be various such heuristic search algorithms, such as the A* algorithm, the IDA* algorithm (iterative deepening A* algorithm), etc. The A* algorithm uses a heuristic function Estimate the cost from the current node v to the target node t. The IDA* algorithm searches for a path by restricting the search depth and gradually increasing the depth limit. In each iteration, IDA* uses a heuristic function to limit the total cost of the current path , so as to reduce memory usage while ensuring the shortest path is found, and has good adaptability for the case where the network scale of the integrated network reaches a certain scale. Preferably, the above A* algorithm is applied to an integrated network with less than 100,000 nodes, and the above IDA algorithm is applied to an integrated network with more than 100,000 nodes.

[0073] The following takes the A* algorithm as an example for illustration. The A* algorithm uses a heuristic function to estimate the cost of the node to the target node , and combines the actual path cost to calculate the total estimated cost:

[0074] wherein, is the actual path cost from the starting node s to the node v, is the heuristic function, which can be based on the Euclidean distance between nodes, the accumulation of cooperation intensity, etc. During the search process, the node with the smallest is preferentially expanded as the next node, so as to find the optimal path more efficiently.

[0075] According to the embodiments provided in the present application, during a single pathfinding process for a node representing a target object, the search range will be gradually expanded in the integrated network according to a certain number of hops, and path pruning will be performed on multiple candidate paths based on the path scores respectively corresponding to the multiple candidate paths to obtain a target path, which may include the following steps: For the current expanded search, sort the multiple candidate paths according to the path scores respectively corresponding to the multiple candidate paths; Retain the candidate paths whose rankings meet the predetermined conditions to obtain the current search result of the current expanded search; the current search result is the starting path for the next expanded search; Until any path in the new search result includes the target node, the retained candidate paths are used as the target path.

[0076] In the current expanded search stage, the multiple candidate paths will be sorted according to the path scores of each candidate path. The path score is a comprehensive index that reflects the overall quality of the path. Retaining those candidate paths whose rankings meet the predetermined conditions constitutes the search result of the current expanded search, which will be used as the starting path for the next stage of expanded search. The search process is iterative, and each round of search is based on the search result of the previous round. This process will be repeated continuously until at least one path in the new search result contains the target node.

[0077] Exemplarily, the above-mentioned screening of multiple candidate paths according to predetermined conditions is a pruning process. Pruning is dynamically performed according to the path scores, and the top several paths with the highest scores are retained. The path score can be defined as:

[0078] wherein, represents the edge weight from node to node and factors such as the strength of the cooperation relationship, the quantity and quality of cooperation results can be comprehensively considered. During the search process, a priority queue is maintained, sorted according to the path scores, and the path with the highest score is expanded each time. When a path reaching the target node is found, the top k paths with the highest scores are retained.

[0079] It should be noted that each cross-network node has an independent digital identity, which is generated and stored through the encryption technology of the blockchain. The identity verification mechanism ensures that the identity of the cross-network node cannot be tampered with and is easy to verify. The permissions of the cross-network nodes are managed through smart contracts. The smart contract defines the range of sub-networks, data types, and operation permissions that the node can access. Each time a node accesses a sub-network, it needs to pass the authorization verification of the smart contract. To further enhance security, the cross-network nodes use end-to-end encryption technology when transmitting data. All data is encrypted before being sent, and only the target node can decrypt it, ensuring that the data is not stolen or tampered with during the transmission process. In some cases, when the cross-network node needs to calculate encrypted data, homomorphic encryption technology can be used to allow the node to calculate the data without decrypting it, thereby protecting the privacy of the data. During the path search process of the cross-network node, all operations are automatically executed through the smart contract. Each time the cross-network node performs data access or path calculation, the smart contract automatically verifies whether its operation complies with the predefined rules, and the automatic verification mechanism ensures the compliance of the node's behavior and prevents malicious operations. The code of the smart contract is publicly transparent on the blockchain, ensuring the fairness and transparency of the operation.

[0080] In the embodiment of the present application, the historical path records are stored in the historical nodes set in the blockchain, and the method further includes the following steps: When the talent recruitment result for the current target object according to the final path is successful, verify the credibility of the talent recruitment result; When the talent recruitment result passes the credibility verification, store the final path in the historical node and update the historical path record.

[0081] In the embodiments provided by the present application, historical path records are stored in specially set historical nodes in the blockchain. The characteristics of the blockchain ensure the security, immutability, and traceability of these historical data. The path information of successful talent recruitment in the past is properly saved to provide a basis for subsequent adjustment of association strength and path planning. When the talent recruitment result for the current target object shows success according to the final path, the path will not be immediately recorded in the historical node. Instead, the credibility of the talent recruitment result will be verified first. This verification process can be achieved through various methods, such as verifying key information during the talent recruitment process, like whether a cooperation intention has truly been reached with the target object and whether relevant agreements are authentic and valid. If the talent recruitment result passes the credibility verification, it indicates that this talent recruitment path is real and effective. At this time, the final path is stored in the historical node, and the historical path record is updated so that when subsequent path planning and association strength calculation are carried out, the latest successful talent recruitment experience can be taken into account.

[0082] Historical path records are an important basis for adjusting the association strength between candidate objects and can more reasonably reflect the actual association degree between different objects. When new successful talent recruitment paths are correctly recorded and updated in the historical path record, the impact of these successful cases on the association relationship can be more accurately reflected when adjusting the association strength, making the association strength more in line with the actual situation, and thus optimizing the result of path planning. Using the historical nodes of the blockchain to store and manage talent recruitment path records, combined with the credibility verification mechanism, further enhances the security and reliability of the entire system. The immutability of the blockchain ensures that historical records will not be maliciously tampered with, while credibility verification ensures the authenticity of the records. This dual guarantee makes the system more stable and trustworthy when handling important tasks such as talent recruitment path planning.

[0083] According to some embodiments provided by the present application, there are multiple known objects, and the identity information of the multiple known objects is stored in the inventory talent nodes in the blockchain in a predetermined order. The method further includes the following steps: Based on the email domains, default records, and response delays respectively corresponding to the multiple known objects, determine the priorities of the multiple known objects. The priorities are used to indicate the path search order based on the multiple known objects. The default record is a predetermined discreditable event recorded by the smart contract in the blockchain; Update the predetermined order stored in the inventory talent nodes according to the priorities of the multiple known objects.

[0084] In the embodiments provided by the present application, different email domain names represent different institutions, reputations or activity levels. For example, the email domain names of some professional institutions imply that the known object is more professional and reliable, and is more worthy of being given priority in path search. A default record is a predetermined discreditable event recorded by a smart contract in the blockchain. The immutable feature of the smart contract ensures the authenticity and reliability of the default record. A known object with a default record will be demoted in the priority ranking to reduce the impact on the effectiveness and reliability of the talent recruitment path. Response latency reflects the response speed of the known object to relevant information. A known object with a lower response latency means that it is more proactive and can participate in the talent recruitment process faster, so it will be ranked higher in the priority ranking. According to the determined priority of the known object, the predetermined order stored in the inventory talent node is updated. This enables the known objects to be used for search in sequence from high to low priority during subsequent path search, improving the efficiency and success rate of path search.

[0085] By determining the priority of the known object and performing path search according to the priority, more reliable and proactive known objects can be used preferentially. The time and resources wasted on low-quality known objects are reduced, enabling the path search to find a suitable path faster and improving the overall search efficiency. Considering factors such as default records, the use of known objects with bad records is avoided, thereby reducing the likelihood of risks occurring during the talent recruitment process. Preferentially selecting known objects with good reputations and positive responses can improve the reliability and success rate of the talent recruitment path.

[0086] Exemplarily, according to a predetermined verification period, as information such as the email domain name, default record and response latency of the known object changes, the priority is adjusted accordingly. This enables the predetermined order of the known objects in the inventory talent node to dynamically adapt to changes in the actual situation, always maintaining the optimal path search order, and further enhancing the performance and adaptability of the system.

[0087] According to the embodiments provided by the present application, for step S104: Based on the target paths respectively obtained by multiple cross-network nodes, the final path is determined by adopting a majority consensus method. A cross-network node can be regarded as a special node authorized to access all sub-network data (with the highest authority in the blockchain), and its identity can be verified through a smart contract. Each cross-network node performs parallel search based on the fused sub-network data (i.e., the integrated network) and can obtain one or more target paths. For example, node α submits: path P1 (score 92), path P2 (score 85), and node β submits: path P1 (score 90), path P3 (score 88).

[0088] The smart contract collects the target paths and their scores written into the blockchain by each cross-network node, and adopts the majority consensus voting method. The target path with the highest number of votes is recorded as the final path. Each cross-network node processes the vote for each target path separately, and a target path that obtains a weighted approval vote greater than the first predetermined threshold (such as greater than 50%) and the number of disapproval votes is less than the second predetermined threshold (such as less than 30%) is determined as the final path. If no path meets the standard after voting, the first predetermined threshold and / or the second predetermined threshold are modified to allow the submission of the target path with the second-highest score for re-voting. The obtained final path is processed for uploading to the blockchain. Optionally, the following information can also be stored on the blockchain: the list of nodes participating in the voting, the score details of each target path, and the voting weight distribution. Through the finality mechanism of the blockchain, it is ensured that the final path result cannot be rolled back.

[0089] According to the above embodiments, the present application also provides an alternative implementation manner to achieve secure sharing and intelligent collaboration of multi-institutional data through blockchain technology. The system adopts a hierarchical architecture design, including three core components: a data layer, a computing layer, and an application layer. In the data layer, each cooperating institution maintains an independent professional sub-network, including a cooperation relationship network, an educational background network, a project cooperation network, etc. After the data of these sub-networks is standardized, it is stored in the dedicated nodes of the blockchain, forming a physically isolated data storage unit. Each sub-network node is equipped with a dedicated edge server responsible for the path search task within the network. This design not only ensures data privacy but also realizes the reasonable distribution of computing resources.

[0090] The system working process is divided into two main stages: network initialization and path discovery. In the network initialization stage, each cooperating institution submits expert data through a standardized interface, and the smart contract automatically verifies the data format and constructs an association network. The cooperation network takes scholars as nodes and calculates the edge weights according to indicators such as the number of co-authored papers and citation times; the education network constructs an alumni relationship graph to distinguish the association strengths at different levels. All network data is processed through an encryption algorithm and distributedly stored in the blockchain nodes. Any data update requires the consensus confirmation of the majority of verification nodes.

[0091] When a talent recruitment request is initiated, the system first performs a local search on the edge server of the relevant sub-network. If no single-network path is found, the search of the cross-network nodes will be triggered. Each cross-network node will integrate multiple sub-networks into a comprehensive network. According to the network scale of the comprehensive network, for a comprehensive network with a network scale smaller than the predetermined threshold, a hybrid search strategy combining depth-first search and breadth-first search is adopted. For a comprehensive network with a network scale greater than or equal to the predetermined threshold, an improved A* search algorithm can be adopted, comprehensively considering the actual path cost of the path and the remaining path cost to find the target path.

[0092] To ensure the credibility of the results, the system has designed a multi-level consensus verification mechanism. The target paths submitted by each cross-network node need to go through cluster analysis and weighted voting. Only the target paths recognized by the majority of verification nodes will be confirmed as the final results. The entire decision-making process and its basis are completely recorded on the blockchain, forming an immutable audit trail. The system has established a dynamic adjustment mechanism. The paths for successfully attracting talents will be incorporated into the historical records and used to optimize the calculation of future association strengths, forming a closed-loop system for continuous improvement.

[0093] In the following example, each cross-network node is processed according to the method provided in this optional embodiment. When planning to introduce the team of Wang XX in the field of life and health, the system reveals potential talent attraction paths by analyzing the association network of the existing talent Zhang XX. Zhang XX has a classmate relationship with Wu XX, and Wu XX has a cooperative relationship with Wang XX.

[0094] The first-level path analysis shows a direct cooperative relationship: Wang XX and Wu XX have co-authored 8 papers in the past 5 years. These papers have been cited a total of 623 times, and their cooperation intensity score has reached 92 points (out of 100).

[0095] MATCH (w:Wang XX)-[r:COAUTHOR]-(k:Wu XX) WHERE r.strength>85 RETURN r.papers, r.citations The MATCH keyword is used to specify the matching pattern and attempt to find matching nodes and relationships in the network. (w:Wang XX) represents matching the node with the label "Wang XX". (k:Wu XX) represents matching the node with the label "Wu XX", which is automatically filled in after the search results are obtained and is the result of the program execution. During the search process, k can represent multiple candidate nodes. [r:COAUTHOR] represents the relationship with the type "COAUTHOR", connecting the two nodes, indicating that these two candidate objects are co-author relationships. The WHERE keyword is used to filter the matching results and only select the records that meet specific conditions. r.strength>85 is a conditional expression, indicating that only the records with a cooperation intensity "strength" greater than 85 are selected. The RETURN keyword is used to specify the fields to be returned in the query results. r.papers and r.citations are the attributes of the relationship r, representing the number of co-authored papers and the number of citations respectively. Return the number of papers and the number of citations corresponding to all cooperative relationships that meet the conditions. Running the above processing shows that Wang XX and Wu XX have co-authored 8 papers in the past 5 years (total citations 623 times), and the cooperation intensity score is 92 / 100.

[0096] The secondary path analysis expands the alumni network: Zhang and Wu are not only alumni of the university but also pursued their doctoral degrees in the same cohort from 2008 to 2012. In addition, the two jointly applied for 2 patents in the field of life and health.

[0097] MATCH (z:Zhang)-[edu:ALUMNI]-(w:Wu) WHERE edu.school="xx University" AND edu.duration OVERLAPS 2008-2012 WITH z,w MATCH (w)-[proj:SHARED_PROJECT]-(Wang) WHERE proj.field = "Life and Health" RETURN z,w,Wang (z:Zhang) represents matching the node with the label "Zhang", which is the result obtained after the query is executed and is automatically filled in. During the search process, z can represent multiple candidate nodes. [edu:ALUMNI] represents matching the educational relationship of type "ALUMNI", indicating that the two candidate objects are alumni. "xx University" AND edu.duration OVERLAPS 2008-2012 is a conditional expression that filters out the alumni relationships where the school is "xx University" and the study time overlaps between 2008 and 2012.

[0098] WITH z,w means passing the nodes z and w matched according to the alumni conditions to the next part of the query. MATCH (w)-[proj:SHARED_PROJECT]-(Wang) means finding the node connected to node w (which is not yet determined to be Wu during the search process) through the relationship proj of type SHARED_PROJECT (shared project). The SHARED_PROJECT relationship indicates that the two candidate objects jointly participated in a certain project. WHERE proj.field = "Life and Health" is a conditional expression that filters out the shared project relationships where the project field is "Life and Health".

[0099] Based on this information, the system discovered a key talent recruitment path: Through the classmate relationship between Zhang and Wu, and the cooperation relationship between Wu and Wang, Zhang can serve as a bridge connecting the three, and a target path can be obtained.

[0100] During this process, Zhang can utilize his resources in the university to actively contact Wu and express his intention to recruit the team of Wang, and explore possible cooperation opportunities. At the same time, Wu can act as an intermediary to facilitate the communication and cooperation between Zhang and Wang, thereby reducing communication costs and trust barriers, and creating favorable conditions for the successful recruitment of Wang's team.

[0101] According to the above optional implementation manner, through the distributed characteristics of the blockchain, the secure sharing of cross-institutional data is realized, and the data island problem in related technologies is solved. The consensus mechanism of multiple cross-network nodes ensures the objectivity and fairness of the results and prevents the risk of single-point manipulation. The whole-process blockchain evidence preservation meets the compliance requirements and provides reliable data support for the talent recruitment decision. Compared with the centralized solution provided by related technologies, this optional implementation manner has significantly improved in terms of data security, path quality, and system reliability, providing a new technical paradigm for intelligent talent recruitment.

[0102] Corresponding to the application scenario and method of the method provided in the embodiment of the present application, Figure 3 The schematic block diagram of the path planning device according to the embodiment of the present application is shown. As Figure 3 shown, the embodiment of the present application further provides a path planning device, including: A sub-network generation module 301, configured to establish multiple sub-networks, where the multiple sub-networks are respectively generated based on the information of candidate objects provided by corresponding cooperation institutions, and the multiple sub-networks are respectively stored in different network storage nodes in the blockchain; A cross-network acquisition module 302, configured to, when there is no single-network path from a known object to the current target object in the multiple sub-networks, respectively acquire the multiple sub-networks by each cross-network node set in the blockchain, where the known object is a candidate object with a contact way among the multiple candidate objects, and the current target object is an object that needs to establish contact among the multiple candidate objects; A cross-network search module 303, configured to use each cross-network node to search according to the multiple sub-networks to determine a target path from the known object to the current target object; A consensus voting module 304, configured to determine a final path by adopting a majority consensus manner based on the target paths respectively obtained by the multiple cross-network nodes.

[0103] The functions of the modules in each device of the embodiment of the present application can refer to the corresponding descriptions in the above method and have corresponding beneficial effects, which will not be elaborated here.

[0104] Figure 4 It is a block diagram of an electronic device for implementing the embodiment of the present application. As Figure 4As shown, the electronic device includes: a memory 401 and a processor 402. The memory 401 stores a computer program that can run on the processor 402. When the processor 402 executes the computer program, the methods in the above embodiments are implemented. The number of the memory 401 and the processor 402 can be one or more. In a specific implementation, the electronic device may further include a communication interface 403 for communicating with external devices and performing data interaction and transmission.

[0105] In a specific implementation, if the memory 401, the processor 402, and the communication interface 403 are independently implemented, the memory 401, the processor 402, and the communication interface 403 can be interconnected through a bus and complete communication with each other. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of representation, Figure 4 only a thick line is used to represent it in the figure, but it does not mean that there is only one bus or one type of bus.

[0106] Optionally, in a specific implementation, if the memory 401, the processor 402, and the communication interface 403 are integrated on a chip, the memory 401, the processor 402, and the communication interface 403 can complete communication with each other through an internal interface.

[0107] An embodiment of the present application provides a computer-readable storage medium that stores a computer program. When the program is executed by a processor, the path planning method provided in the embodiment of the present application is implemented.

[0108] An embodiment of the present application provides a computer program product, including a computer program. When the program is executed by a processor, the path planning method provided in the embodiment of the present application is implemented.

[0109] An embodiment of the present application further provides a chip, which includes a processor for calling and running an instruction stored in a memory from the memory, so that a communication device installed with the chip executes the path planning method provided in the embodiment of the present application.

[0110] An embodiment of the present application further provides a chip, including: an input interface, an output interface, a processor, and a memory. The input interface, the output interface, the processor, and the memory are connected through an internal connection path. The processor is configured to execute the code in the memory. When the code is executed, the processor is configured to execute the path planning method provided by the embodiment of the application.

[0111] It should be understood that the above-mentioned processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. It is worth noting that the processor may be a processor that supports the advanced reduced instruction set machine (ARM) architecture.

[0112] Further, optionally, the above-mentioned memory may include a read-only memory and a random access memory. The memory may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memories. Among them, the non-volatile memory may include a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory may include a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available. For example, static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct rambus random access memory (DR RAM).

[0113] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the processes or functions according to the present application are generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium.

[0114] In the description of this specification, the descriptions referring to terms such as "one embodiment", "some embodiments", "examples", "specific examples", or "some examples", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of this application. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0115] In addition, the terms "first" and "second" are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, the features defined with "first" and "second" may explicitly or implicitly include at least one of these features. In the description of this application, "a plurality of" means two or more unless otherwise specifically defined.

[0116] Any process or method described in the flowchart or otherwise described herein can be understood to represent a module, segment, or portion of code including one or more executable instructions for implementing a specific logical function or process. And the scope of the preferred embodiments of this application includes additional implementations, where the functions can be executed in a substantially simultaneous manner or in a reverse order according to the involved functions, rather than in the order shown or discussed.

[0117] The logic and / or steps described in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing a logical function, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or in conjunction with these instruction execution systems, apparatuses, or devices.

[0118] It should be understood that each part of this application can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. All or part of the steps of the method in the above embodiments can be completed by a program instructing relevant hardware. The program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0119] In addition, each functional unit in various embodiments of the present application may be integrated into a processing module, may exist separately physically for each unit, or two or more units may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of a software functional module. If the above-mentioned integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium. The storage medium may be a read-only memory, a magnetic disk, an optical disc, or the like.

[0120] The above is only an exemplary embodiment of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope recorded in the present application can easily think of various changes or substitutions thereof, and these should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A path planning method, characterized in that, Including: Establishing a plurality of sub-networks, which are respectively generated based on information of candidate objects provided by corresponding cooperation institutions, and the plurality of sub-networks are respectively stored in different network storage nodes in the blockchain; In the case that there is no single-network path from a known object to the current target object in the plurality of sub-networks, each cross-network node set in the blockchain respectively obtains the plurality of sub-networks; The known object is a candidate object with a contact path among the plurality of candidate objects, and the current target object is an object among the plurality of candidate objects that needs to establish a connection; Using each cross-network node, searching according to the plurality of sub-networks to determine a target path from the known object to the current target object; Based on the target paths respectively obtained by the plurality of cross-network nodes, determining the final path by means of a majority consensus method.

2. The method according to claim 1, wherein Each of the plurality of sub-networks corresponds to an edge server, and the permission of the edge server is set to allow access to the network storage node of the corresponding sub-network in the blockchain. After establishing the plurality of sub-networks, the method further includes: Using a plurality of edge servers to respectively search in the corresponding sub-networks to determine whether there is the single-network path; In the case that there is a unique single-network path, determining the single-network path as the final path; In the case that there are a plurality of single-network paths, determining path scores respectively corresponding to the plurality of single-network paths; Determining the one with the highest path score among the plurality of single-network paths as the final path.

3. The method according to claim 1, characterized in that, There are a plurality of known objects, and the identity information of the plurality of known objects is stored in the inventory talent node in the blockchain in a predetermined order. The method further includes: Determining the priorities of the plurality of known objects based on the email domains, default records, and response delays respectively corresponding to the plurality of known objects. The priorities are used to indicate the path search order based on the plurality of known objects, and the default record is a predetermined discreditable event recorded by a smart contract in the blockchain; Updating the predetermined order stored in the inventory talent node according to the priorities of the plurality of known objects.

4. The method according to any one of claims 1 to 3, characterized in that Using each cross-network node, searching according to the plurality of sub-networks to determine a target path from the known object to the current target object, including: Fusing the plurality of sub-networks to obtain a comprehensive network; the nodes included in the comprehensive network are the plurality of candidate objects, and the edges included in the comprehensive network represent different types of association relationships of the plurality of candidate objects; Using a path search strategy to search the comprehensive network and recording the paths passed during the search as the plurality of candidate paths, and the path search strategy is determined based on the network scale of the comprehensive network; Using each of the cross-network nodes, determine the path scores corresponding to the multiple candidate paths according to the association strength between the multiple candidate objects and a predetermined attenuation factor; the attenuation factor is used to represent the change manner in which the path score of the corresponding candidate path decreases as the path length increases, the path length represents the number of candidate objects passed through, and the association strength is adjusted based on historical path records, and the historical path records include the historical paths used to successfully contact historical target objects; Based on the path scores corresponding to the multiple candidate paths, perform path pruning on the multiple candidate paths to determine the target paths obtained by each cross-network node.

5. The method according to claim 4, characterized in that The historical path records are stored in historical nodes set in the blockchain, and the method further includes: When the talent recruitment result for the current target object is successful according to the final path, verify the credibility of the talent recruitment result; When the talent recruitment result passes the credibility verification, store the final path in the historical node and update the historical path records.

6. The method according to claim 4, characterized in that The adopting a path search strategy to search the integrated network includes: When the network scale is smaller than a predetermined threshold, determine the path search strategy as: starting from the starting node representing the known object in the integrated network, search whether there is a target node representing the target object within a predetermined range; When the target node does not exist within the predetermined range, search based on the adjacent nodes of the nodes within the predetermined range to determine whether the target node exists among the adjacent nodes; When the target node does not exist among the adjacent nodes, start with the adjacent nodes as new starting nodes and search the nodes within the predetermined range; Repeat the above process until the target node exists within the predetermined range or among the adjacent nodes, and stop executing the search process.

7. The method according to claim 4, wherein The adopting a path search strategy to search the integrated network includes: When the network scale is greater than or equal to a predetermined threshold, determine the path search strategy as: start searching in the integrated network from the starting node representing the known object; Based on the actual path cost from the starting node to the current node and the remaining path cost, determine the next node that minimizes the cumulative path cost from the starting node to the target node; the target node is used to represent the target object in the integrated network, and the remaining path cost is used to estimate the cost required to reach the target node from the current node; Process in the manner of determining the next node until the new next node is the target node, and stop executing the search process.

8. An electronic device, including a memory, a processor, and a computer program stored on the memory, and the processor implements the method according to any one of claims 1 to 7 when executing the computer program.

9. A computer-readable storage medium storing a computer program therein, and the computer program, when executed by a processor, implements the method according to any one of claims 1 to 7.

10. A computer program product comprising a computer program which, when executed by a processor, implements the method according to any one of claims 1 to 7.

Images