A task allocation method, system, and storage medium based on an intelligent task searcher

Abstract

The present invention relates to a task allocation method, system and storage medium based on an intelligent task searcher, belonging to the technical field of intelligent task allocation, including: initializing tasks and the intelligent task searcher: establishing an employment task linked list and an intelligent task searcher matrix, digitizing tasks and personnel; the intelligent task searcher performing a displacement iteration algorithm; performing task matching on the intelligent task searcher; loading new personnel and new tasks into the intelligent task searcher; recommending the matching results and feeding them back to the employment demand unit. By establishing an employment task linked list and an intelligent task searcher matrix, digitizing tasks and personnel, and controlling the search method of the intelligent task searcher through the displacement iteration algorithm of the intelligent task searcher, the present invention realizes the matching processing of a large number of clustered tasks, and improves the efficiency of personnel task matching.

Description

Technical Field

[0001] The present invention relates to a task allocation method, system and storage medium based on an intelligent task searcher, and belongs to the technical field of intelligent task allocation. Background Art

[0002] Under the wave of "mass entrepreneurship and innovation", the huge potential contained in the wisdom of the public has been explored. The crowdsourcing model represented by flexible employment and human resource sharing centers advocates breaking through the traditional organizational boundaries and outsourcing tasks to the general public who are not specific. In recent years, it has received close attention and wide promotion. How to adopt an effective systematic method to reasonably match employees with work tasks is an important problem faced by the crowdsourcing model represented by flexible employment and human resource sharing centers.

[0003] The current research methods mainly focus on achieving one-to-one matching of single threads and cannot solve the problem of big data matching of employees and tasks in the system. Summary of the Invention

[0004] In view of the deficiencies of the prior art, the present invention provides a task allocation method, system and storage medium based on an intelligent task searcher. By establishing an employment task linked list and an intelligent task searcher matrix, tasks and personnel are digitized. Through the displacement iteration algorithm of the intelligent task searcher, the search method of the intelligent task searcher is controlled, so as to realize the matching processing of large-scale clustered tasks and improve the matching efficiency of personnel tasks.

[0005] The present invention adopts the following technical solutions:

[0006] A task allocation method based on an intelligent task searcher includes the following steps:

[0007] (1) Initialize the task and the intelligent task searcher: establish an employment task linked list and an intelligent task searcher matrix, and digitize the tasks and personnel;

[0008] (2) The intelligent task searcher performs a displacement iteration algorithm;

[0009] (3) Perform task matching on the intelligent task searcher;

[0010] (4) Load new personnel and new tasks into the intelligent task searcher;

[0011] (5) Recommend the matching results. The system regularly feeds back the matching result form FK to the employment demand unit as a personnel task recommendation form, which can be displayed on the front-end interface of the system.

[0012] Preferably, the step (1) is specifically:

[0013] Establish a task coordinate system X (this coordinate system is a virtual one-dimensional coordinate system with only the X-axis, and the X-axis represents the position where the task is randomly assigned). Initialize the range of the abscissa X of the task coordinate system to be between 0 and a, where a is an integer within 100, and the default initialization is 100. Establish a labor task linked list Y b (x b ,c b ,d b ), where b is the labor task number, x b is the abscissa value of the labor task in the task coordinate system, c b is the text information description of the task, d b is the status of the task. Assume the labor task number b = 1008, x b = 70, c b = "Position: Welding technology; Education: Above high school; Working hours: 4 hours; Age requirement: None; Household register requirement: None; Other requirements:...", d b = "Pending assignment of personnel". Initialize all task requirements of the labor system to the labor task linked list Y b (x b ,c b ,d b ) (data initialization can be performed through basic operations such as addition, deletion, modification, and query). At the same time, use a random number generation algorithm to randomly generate an abscissa x b , 0 ≤ x b ≤ a;

[0014] Randomly generate n intelligent task searchers. The intelligent task searcher is similar to an intelligent task execution unit, which is virtual and can be regarded as an independent edge computing body or a computer host. Each intelligent task searcher is allocated computing resources such as CPU, GPU, memory, and hard disk. Each intelligent task searcher undertakes a task. The intelligent task searcher shares an abscissa with the initialized task. Allocate computing resources for each intelligent task searcher and establish an intelligent task searcher matrix P m (X p ,C p ,d p ,e p ), where m is the intelligent task searcher number, X p is the position of the intelligent task searcher, C p is the text description of the personnel information loaded by the intelligent task searcher, d p is the status of the intelligent task searcher, e p is the historical average matching rate of the intelligent task searcher. Each intelligent task searcher can perform functions such as displacement iteration calculation, task matching calculation, and historical average matching rate calculation;

[0015] Load all the personnel with tasks to be assigned in the employment system into the intelligent task searcher matrix P m (X p ,C p ,d p ,e p ). (It can be loaded one by one through the polling method), and the intelligent task searcher shares an abscissa with the initialized task.

[0016] Preferably, in the displacement iteration algorithm in step (2), the upper limit of the number of iterations is set to k. At the jth iteration, the position of the ith intelligent task searcher is denoted as X j (i), the position of the task with the smallest distance from the ith intelligent task searcher is denoted as X′ j (i), and the position X of the intelligent task searcher at the (i + 1)th iteration j (i + 1) is calculated as follows:

[0017] X j (i + 1) = αX j (i) - βS j |T j *X j (i) - X′ j (i)|

[0018] Among them, |T j *X j (i) - X′ j (i)| represents the coordinate difference between the ith intelligent task searcher and the task with the smallest distance from the ith intelligent task searcher as the displacement step size;

[0019] α is the first weight of the displacement iteration algorithm. When the personnel information description text C p and the task text information description c b have a similarity greater than 60%, the current position of the intelligent task searcher is in a better position, and α takes a random value in the following interval: [1.1, 1.3]; when the personnel information description text C p and the task text information description c b have a similarity less than 60%, the current position of the intelligent task searcher is in a worse position, and at this time it is suitable to increase the displacement, so the value of α is suitable for a larger space, and α takes a random value in the following interval: [1.6, 1.9];

[0020] β is the second weight of the displacement iteration algorithm. In the process of algorithm iteration, a reasonable displacement amount is the key to the algorithm to seek the optimal solution. At the beginning of the iterative calculation, a larger displacement amount helps to seek the optimal solution. As the number of iterations increases and the iterative calculation enters the later stage, a smaller displacement amount helps to seek the optimal solution. Therefore, the value of β is:

[0021]

[0022] Where g and L are constants, both defaulting to 1, m is the current iteration number, and k is the upper limit of the iteration number;

[0023] S j is the first control factor, and T j is the second control factor, and the calculation methods are as follows:

[0024] S j = 2λ * u - λ

[0025] T j = 2u

[0026] Where u is a random number with a value range of [0, 1], and λ is the convergence coefficient, m and k have the same meanings as in the above formula for calculating β, that is, m is the current iteration number, and k is the upper limit of the iteration number.

[0027] The present invention uses two weight coefficients as a distinction to establish two intelligently switchable search modes, intelligently judge the reasonable displacement amount, and use it as a method to seek the optimal solution. At the beginning of the iterative calculation, a larger displacement amount helps to seek the optimal solution. As the number of iterations increases and the iterative calculation enters the later stage, a smaller displacement amount helps to seek the optimal solution.

[0028] Preferably, step (3) is further:

[0029] After each position movement of the intelligent task searcher, it is matched with the nearest task, and the intelligent task searcher matrix P m (X p , C p , d p , e p )'s description text C of the loaded personnel information p is compared with the task text information description c of the labor task linked list Y b (x b , c b , d b ) (the comparison can be performed using the word segmenter integrated in java and the Jaccard similarity algorithm, which is an existing technology and will not be elaborated here). When the similarity is greater than 80%, the description text C of the personnel information b and the task text information description c p are added to the matching result FK (the matching result FK can be regarded as a table in the database), and the status d of the intelligent task searcher in the intelligent task searcher matrix P b (X m , C p , d p , e p ) p is set to pMarked as the status of personnel to be assigned, and the employment task linked list Y b (x b ,c b ,d b ) The task status d b Marked as assigned personnel, when the similarity is less than 80%, the intelligent task searcher continues to participate in the displacement iteration calculation.

[0030] Preferably, in step (4), when there are new personnel to be worked, automatically traverse the intelligent task searcher matrix P m (X p ,C p ,d p ,e p ) and select d p As the intelligent task searcher with the status of personnel to be assigned, perform personnel loading;

[0031] When there is a new task, add the task to the employment task linked list Y b (x b ,c b ,d b ) where x b The initial value selects the intelligent task searcher matrix P in the current system m (X p ,C p ,d p ,e p ) The historical average matching rate e of the intelligent task searcher p (The historical average matching rate is to count the matching rate of each intelligent task searcher for each matching work and record it, and finally extract an average value calculated in real time) The X value of the intelligent task searcher corresponding to the maximum value p value, that is, assign this X p value to the corresponding x of the task linked list b value to solve the cold start problem.

[0032] A task assignment system based on an intelligent task searcher, including an initialization unit, a displacement iteration algorithm unit, a task matching unit, a loading unit, and a matching result recommendation unit. The initialization unit is used to initialize tasks and intelligent task searchers. The displacement iteration algorithm unit is used to perform displacement iteration algorithm calculations on intelligent task searchers. The task matching unit is used to match tasks for intelligent task searchers. The loading unit is used to load new personnel and new tasks into intelligent task searchers. The matching result recommendation unit is used to recommend the matching results and feedback them to the employment demand unit.

[0033] A computer-readable storage medium stores a computer program thereon, and the computer program is used to execute the above method.

[0034] For those aspects of the present invention that are not elaborated in detail, reference can be made to the prior art.

[0035] The beneficial effects of the present invention are as follows:

[0036] 1) By establishing an employment task linked list and an intelligent task searcher matrix, the present invention realizes the digitization and dataization of tasks and personnel.

[0037] 2) The present invention performs a displacement iteration algorithm through an intelligent task searcher, and intelligently adjusts two search modes of the displacement algorithm through the first weight and the second weight of the displacement iteration algorithm. When the similarity between the personnel information description text C p and the task text information description c b is greater than 60%, the current position of the searcher is in a relatively optimal position, and small-step adjustment can improve the retrieval efficiency. When the similarity between the personnel information description text C p and the task text information description c b is less than or equal to 60%, the current position of the searcher is in a relatively poor position, and restarting the allocation of a random position can better cooperate with the searcher to search for tasks. Through dynamic adjustment, large-scale clustered task matching processing is realized, and the execution efficiency of personnel task calculation matching is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 is a flowchart of a task allocation method based on an intelligent task searcher;

[0039] Figure 2 is a schematic diagram of the system composition. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0040] To make the technical problems, technical solutions, and advantages to be solved by the present invention clearer, the following will be described in detail with reference to the accompanying drawings and specific embodiments, but not limited thereto. For those aspects of the present invention that are not elaborated in detail, they are all in accordance with the conventional techniques in the art.

[0041] Embodiment 1

[0042] A task allocation method based on an intelligent task searcher includes the following steps:

[0043] (1) Initialize the task and the intelligent task searcher: establish an employment task linked list and an intelligent task searcher matrix, and digitize the task and personnel;

[0044] Establish a task coordinate system X (this coordinate system is a virtual one-dimensional coordinate system with only the X-axis, and the X-axis represents the position where the task is randomly allocated), initialize the range of the abscissa X of the task coordinate system to be between 0 and a, where a is an integer within 100, and the default initialization is 100, and establish an employment task linked list Y b (x b ,c b ,db ), where b is the labor task number, and x b is the abscissa value of the labor task in the task coordinate system, c b is the text information description of the task, d b is the status of the task. Assume that the labor task number b = 1008, and x b = 70, c b = "Position: Welding Technology; Education: Above high school; Working hours: 4 hours; Age requirement: None; Household registration requirement: None; Other requirements:...", d b = "Personnel to be assigned", and initialize all task requirements of the labor system to the labor task linked list Y b (x b , c b , d b ) (data initialization can be performed through basic operations such as addition, deletion, modification, and query). At the same time, use a random number generation algorithm to randomly generate an abscissa x b for each task, where 0 ≤ x b ≤ a;

[0045] Randomly generate n intelligent task searchers. The intelligent task searcher is similar to an intelligent task execution unit, which is virtual and can be regarded as an independent edge computing body or a computer host. Each intelligent task searcher is allocated computing resources such as CPU, GPU, memory, and hard disk. Each intelligent task searcher carries a task, and the intelligent task searcher shares an abscissa with the initialized task. Allocate computing resources for each intelligent task searcher and establish an intelligent task searcher matrix P m (X p , C p , d p , e p ), where m is the intelligent task searcher number, X p is the position of the intelligent task searcher, C p is the text description of the personnel information loaded by the intelligent task searcher, d p is the status of the intelligent task searcher, and e p is the historical average matching rate of the intelligent task searcher. Each intelligent task searcher can perform functions such as displacement iteration calculation, task matching calculation, and historical average matching rate calculation;

[0046] Load all personnel to be assigned tasks in the labor system into the intelligent task searcher matrix P m (X p , C p , d p , e p ) (loading can be done one-to-one through polling). The intelligent task searcher shares an abscissa with the initialized task.

[0047] (2) The intelligent task searcher performs a displacement iteration algorithm;

[0048] In the displacement iteration algorithm, the upper limit of the number of iterations is set to k. At the jth iteration, the position of the i-th intelligent task searcher is recorded as X j (i), the task position with the shortest distance to the i-th intelligent task searcher is denoted as X′ j (i), the position X of the intelligent task searcher at the i+1th iteration j (i+1) is calculated as follows:

[0049] X j (i+1)=αX j (i)-βS j |T j *X j (i) -X′ j (i)|

[0050] Among them, |T j *X j (i) -X′ j (i)| denotes that the coordinate difference between the i-th intelligent task searcher and the task with the smallest distance from the i-th intelligent task searcher is used as the displacement step;

[0051] α is the first weight of the displacement iteration algorithm. p With the task text information description c b When the similarity is greater than 60%, the current position of the intelligent task searcher is in a better position, and α takes a random value in the following interval: [1.1, 1.3]; when the personnel information description text C p With the task text information description c b When the similarity is less than 60%, the current position of the intelligent task searcher is in a bad position. At this time, it is suitable to increase the displacement. Therefore, the α value is suitable for a larger space. α takes a random value in the following interval: [1.6, 1.9];

[0052] β is the second weight of the displacement iteration algorithm. In the algorithm iteration process, a reasonable displacement is the key to the algorithm to find the optimal solution. At the beginning of the iterative calculation, a larger displacement helps to find the optimal solution. As the number of iterations increases and the iterative calculation enters the later stage, a smaller displacement helps to find the optimal solution. Therefore, the value of β is:

[0053]

[0054] Where g and L are constants, both of which are 1 by default, m is the current number of iterations, and k is the upper limit of the number of iterations;

[0055] S j is the first control factor, T jis the second control factor, and its calculation method is as follows:

[0056] S j = 2λ * u - λ

[0057] T j = 2u

[0058] where u is a random number with a value range of [0, 1], and λ is the convergence coefficient. m and k have the same meanings as in the above formula for calculating β, that is, m is the current iteration number, and k is the upper limit of the iteration number.

[0059] The present invention uses two weight coefficients as a distinction to establish two intelligently switchable search modes, intelligently judge a reasonable displacement amount, and use it as a method to seek the optimal solution. At the beginning of the iterative calculation, a larger displacement amount helps to seek the optimal solution. As the number of iterations increases and the iterative calculation enters the later stage, a smaller displacement amount helps to seek the optimal solution.

[0060] (3) Perform task matching on the intelligent task searcher;

[0061] After each position movement of the intelligent task searcher, it is matched with the nearest task. The intelligent task searcher matrix P m (X p , C p , d p , e p ) describes the text C of the personnel information loaded p and the task text information description c of the labor task linked list Y b (x b , c b , d b ) are compared in text (the comparison can be performed using the word segmenter integrated in java and the Jaccard similarity algorithm, which is an existing technology and will not be elaborated here). When the similarity is greater than 80%, the personnel information description text C b and the task text information description c p are added to the matching result FK (the matching result FK can be regarded as a table in the database). The status d of the intelligent task searcher in the intelligent task searcher matrix P b m (X p , C p , d p , e p ) is marked as the status of the personnel to be assigned, and the status d of the task in the labor task linked list Y p b b b (x b , c b , d b ) is bMark the allocated personnel. When the similarity is less than 80%, the intelligent task searcher continues to participate in the displacement iteration calculation.

[0062] (4) Load the new personnel and new tasks into the intelligent task searcher.

[0063] When there are new personnel to be worked, automatically traverse the intelligent task searcher matrix P m (X p ,C p ,d p ,e p ), select d p as the intelligent task searcher in the state of personnel to be allocated, and perform personnel loading;

[0064] When there is a new task, add the task to the employment task linked list Y b (x b ,c b ,d b ), where x b The initial value selects the intelligent task searcher matrix P in the current system m (X p ,C p ,d p ,e p ) The historical average matching rate e of the intelligent task searcher p (The historical average matching rate is to count and record the matching rate of each intelligent task searcher for each matching work, and finally extract a real-time calculated average value) The X value of the intelligent task searcher corresponding to the maximum value p value, that is, assign this X p value to the corresponding x of the task linked list b value to solve the cold start problem.

[0065] (5) Recommend the matching result. The system regularly feeds back the matching result feedback form FK to the employment demand unit as the personnel task recommendation form, which can be displayed on the front-end interface of the system.

[0066] Embodiment 2

[0067] A task allocation system based on an intelligent task searcher, including an initialization unit, a displacement iteration algorithm unit, a task matching unit, a loading unit, and a matching result recommendation unit. The initialization unit is used to initialize the task and the intelligent task searcher. The displacement iteration algorithm unit is used to perform displacement iteration algorithm calculation on the intelligent task searcher. The task matching unit is used to perform task matching on the intelligent task searcher. The loading unit is used to load new personnel and new tasks into the intelligent task searcher. The matching result recommendation unit is used to recommend the matching result and feed it back to the employment demand unit.

[0068] Embodiment 3

[0069] A computer-readable storage medium has a computer program stored thereon, and the computer program is used to execute the method of Embodiment 1.

[0070] The above are the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A task assignment method based on an intelligent task searcher, characterized in that, It includes the following steps: (1) Initialize the task and the intelligent task searcher: establish an employment task linked list and an intelligent task searcher matrix, and digitize the tasks and personnel; (2) The intelligent task searcher performs a displacement iteration algorithm; (3) Perform task matching on the intelligent task searcher; (4) Load new personnel and new tasks into the intelligent task searcher; (5) Recommend the matching results and feedback them to the employment demand unit; The specific content of step (1) is as follows: Establish a task coordinate system X, initialize the range of the abscissa X of the task coordinate system to be between 0 and a, where a is an integer within 100, and establish a labor task linked list Y b (x b ,c b ,d b ), where b is the labor task number, x b is the abscissa value of the labor task in the task coordinate system, c b is the text information description of the task, and d b is the status of the task. Initialize all task requirements of the labor system to the labor task linked list Y b (x b ,c b ,d b ). At the same time, use a random number generation algorithm to randomly generate an abscissa x for each task b , 0 ≤ x b ≤ a; Randomly generate n intelligent task searchers, each intelligent task searcher carries a task, the intelligent task searcher shares an abscissa with the initialized task, allocate computing resources to each intelligent task searcher, and establish an intelligent task searcher matrix P m (X p ,C p ,d p ,e p ), where m is the intelligent task searcher number, X p is the position of the intelligent task searcher, C p is the text description of the personnel information loaded by the intelligent task searcher, d p is the status of the intelligent task searcher, e p is the historical average matching rate of the intelligent task searcher. Each intelligent task searcher can perform displacement iteration calculation, task matching calculation, and historical average matching rate calculation; Load all the personnel with tasks to be assigned in the labor employment system into the intelligent task searcher matrix P m (X p ,C p ,d p ,e p ) In this case, the intelligent task searcher shares an abscissa with the initialized task; In the displacement iteration algorithm in step (2), the upper limit of the number of iterations is set to k. At the j-th iteration, the position of the i-th intelligent task searcher is denoted as X j (i), and the position of the task with the smallest distance from the i-th intelligent task searcher is denoted as X′ j (i). The position X j (i + 1) of the intelligent task searcher at the (i + 1)-th iteration is calculated as follows: X j (i + 1) = αX j (i) - βS j |T j *X j (i) - X′ j (i)| Among them, |T j *X j (i)-X′ j (i)| represents using the coordinate difference between the i-th intelligent task searcher and the task with the smallest distance from the i-th intelligent task searcher as the displacement step size; α is the first weight of the displacement iteration algorithm. When the personnel information description text C p and the task text information description c b have a similarity greater than 60%, the current position of the intelligent task searcher is in a better position, and α takes a random value in the following interval: [1.1, 1.3]; when the personnel information description text C p and the task text information description c b have a similarity less than 60%, the current position of the intelligent task searcher is in a worse position, and α takes a random value in the following interval: [1.6, 1.9]; β is the second weight of the displacement iteration algorithm, and the value of β is: where g and L are constants, both default to 1, m is the current iteration number, and k is the upper limit of the iteration number; S j is the first control factor, and T j is the second control factor. The calculation method is as follows: S j = 2λ * u - λ T j = 2u Among them, u is a random number with a value range of [0, 1], and λ is a convergence coefficient.

2. The task allocation method based on an intelligent task searcher according to claim 1, wherein Step (3) is further as follows: Each time the intelligent task searcher moves, it matches the task closest to it and converts the intelligent task searcher matrix P m (X p ,C p ,d p ,e p ) The loaded personnel information description text C p And the labor task list Y b (x b ,c b ,d b ) task text information description c b Perform text comparison. When the similarity is greater than 80%, the personnel information description text C p , Task text information description c b Add to the matching result FK and convert the intelligent task searcher matrix P m (X p ,C p ,d p ,e p ) in the Intelligent Task Finder state d p Mark as the personnel to be assigned status, and link the labor task list Y b (x b ,c b ,d b ) in task status d b When the similarity is less than 80% and the person is marked as assigned, the intelligent task searcher continues to participate in the displacement iteration calculation.

3. The task allocation method based on an intelligent task searcher according to claim 2, wherein In step (4), when there is a new staff member to be processed, automatically traverse the intelligent task searcher matrix P m (X p ,C p ,d p ,e p ), select d p as the intelligent task searcher for the staff member to be assigned, and perform staff loading; When there is a new task, add the task to the employment task linked list Y b (x b ,c b ,d b ), where x b The initial value selects the intelligent task searcher matrix P in the current system m (X p ,C p ,d p ,e p ) the historical average matching rate e of the intelligent task searcher p The X value of the intelligent task searcher corresponding to the maximum value p value.

4. A task allocation system based on an intelligent task searcher, characterized in that, To implement the method described in any one of claims 1-3, it includes an initialization unit, a displacement iteration algorithm unit, a task matching unit, a loading unit, and a matching result recommendation unit. The initialization unit is used to initialize the task and the intelligent task searcher. The displacement iteration algorithm unit is used to perform displacement iteration algorithm calculations on the intelligent task searcher. The task matching unit is used to perform task matching on the intelligent task searcher. The loading unit is used to load new personnel and new tasks into the intelligent task searcher. The matching result recommendation unit is used to recommend the matching results and feedback them to the employment demand unit.

5. A computer-readable storage medium, characterized in that, A computer program is stored on a computer-readable storage medium, and the computer program is used to execute the method described in any one of claims 1-3.

Images