A distributed computing scheduling method for resource competition

Abstract

The present invention discloses a distributed computing scheduling method for resource competition. The distributed computing scheduling includes: each computing terminal constituting the distributed computing scheduling is composed of electronic devices with computing and data processing capabilities, including but not limited to servers, personal computers, edge computers, and smart devices; each computing terminal sends the scheduling plan of the subtask obtained by its scheduled operation to the next computing terminal for scheduling operation through a preset encryption communication method until the scheduling operation is completed; the scheduling method includes: initializing the scheduling system and modeling related predictions, inputting the task plan and entering the scheduling operation; the present invention makes up for the technical deficiencies of the prior art, such as the inability to reach a consensus on the scheduling algorithm between supply chain partners, the inability to make a scheduling response in real time according to the production site situation, and the inability to predict the value change of resources in real time, and provides a distributed computing scheduling method for resource competition.

Description

Technical Field

[0001] The present application relates to the field of informatization and digitization, and specifically to a distributed computing scheduling method for resource competition. Background Art

[0002] APS is the abbreviation of Advanced Planning and Scheduling. The detailed scheduling of APS is far more than that of general ERP system. It is very important for some large-scale discrete enterprises. There are many APS scheduling methods, such as first-in-first-out, minimum processing time, minimum completion time first, maximum benefit method, critical path method, etc.

[0003] At present, cross-enterprise planning collaboration has not been widely used. Basically, each enterprise still arranges its own production plan and project plan, and the plans are independent of each other. A large part of the reason for this is that the various enterprises in the supply chain lack a scheduling calculation method that is easy for each other to agree on. How to use advanced technology and scheduling methods (algorithms) that can be recognized by various enterprises in the supply chain is undoubtedly a technical problem at present.

[0004] With the continuous development of artificial intelligence algorithms, neural networks, genetic algorithms, etc. have also been applied to scheduling, but the requirements for computing power are getting higher and higher. However, there is no effective solution in scheduling for how to apply artificial intelligence technology and at which stage of scheduling. In addition, the requirements for computing power of artificial intelligence technology are getting higher and higher, so there is no effective solution for which part of the scheduling plan needs artificial intelligence technology the most.

[0005] In addition, the current scheduling system will basically not actively modify the scheduling plan if the order plan has not changed. However, the market situation is changing rapidly, and the production site conditions of enterprises are also subject to variables at all times. When should the plan be triggered to modify and on what basis should it be modified actively? This is also a technical problem.

[0006] The above problems hinder the development of scheduling technology and affect the efficient collaboration of the supply chain. They also create considerable technical resistance to the development of Industry 4.0 and Industrial Internet technologies, and hinder the development of our intelligent manufacturing level. Summary of the invention

[0007] The purpose of the present invention is to make up for the technical deficiencies of the above-mentioned prior art, such as the inability to obtain the optimal combination of conditions among numerous scheduling constraints, the inability to reach a consensus on the scheduling algorithm among supply chain partners and the inability to form a unified scheduling operation network, the inability to effectively solve the scheduling computing power of the server, the inability to make scheduling responses in real time according to the production site conditions, the inability to integrate artificial intelligence technology well, and the inability to predict the value changes of various resources and task routes in real time, thereby providing a distributed computing scheduling method for resource competition.

[0008] The purpose of the present invention is achieved through the following technical solutions:

[0009] Each computing terminal that constitutes the distributed computing scheduling network is composed of electronic devices with computing and data processing capabilities, including but not limited to servers, personal computers, edge computers, and smart devices; a subtask list of all task routes that each computing terminal is responsible for scheduling operations is preset; each computing terminal sends the scheduling plan of the subtask obtained by its scheduled operation to the next computing terminal for scheduling operations through a preset encrypted communication method; further, the computing terminal that receives the information continues the scheduling operation until the scheduling operation is completed; the computing terminals include not only the computing terminals within the subject, but also the computing terminals of the cooperative subjects; when a subtask is undertaken by other subjects, the computing terminal responsible for the calculation of this subtask will join the computing network of the task subject together to form a distributed computing scheduling network.

[0010] The tasks mentioned above are the work to achieve a certain goal, including but not limited to product production and project tasks; the task plan of this application refers to the work plan for the tasks that need to be delivered in the end;

[0011] The task route is a collection of all the steps required to complete a task. A task can have multiple completion plans, that is, there can be multiple task routes. The subtasks are the steps that can be accepted in the task. Each task has only one end subtask, but multiple start subtasks are allowed. When a subtask is completed, it is the turn of the next subtask to start work. Then the previous subtask is the immediate previous subtask of the next subtask, and the next subtask is the immediate next subtask of the previous subtask.

[0012] The scheduling plan mentioned above refers to the work plan for each step, i.e. each subtask, in the production or service process;

[0013] This application outputs the scheduling parameters of each subtask, including but not limited to scheduled tasks, subtasks, work centers, scheduled quantity or progress, subtask quantity or progress, subtask duration, earliest start time, latest start time, earliest finish time, and latest finish time;

[0014] The scheduling tasks are tasks that participate in the scheduling calculation and are the tasks that need to be delivered in the end;

[0015] The work center is an operating unit that is capable of performing certain types of work. A work center can be one or more direct production personnel, one or several machines with the same function, or a combination of people and equipment.

[0016] The scheduled quantity or progress mentioned above refers to the workload of the scheduled tasks that have been scheduled; if all tasks are scheduled, the scheduled quantity or progress is equal to the total quantity or progress of the scheduled tasks;

[0017] The number or progress of subtasks mentioned refers to the workload of the subtask; due to different measurement units and work content, the number of subtasks and scheduled tasks may not be equal.

[0018] A resource competition distributed computing scheduling method, the specific steps include:

[0019] Step 101: Initialize the scheduling system and model related predictions;

[0020] Specifically, it includes:

[0021] (1) Preset work centers and work calendars of work centers;

[0022] The work completed by a work center in a task route is the subtask of the task route. There is a one-to-many relationship between subtasks and work centers, that is, each specific subtask corresponds to a unique work center; but a work center can correspond to multiple subtasks, that is, a work center can complete different specific tasks.

[0023] The work calendar of the work center is a list that defines the working hours and rest time of the work center; the tasks performed by the work center can only be scheduled to be performed within the working hours of the work center.

[0024] (2) Preset the task route parameters for all tasks;

[0025] All task route parameters described, including but not limited to task route name, subtask, subtask duration, work center, computing terminal, and next subtask;

[0026] The computing terminal is responsible for scheduling the subtask, that is, the computing terminal responsible for arranging tasks for the work center;

[0027] Each specific subtask corresponds to a specific work center and a unique computing terminal.

[0028] (3) Preset the communication method and encryption method between each computing terminal and other computing terminals.

[0029] (4) Preset the calculation logic and / or algorithm rules for the earliest start time of the starting subtask of each task. According to this calculation logic and / or algorithm rules, the earliest start time of the starting subtask of each task can be calculated.

[0030] (5) Model the revenue values ​​of all subtasks of the predicted task routes and obtain the revenue value model of the task routes;

[0031] The benefit value of the subtask mentioned above refers to the measurable benefit or importance brought to the subject by a subtask of a task route completing a unit task quantity or a unit task progress within a unit time; since the subtask is completed by a specific work center, it is equivalent to the work center completing a unit task quantity or a unit task progress within a unit time, bringing a measurable benefit or importance to the subject;

[0032] The model for predicting the revenue value of subtasks of all task routes is a model established by using artificial intelligence technology to predict the revenue value of subtasks of all task routes;

[0033] This application does not involve specific algorithms, but is intended to provide methodological support for how this application uses artificial intelligence technology;

[0034] The steps include:

[0035] ① The key factors of the benefit value of the preset subtask, that is, the factors that have a key impact on the benefit value of the subtask and can be measured;

[0036] ② Obtain the historical record set big data of the benefit value of each subtask and the historical record set big data of the key factor value of the benefit value of each subtask from the management department of the relevant subject;

[0037] ③ Take the key factors of the subtask's benefit value as the independent variable and the subtask's benefit value as the dependent variable; use artificial intelligence technology to find the relationship model between the dependent variable and the independent variable; use this model to input the value of the key factors of the subtask's benefit value and output the subtask's benefit value;

[0038] The revenue value of the task route is the sum of the revenue values ​​of all subtasks of the task route. The calculation formula of the revenue value of the task route is:

[0039] In the formula:

[0040] ①n is the number of subtasks in the task route;

[0041] ②i is the i-th subtask of the task route;

[0042] ③V i is the revenue value of the ith subtask of the task route, which is calculated according to the revenue value model of the subtasks of the task route;

[0043] ④R is the revenue value of the task route;

[0044] Selecting the task route with the highest benefit value to participate in the scheduling calculation will achieve resource efficiency competition and survival of the fittest in work centers.

[0045] (6) Preset a model for calculating task importance scores;

[0046] The task importance score is the result of evaluating the importance of each planned task; the ranking of the task importance score is the ranking of the importance of each planned task; the task importance score is the basis for the order of scheduling calculation. When the work center is overloaded during a certain period of time, task plans with high task importance scores are given priority.

[0047] (7) Calculation method for pre-setting work center capacity verification;

[0048] The work center capacity refers to the ability of the work center to complete tasks; the work center load refers to the workload assigned to the work center; the work center capacity and load must be measured in a unified unit; the work center capacity verification is to determine the relationship between the work center capacity and load and output the result; when the capacity is greater than or equal to the load, it means that the work center capacity meets the demand, otherwise it means that the work center capacity does not meet the demand; the capacity not meeting the demand is also called overload;

[0049] The technical requirements for work center capability verification are:

[0050] ① The capacity of the work center must give priority to the task plans with high task importance scores, that is, the capacity of the work center must be checked in the order of task importance scores from high to low;

[0051] ② Verify the capacity of the work center for each subtask based on the existing planning and scheduling capacity verification technology, especially the capacity verification technology of the ERP system and / or APS system, and output the number or progress of all task plans that have passed the verification and the number or progress of all task plans that have not passed the verification.

[0052] (8) Preset the pre-processing method for each task rescheduling;

[0053] There are two pre-processing methods for rescheduling each task: one is to modify the task plan, and the other is to reselect the task route; the pre-processing method for rescheduling each task is to pre-set one of the two processing methods for each task plan. When the task plan has "the number or progress that has not passed the verification", the pre-set processing method is installed for processing;

[0054] The output of the pre-processing method for rescheduling each task is preset to be a task route and a task plan;

[0055] The specific processing is as follows:

[0056] ① The modification of the task plan is performed by presetting the calculation logic and / or algorithm rules for processing the task quantity or progress, and presetting the calculation logic and / or algorithm rules for processing the latest completion time of the task, respectively, to generate a new task quantity or progress, and a new latest completion time of the task; further, the new task quantity or progress, and the new task completion time are combined with other parameters of the task plan to output a new task plan, and refresh the task importance score of the new task; further, task route selection is performed; when a task has multiple task routes, the task route with the largest benefit value is selected as the task route for scheduling the task;

[0057] Finally, the following can be output: task route and new task plan;

[0058] ② The reselection of the task route does not change the task plan, but only changes the task route; specifically, from all the task routes available in the task plan, the task routes are sorted in descending order according to their benefit values, and the task route that follows the task route currently participating in the scheduling operation is selected:

[0059] a. If the task route that follows the task exists, then output the task route that follows the task and output the current task plan at the same time;

[0060] b. If the task route that follows does not exist, enter the first processing method of "modifying the task plan" mentioned above, and finally output the task route and the new task plan.

[0061] (9) Presetting the conditions that trigger the rescheduling operation;

[0062] The specific processing is as follows:

[0063] ① Preset the conditions for triggering the scheduling operation of the subtask revenue value of each task route;

[0064] Specifically, the maximum and minimum trigger values ​​of the revenue value of each subtask of each task route are preset, and when the revenue value of the subtask is greater than the maximum value, or less than the minimum value, the rescheduling operation is triggered;

[0065] ② Preset the task importance score of each task plan to trigger the rescheduling calculation conditions;

[0066] Specifically, the maximum and minimum trigger values ​​of the task importance score of the task plan are preset, and when the task importance score of the task plan is greater than the maximum value, or less than the minimum value, the rescheduling operation is triggered.

[0067] Step 102: Inputting a mission plan;

[0068] The task plan includes but is not limited to the following parameters: task plan name, task quantity or progress, and latest completion time.

[0069] Step 103: Calculate the task importance scores of all task plans and sort them from high to low, and then schedule each task plan in turn according to the sorting;

[0070] Specifically, according to the calculation method of the preset task importance scores in step 101, the task importance scores of all task plans are calculated and sorted from high to low, and the scheduling operation is performed in step 104 in sequence according to the sorting.

[0071] Step 104: Enter the task scheduling calculation process;

[0072] The specific steps are as follows:

[0073] ① The task plans involved in the scheduling calculate the revenue values ​​of each task route of all task plans according to the revenue value model of the subtasks of all task routes predicted in step 101;

[0074] ② Further, the task route with the largest benefit value is selected as the task route for scheduling calculation of the task plan;

[0075] ③ Enter the task scheduling calculation process, that is, enter step S201.

[0076] The task scheduling calculation process specifically includes the following steps:

[0077] Step S201: start calculating the subtasks at both ends of the task route at the same time;

[0078] The calculation starts from the subtasks at the beginning and end of the task route at the same time. The calculation starts from the starting subtask of the task route in the forward order and starts from the completed subtask of the task route in the reverse order.

[0079] The sequential scheduling is a scheduling method that starts with the starting subtask of the task route according to the selected task route, and plans and arranges tasks in the direction of task completion according to the task route;

[0080] The reverse scheduling is a scheduling method that starts with the completed subtasks of the task route according to the selected task route, and schedules tasks in the direction of starting work according to the task route.

[0081] Specifically, according to the task plan and task route of the current scheduling operation, the following processing is performed:

[0082] (1) Execute in order:

[0083] ① The "scheduled quantity or progress" of the starting subtask of the task plan is equal to the quantity or progress of the task plan;

[0084] ② Calculate the “earliest start time” of the start subtask of the task plan; specifically, according to the calculation logic and / or algorithm rules of the earliest start time of each task’s start subtask preset in step 101, the earliest start time of the start subtask of the task plan can be calculated according to this calculation logic and / or algorithm rules;

[0085] ③ The "earliest completion time" of the starting subtask of the task plan is equal to the "earliest start time" plus the "subtask duration" of the subtask;

[0086] ④ The "work center, subtask duration, subtask quantity or progress" of the starting subtask of the task plan is calculated based on the calculation technology of the task route and the existing plan scheduling, especially the calculation technology of the ERP system and / or APS system;

[0087] (2) Execute when reversing:

[0088] ① The "scheduled quantity or progress" of the completed subtasks of the task plan is equal to the quantity or progress of the task plan;

[0089] ② The "latest completion time" of the completed subtask of the task plan is equal to the "latest completion time" of the task plan;

[0090] ③ The "latest start time" of the completed subtask in the task plan is equal to the "latest completion time" minus the "subtask duration" of the subtask;

[0091] ④ The "work center, subtask duration, subtask quantity or progress" of the completed subtask of the task plan is calculated based on the calculation technology of the task route and the existing plan scheduling, especially the calculation technology of the ERP system and / or APS system;

[0092] Further, the process proceeds to step S202.

[0093] Step S202: Performing capability check on the work center of the subtask;

[0094] Specifically, according to a preset work center capacity verification calculation method, a capacity verification is performed on the work center of the subtask to verify the number or progress of the output task plan that has passed the verification and the number or progress of the task plan that has not passed the verification.

[0095] Step S203: Is it a task plan that has passed the verification? If yes, proceed to step S205; if no, proceed to step S204;

[0096] Specifically, according to the output result of step 202, the processing of the number or progress of task plans that pass the verification enters step S205; if the number or progress of task plans that fail the verification enters step S204.

[0097] Step S204: output the task route and the new task plan according to the pre-set pre-processing method for rescheduling each task; and respond to the reselected task route; further, re-enter step S201;

[0098] Specifically, the task route and task plan are output according to the pre-processing method of each task rescheduling; if the task route changes, the following processing is further performed:

[0099] ① "Scheduled quantity or progress" is changed to "Quantity or progress that has passed verification";

[0100] ② “Subtask quantity or progress” is calculated based on the task route, the scheduled quantity or progress, and the existing planning and scheduling calculation technology, especially the calculation technology of the ERP system and / or APS system;

[0101] Further, by using the output task route and task plan, re-enter step S201.

[0102] Step S205: Is the plan confirmed? If yes, proceed to step S206; if no, proceed to step S204.

[0103] Step S206: obtaining a scheduling plan for the subtask;

[0104] Specifically, the confirmed forward and reverse scheduling plans of the subtask are obtained; the scheduling plan parameters include but are not limited to: scheduled tasks, subtasks, work centers, scheduled quantity or progress, subtask quantity or progress, subtask duration, earliest start time, latest start time, earliest completion time, and latest completion time.

[0105] Step S207: Is it a task completion node? If yes, proceed to step S209; if not, proceed to step S208.

[0106] Step S208: the computing terminal sends the scheduling plan of the subtask to the next computing terminal for scheduling calculation;

[0107] Specifically, the process is as follows:

[0108] (1) Execute in order:

[0109] ① Send the scheduling information of the currently scheduled task plan obtained in S206 to the computing terminal corresponding to the subtask immediately following the current subtask according to the preset communication method and encryption method;

[0110] ② Further, determine whether the next subtask is an intersection subtask; if so, take the maximum value of the "earliest completion time" of all the previous subtasks as the "earliest start time" of the intersection subtask; otherwise, directly take the "earliest completion time" of the previous subtask as the "earliest start time" of the next subtask;

[0111] The intersection subtask refers to a subtask in which two or more subtasks in the same task route converge into the same subtask after completion. This subtask is the intersection subtask; that is, an intersection subtask has multiple immediate predecessor subtasks;

[0112] ③ Further, the "scheduled quantity or progress" of the previous subtask is used as the "scheduled quantity or progress" of the next subtask;

[0113] ④ Further, the “earliest completion time” of the next subtask is equal to the “earliest start time” of the next subtask plus the “subtask duration” of the subtask;

[0114] ⑤ The "work center, subtask duration, subtask quantity or progress" of the immediately following subtask is calculated based on the task route and existing planning and scheduling technology, especially using the calculation technology of the ERP system and / or APS system;

[0115] ⑥ Further, proceed to step 202;

[0116] (2) Execute when reversing:

[0117] ① Send the scheduling information of the currently scheduled task plan obtained in S206 to the computing terminal corresponding to the immediately preceding subtask of the current subtask according to the preset communication method and encryption method;

[0118] ② Further, the "latest start time" of the current subtask is used as the "latest completion time" of the previous subtask;

[0119] ③ Further, the "scheduled quantity or progress" of the current subtask is used as the "scheduled quantity or progress" of the previous subtask;

[0120] ④ Further, the "latest start time" of the immediate predecessor subtask is equal to the "latest completion time" of the subtask minus the "subtask duration" of the subtask;

[0121] ⑤ The "work center, subtask duration, subtask quantity or progress" of the immediate predecessor subtask is calculated based on the task route and existing planning and scheduling technology, especially using the calculation technology of the ERP system and / or APS system;

[0122] ⑥ Further, proceed to step 202.

[0123] Step S209: The current task plan scheduling ends.

[0124] Step 105: Check whether the scheduling calculation of all task plans is completed; if yes, proceed to step 106; if no, proceed to step 104;

[0125] Specifically, whether the scheduling calculation of all task plans in step 103 is completed, if yes, proceed to step 106; if not, proceed to the scheduling calculation of the next task plan that follows in step 103 according to the order of the task importance scores, that is, proceed to step 104.

[0126] Step 106: Execute all task plans and monitor all preset triggering rescheduling conditions in real time until each task plan is executed.

[0127] Step 107: Whether the trigger is satisfied; if yes, proceed to step 108; if no, proceed to step 106.

[0128] Step 108: The task plan enters the task scheduling operation;

[0129] Specifically, the task plan that is triggered to be rescheduled enters into the task scheduling operation, that is, enters step S201 to perform scheduling operation; after the scheduling is completed, it enters step 106.

[0130] The present application has the following advantages over the prior art:

[0131] (1) Among the numerous scheduling constraints, artificial intelligence technology is integrated to find the optimal weight relationship between various scheduling constraints and benefits, which makes up for the defect that various scheduling methods cannot obtain the optimal condition combination;

[0132] (2) The best scheduling solution is selected by ranking the net benefits of task routes and task importance, which makes it easier to reach a consensus among supply chain partners. Distributed computing is used to form a unified scheduling computing network among the partners, effectively making up for the technical defects of different partners being unable to unify algorithms and schedules.

[0133] (3) Using computing terminals near the physical end, it is convenient to directly use the Internet of Things and device data to trigger scheduling, which makes up for the defects of centralized server scheduling in the past that is out of the field and cannot trigger scheduling in real time;

[0134] (4) The computing network composed of distributed computing terminals has enhanced computing power and made up for the shortcomings of the previous centralized server scheduling in terms of computing power requirements;

[0135] (5) Integrating artificial intelligence technology, it can predict the value changes of resources and task routes in real time and trigger them automatically, making up for the defect that the previous server scheduling could not respond in real time according to the production site conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0136] The drawings constituting part of the present application are used to provide a further understanding of the present application. The exemplary embodiments and descriptions of the present application are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0137] Figure 1 A network topology diagram of a distributed computing scheduling method for resource competition provided in an embodiment of the present application;

[0138] Figure 2 A schematic diagram of a distributed computing scheduling method for resource competition provided in an embodiment of the present application;

[0139] Figure 3 A schematic diagram of the task scheduling operation process provided in an embodiment of the present application. DETAILED DESCRIPTION

[0140] The present application is further described in detail below in conjunction with the accompanying drawings and embodiments. Obviously, the described embodiments are part of the embodiments of the present application, not all of them; all other embodiments obtained based on the embodiments of the present application without creative work are within the protection scope of the present application; it should be noted that the table data provided in this embodiment are all samples, and corresponding adjustments should be made according to the subject's own situation in actual application.

[0141] The entities mentioned in this application include but are not limited to enterprises, social organizations, public institutions and government agencies;

[0142] As introduced above, in order to address technical issues such as how to reach a consensus on scheduling algorithms among supply chain partners, how to effectively solve the scheduling computing power of servers, how to make scheduling responses in real time based on production site conditions, and how to integrate artificial intelligence technology and predict the value changes of various resources and task routes in real time, the embodiments of the present application provide a distributed computing scheduling method with resource competition.

[0143] The embodiments of the present application take the scheduling of product processing in the manufacturing industry as an example, which is also applicable to the task scheduling of project management.

[0144] For example, see Figure 1 , which is a network topology diagram of a distributed computing scheduling method for resource competition provided in an embodiment of the present application;

[0145] Each computing terminal that constitutes the distributed computing scheduling network is composed of electronic devices with computing and data processing capabilities, including but not limited to servers, personal computers, edge computers, and smart devices; a subtask list of all task routes that each computing terminal is responsible for scheduling operations is preset; each computing terminal sends the scheduling plan of the subtask obtained by its scheduled operation to the next computing terminal for scheduling operations through a preset encrypted communication method; further, the computing terminal that receives the information continues the scheduling operation until the scheduling operation is completed; the computing terminals include not only the computing terminals within the subject, but also the computing terminals of the cooperative subjects; when a subtask is undertaken by other subjects, the computing terminal responsible for the calculation of this subtask will join the computing network of the task subject together to form a distributed computing scheduling network.

[0146] For example, see Figure 1 , the computing terminal is an edge computer, and each subtask corresponds to a computing terminal; the process routes in the figure are "Route A", "Route B1", "Route B2", and "Route B3"; among them, "Route B1", "Route B2", and "Route B3" are different process routes for the same product; the computing terminal "WC10" in the figure is the computing terminal of Company B in the supply chain cooperation, and the other computing nodes are all Company A for scheduling calculations; Company A is the main body responsible for production tasks, and Company B is the supply chain cooperation processing unit, responsible for assisting Company A in completing tasks;

[0147] For example, see Figure 1 ,In route A, the subtask corresponding to the WC1 edge computer is the subtask corresponding to the WC6 edge computer; the subtask corresponding to the WC6 edge computer is the subtask corresponding to the WC1 and WC2 edge computers.

[0148] The tasks mentioned above are the work to achieve a certain goal, including but not limited to product production and project tasks; the task plan of this application refers to the work plan for the tasks that need to be delivered in the end;

[0149] The task route is a collection of all the steps required to complete a task. A task can have multiple completion plans, that is, there can be multiple task routes. The subtasks are the steps that can be accepted in the task. Each task has only one end subtask, but multiple start subtasks are allowed. When a subtask is completed, it is the turn of the next subtask to start work. Then the previous subtask is the immediate previous subtask of the next subtask, and the next subtask is the immediate next subtask of the previous subtask.

[0150] The scheduling plan mentioned above refers to the work plan for each step, i.e. each subtask, in the production or service process;

[0151] This application outputs the scheduling parameters of each subtask, including but not limited to scheduled tasks, subtasks, work centers, scheduled quantity or progress, subtask quantity or progress, subtask duration, earliest start time, latest start time, earliest finish time, and latest finish time;

[0152] The scheduling tasks are tasks that participate in the scheduling calculation and are the tasks that need to be delivered in the end;

[0153] The work center is an operating unit that is capable of performing certain types of work. A work center can be one or more direct production personnel, one or several machines with the same function, or a combination of people and equipment.

[0154] The scheduled quantity or progress mentioned above refers to the workload of the scheduled tasks that have been scheduled; if all tasks are scheduled, the scheduled quantity or progress is equal to the total quantity or progress of the scheduled tasks;

[0155] The number or progress of subtasks mentioned refers to the workload of the subtask; due to different measurement units and work content, the number of subtasks and scheduled tasks may not be equal.

[0156] See also Figure 2 , is a schematic diagram of a distributed computing scheduling method for resource competition provided in an embodiment of the present application; the specific steps of the embodiment include:

[0157] Step 101: Initialize the scheduling system and model related predictions;

[0158] Specifically, it includes:

[0159] (1) Preset work centers and work calendars of work centers;

[0160] The work completed by a work center in a task route is the subtask of the task route. There is a one-to-many relationship between subtasks and work centers, that is, each specific subtask corresponds to a unique work center; but a work center can correspond to multiple subtasks, that is, a work center can complete different specific tasks.

[0161] The work calendar of the work center is a list that defines the working hours and rest time of the work center; the tasks performed by the work center can only be scheduled to be performed within the working hours of the work center.

[0162] (2) Preset the task route parameters for all tasks;

[0163] All task route parameters described, including but not limited to task route name, subtask, subtask duration, work center, computing terminal, and next subtask;

[0164] The computing terminal is responsible for scheduling the subtask, that is, the computing terminal responsible for arranging tasks for the work center;

[0165] Each specific subtask corresponds to a specific work center and a unique computing terminal;

[0166] Embodiment, see Table 1, which is a sample of the process route of product A. This is just one sample of the process route of product A;

[0167] Table 1

[0168]

[0169] (3) Preset the communication method and encryption method between each computing terminal and other computing terminals;

[0170] In the embodiment, the communication between the computing terminals adopts HTTP communication, and the encryption method is asymmetric encryption.

[0171] (4) Presetting the calculation logic and / or algorithm rules for the earliest start time of the starting subtask of each task, and according to the calculation logic and / or algorithm rules, the earliest start time of the starting subtask of each task can be calculated;

[0172] In the embodiment, the first process is the starting point of the task route, and its earliest start time is the current time of the scheduling plan calculation.

[0173] (5) Model the revenue values ​​of all subtasks of the predicted task routes and obtain the revenue value model of the task routes;

[0174] The benefit value of the subtask mentioned above refers to the measurable benefit or importance brought to the subject by a subtask of a task route completing a unit task quantity or a unit task progress within a unit time; since the subtask is completed by a specific work center, it is equivalent to the work center completing a unit task quantity or a unit task progress within a unit time, bringing a measurable benefit or importance to the subject;

[0175] The model for predicting the revenue value of subtasks of all task routes is a model established by using artificial intelligence technology to predict the revenue value of subtasks of all task routes;

[0176] This application does not involve specific algorithms, but is intended to provide methodological support for how this application uses artificial intelligence technology;

[0177] The steps include:

[0178] ① The key factors of the benefit value of the preset subtask, that is, the factors that have a key impact on the benefit value of the subtask and can be measured;

[0179] Example, see Table 2, the key factors of the benefit value of subtask "A1" include processing cost and quality pass rate;

[0180] Table 2

[0181]

[0182] ② Obtain the historical record set big data of the benefit value of each subtask and the historical record set big data of the key factor value of the benefit value of each subtask from the management department of the relevant subject;

[0183] ③ Take the key factors of the subtask's benefit value as the independent variable and the subtask's benefit value as the dependent variable; use artificial intelligence technology to find the relationship model between the dependent variable and the independent variable; use this model to input the value of the key factors of the subtask's benefit value and output the subtask's benefit value;

[0184] The revenue value of the task route is the sum of the revenue values ​​of all subtasks of the task route. The calculation formula of the revenue value of the task route is:

[0185] In the formula:

[0186] ①n is the number of subtasks in the task route;

[0187] ②i is the i-th subtask of the task route;

[0188] ③V i is the revenue value of the ith subtask of the task route, which is calculated according to the revenue value model of the subtasks of the task route;

[0189] ④R is the revenue value of the task route;

[0190] Selecting the task route with the highest benefit value to participate in the scheduling calculation will achieve resource efficiency competition and survival of the fittest in work centers.

[0191] In the embodiment, the artificial intelligence algorithm is a multiple linear regression algorithm. Other artificial intelligence algorithms can be used in practical applications. The following are the steps for modeling the income value of the subtask in the embodiment:

[0192] ① Divide the data set into training data set and test data set;

[0193] ② Determine the independent variables; in the embodiment, the independent variables are processing cost and quality pass rate;

[0194] ③ Data standardization: standardize the data of different units to ensure that they have the same importance in the model; in the embodiment, the data standardization processing method adopts the Min-Max standardization method, and the calculation method is:

[0195] Normalized data value = (current value of data - minimum value of data) / (maximum value of data - minimum value of data)

[0196] ④Establish a computational model and train the model;

[0197] Embodiment, the multiple linear regression calculation formula is:

[0198] f(c)=w 1 a+w 2 b+c

[0199] in:

[0200] (A) f(c) is the benefit value of the subtask, that is, the benefit value of the subtask per unit time, unit task quantity or unit progress;

[0201] (B) a is the processing cost, b is the quality pass rate, and c is the intercept;

[0202] (C) w 1、 w 2 are the weights found by the multiple linear regression model;

[0203] ⑤ Conduct model evaluation and finally determine the calculation model.

[0204] (6) Preset a model for calculating task importance scores;

[0205] The task importance score is the result of evaluating the importance of each planned task; the ranking of the task importance score is the ranking of the importance of each planned task; the task importance score is the basis for the order of scheduling calculation. When the work center is overloaded during a period of time, the task plan with a high task importance score will be arranged first;

[0206] In the embodiment, for example, the net profit of the product is used as the task importance score.

[0207] (7) Calculation method for pre-setting work center capacity verification;

[0208] The work center capacity refers to the ability of the work center to complete tasks; the work center load refers to the workload assigned to the work center; the work center capacity and load must be measured in a unified unit; the work center capacity verification is to determine the relationship between the work center capacity and load and output the result; when the capacity is greater than or equal to the load, it means that the work center capacity meets the demand, otherwise it means that the work center capacity does not meet the demand; the capacity not meeting the demand is also called overload;

[0209] The technical requirements for work center capability verification are:

[0210] ① The capacity of the work center must give priority to the task plans with high task importance scores, that is, the capacity of the work center must be checked in the order of task importance scores from high to low;

[0211] ② Verify the capacity of the work center for each subtask based on the existing planning and scheduling capacity verification technology, especially the capacity verification technology of the ERP system and / or APS system, and output the number or progress of all task plans that have passed the verification and the number or progress of all task plans that have not passed the verification.

[0212] (8) Preset the pre-processing method for each task rescheduling;

[0213] There are two pre-processing methods for rescheduling each task: one is to modify the task plan, and the other is to reselect the task route; the pre-processing method for rescheduling each task is to pre-set one of the two processing methods for each task plan. When the task plan has "the number or progress that has not passed the verification", the pre-set processing method is installed for processing;

[0214] The output of the pre-processing method for rescheduling each task is preset to be a task route and a task plan;

[0215] The specific processing is as follows:

[0216] ① The modification of the task plan is performed by presetting the calculation logic and / or algorithm rules for processing the task quantity or progress, and presetting the calculation logic and / or algorithm rules for processing the latest completion time of the task, respectively, to generate a new task quantity or progress, and a new latest completion time of the task; further, the new task quantity or progress, and the new task completion time are combined with other parameters of the task plan to output a new task plan, and refresh the task importance score of the new task; further, task route selection is performed; when a task has multiple task routes, the task route with the largest benefit value is selected as the task route for scheduling the task;

[0217] Finally, the following can be output: task route and new task plan;

[0218] In the embodiment, the modification scheme of each task plan is "latest completion date = the latest completion date of the original plan + 1 day"; in actual application, the modification scheme of the task plan can be defined according to the specific task or specific product, including the modification quantity, the required date, etc.;

[0219] ② The reselection of the task route does not change the task plan, but only changes the task route; specifically, from all the task routes available in the task plan, the task routes are sorted in descending order according to their benefit values, and the task route that follows the task route currently participating in the scheduling operation is selected:

[0220] a. If the task route that follows the task exists, then output the task route that follows the task and output the current task plan at the same time;

[0221] b. If the task route that follows does not exist, enter the first processing method of "modifying the task plan" mentioned above, and finally output the task route and the new task plan.

[0222] (9) Presetting the conditions that trigger the rescheduling operation;

[0223] The specific processing is as follows:

[0224] ① Preset the conditions for triggering the scheduling operation of the subtask revenue value of each task route;

[0225] Specifically, the maximum and minimum trigger values ​​of the revenue value of each subtask of each task route are preset, and when the revenue value of the subtask is greater than the maximum value, or less than the minimum value, the rescheduling operation is triggered;

[0226] ② Preset the task importance score of each task plan to trigger the rescheduling calculation conditions;

[0227] Specifically, the maximum and minimum trigger values ​​of the task importance score of the task plan are preset, and when the task importance score of the task plan is greater than the maximum value, or less than the minimum value, the rescheduling operation is triggered.

[0228] Step 102: Inputting a mission plan;

[0229] The task plan includes but is not limited to the following parameters: task plan name, task quantity or progress, and latest completion time;

[0230] Embodiment,Table 3,Task plan example where the task is a product order;

[0231] Table 3

[0232]

[0233] Embodiment, in actual application, the latest completion time is generally not the order delivery date, and factors such as delivery preparation time also need to be considered.

[0234] Step 103: Calculate the task importance scores of all task plans and sort them from high to low, and then schedule each task plan in turn according to the sorting;

[0235] Specifically, according to the calculation method of the preset task importance scores in step 101, the task importance scores of all task plans are calculated and sorted from high to low, and the scheduling operation is performed in step 104 in sequence according to the sorting.

[0236] Step 104: Enter the task scheduling calculation process;

[0237] The specific steps are as follows:

[0238] ① The task plans involved in the scheduling calculate the revenue values ​​of each task route of all task plans according to the revenue value model of the subtasks of all task routes predicted in step 101;

[0239] ② Further, the task route with the largest benefit value is selected as the task route for scheduling calculation of the task plan;

[0240] ③ Enter the task scheduling calculation process, that is, enter step S201.

[0241] For example, see Figure 3 , is a schematic diagram of the task scheduling calculation process, the specific steps include:

[0242] Step S201: start calculating the subtasks at both ends of the task route at the same time;

[0243] The calculation starts from the subtasks at the beginning and end of the task route at the same time. The calculation starts from the starting subtask of the task route in the forward order and starts from the completed subtask of the task route in the reverse order.

[0244] The sequential scheduling is a scheduling method that starts with the starting subtask of the task route according to the selected task route, and plans and arranges tasks in the direction of task completion according to the task route;

[0245] The reverse scheduling is a scheduling method that starts with the completed subtasks of the task route according to the selected task route, and schedules tasks in the direction of starting work according to the task route.

[0246] Specifically, according to the task plan and task route of the current scheduling operation, the following processing is performed:

[0247] (1) Execute in order:

[0248] ① The "scheduled quantity or progress" of the starting subtask of the task plan is equal to the quantity or progress of the task plan;

[0249] ② Calculate the “earliest start time” of the start subtask of the task plan; specifically, according to the calculation logic and / or algorithm rules of the earliest start time of each task’s start subtask preset in step 101, the earliest start time of the start subtask of the task plan can be calculated according to this calculation logic and / or algorithm rules;

[0250] ③ The "earliest completion time" of the starting subtask of the task plan is equal to the "earliest start time" plus the "subtask duration" of the subtask;

[0251] ④ The "work center, subtask duration, subtask quantity or progress" of the starting subtask of the task plan is calculated based on the calculation technology of the task route and the existing plan scheduling, especially the calculation technology of the ERP system and / or APS system;

[0252] In the embodiment, the subtask quantity or progress is the number of products processed in each process. This number is calculated by multiplying the task plan quantity by the expanded usage of the BOM (Bill of Materials) in the ERP system. For example, if the task plan quantity is 100 pieces, then the "scheduled quantity or progress" is 100 pieces; if the BOM usage of the subtask is 2 pieces, then the product quantity of the subtask is 200 pieces.

[0253] (2) Execute when reversing:

[0254] ① The "scheduled quantity or progress" of the completed subtasks of the task plan is equal to the quantity or progress of the task plan;

[0255] ② The "latest completion time" of the completed subtask of the task plan is equal to the "latest completion time" of the task plan;

[0256] ③ The "latest start time" of the completed subtask in the task plan is equal to the "latest completion time" minus the "subtask duration" of the subtask;

[0257] ④ The "work center, subtask duration, subtask quantity or progress" of the completed subtask of the task plan is calculated based on the calculation technology of the task route and the existing plan scheduling, especially the calculation technology of the ERP system and / or APS system;

[0258] Further, the process proceeds to step S202.

[0259] Step S202: Performing capability check on the work center of the subtask;

[0260] Specifically, according to a preset work center capacity verification calculation method, a capacity verification is performed on the work center of the subtask to verify the number or progress of the output task plan that has passed the verification and the number or progress of the task plan that has not passed the verification.

[0261] Step S203: Is it a task plan that has passed the verification? If yes, proceed to step S205; if no, proceed to step S204;

[0262] Specifically, according to the output result of step 202, the processing of the number or progress of task plans that pass the verification enters step S205; if the number or progress of task plans that fail the verification enters step S204.

[0263] Step S204: output the task route and the new task plan according to the pre-set pre-processing method for rescheduling each task; and respond to the reselected task route; further, re-enter step S201;

[0264] Specifically, the task route and task plan are output according to the pre-processing method of each task rescheduling; if the task route changes, the following processing is further performed:

[0265] ① "Scheduled quantity or progress" is changed to "Quantity or progress that has passed verification";

[0266] ② “Subtask quantity or progress” is calculated based on the task route, the scheduled quantity or progress, and the existing planning and scheduling calculation technology, especially the calculation technology of the ERP system and / or APS system;

[0267] Further, by using the output task route and task plan, re-enter step S201.

[0268] Step S205: Is the plan confirmed? If yes, proceed to step S206; if no, proceed to step S204.

[0269] Step S206: obtaining a scheduling plan for the subtask;

[0270] Specifically, the confirmed forward and reverse scheduling plans of the subtask are obtained; the scheduling plan parameters include but are not limited to: scheduled tasks, subtasks, work centers, scheduled quantity or progress, subtask quantity or progress, subtask duration, earliest start time, latest start time, earliest completion time, and latest completion time.

[0271] Step S207: Is it a task completion node? If yes, proceed to step S209; if not, proceed to step S208.

[0272] Step S208: the computing terminal sends the scheduling plan of the subtask to the next computing terminal for scheduling calculation;

[0273] Specifically, the process is as follows:

[0274] (1) Execute in order:

[0275] ① Send the scheduling information of the currently scheduled task plan obtained in S206 to the computing terminal corresponding to the subtask immediately following the current subtask according to the preset communication method and encryption method;

[0276] ② Further, determine whether the next subtask is an intersection subtask; if so, take the maximum value of the "earliest completion time" of all the previous subtasks as the "earliest start time" of the intersection subtask; otherwise, directly take the "earliest completion time" of the previous subtask as the "earliest start time" of the next subtask;

[0277] The intersection subtask refers to a subtask in which two or more subtasks in the same task route converge into the same subtask after completion. This subtask is the intersection subtask; that is, an intersection subtask has multiple immediate predecessor subtasks;

[0278] ③ Further, the "scheduled quantity or progress" of the previous subtask is used as the "scheduled quantity or progress" of the next subtask;

[0279] ④ Further, the “earliest completion time” of the next subtask is equal to the “earliest start time” of the next subtask plus the “subtask duration” of the subtask;

[0280] ⑤ The "work center, subtask duration, subtask quantity or progress" of the immediately following subtask is calculated based on the task route and existing planning and scheduling technology, especially using the calculation technology of the ERP system and / or APS system;

[0281] ⑥ Further, proceed to step 202;

[0282] (2) Execute when reversing:

[0283] ① Send the scheduling information of the currently scheduled task plan obtained in S206 to the computing terminal corresponding to the immediately preceding subtask of the current subtask according to the preset communication method and encryption method;

[0284] ② Further, the "latest start time" of the current subtask is used as the "latest completion time" of the previous subtask;

[0285] ③ Further, the "scheduled quantity or progress" of the current subtask is used as the "scheduled quantity or progress" of the previous subtask;

[0286] ④ Further, the "latest start time" of the immediate predecessor subtask is equal to the "latest completion time" of the subtask minus the "subtask duration" of the subtask;

[0287] ⑤ The "work center, subtask duration, subtask quantity or progress" of the immediate predecessor subtask is calculated based on the task route and existing planning and scheduling technology, especially using the calculation technology of the ERP system and / or APS system;

[0288] ⑥ Further, proceed to step 202.

[0289] Step S209: The current task plan scheduling ends.

[0290] Step 105: Check whether the scheduling calculation of all task plans is completed; if yes, proceed to step 106; if no, proceed to step 104;

[0291] Specifically, whether the scheduling calculation of all task plans in step 103 is completed, if yes, proceed to step 106; if not, proceed to the scheduling calculation of the next task plan that follows in step 103 according to the order of the task importance scores, that is, proceed to step 104.

[0292] Step 106: Execute all task plans and monitor all preset triggering rescheduling conditions in real time until each task plan is executed.

[0293] Step 107: Whether the trigger is satisfied; if yes, proceed to step 108; if no, proceed to step 106.

[0294] Step 108: The task plan enters the task scheduling operation;

[0295] Specifically, the task plan that is triggered to be rescheduled enters into the task scheduling operation, that is, enters step S201 to perform scheduling operation; after the scheduling is completed, it enters step 106.

[0296] Compared with the prior art, the present application has the following beneficial effects:

[0297] (1) Among the numerous scheduling constraints, artificial intelligence technology is integrated to find the optimal weight relationship between various scheduling constraints and benefits, which makes up for the defect that various scheduling methods cannot obtain the optimal condition combination;

[0298] (2) The best scheduling solution is selected by ranking the net benefits of task routes and task importance, which makes it easier to reach a consensus among supply chain partners. Distributed computing is used to form a unified scheduling computing network among the partners, effectively making up for the technical defects of different partners being unable to unify algorithms and schedules.

[0299] (3) Using computing terminals near the physical end, it is convenient to directly use the Internet of Things and device data to trigger scheduling, which makes up for the defects of centralized server scheduling in the past that is out of the field and cannot trigger scheduling in real time;

[0300] (4) The computing network composed of distributed computing terminals has enhanced computing power and made up for the shortcomings of the previous centralized server scheduling in terms of computing power requirements;

[0301] (5) Integrating artificial intelligence technology, it can predict the value changes of resources and task routes in real time and trigger them automatically, making up for the defect that the previous server scheduling could not respond in real time according to the production site conditions.

[0302] The above-mentioned specific implementation modes are preferred embodiments of the present application and cannot be used to limit the present application. Any other changes or other equivalent replacement methods that do not deviate from the technical solutions of the present application are included in the protection scope of the present application.

Claims

1. A distributed computing scheduling method for resource competition, characterized in that: include: Step 101: Initialize the scheduling system and build relevant forecast models: (1) Preset work centers and work calendars of work centers; Each specific subtask corresponds to a unique work center. A work center can correspond to multiple subtasks, and a work center can complete different specific tasks. (2) Preset the task route parameters for all tasks; All task route parameters described, including but not limited to task route name, subtask, subtask duration, work center, computing terminal, and next subtask; The computing terminal is responsible for scheduling the subtask, that is, the computing terminal responsible for arranging tasks for the work center; Each specific subtask corresponds to a specific work center and a unique computing terminal; (3) Preset the communication method and encryption method between each computing terminal and other computing terminals; (4) Presetting the calculation logic and / or algorithm rules for the earliest start time of the starting subtask of each task, and calculating the earliest start time of the starting subtask of each task; (5) Model the revenue values ​​of all subtasks of the predicted task routes and obtain the revenue value model of the task routes; (6) Preset a model for calculating task importance scores; (7) Calculation method for pre-setting work center capacity verification; The technical requirements for work center capability verification are: ① The capacity of the work center must give priority to the task plans with high task importance scores, that is, the capacity of the work center must be checked in the order of task importance scores from high to low; ② Check the capacity of the work center for each subtask. Based on the existing capacity checking technology of planning and scheduling, the capacity checking technology of the ERP system and / or APS system is used to output the number or progress of all task plans that have passed the check and the number or progress of all task plans that have not passed the check; (8) Preset the pre-processing method for each task rescheduling; (9) Presetting the conditions that trigger the rescheduling operation; Step 102: Inputting a mission plan; The task plan includes but is not limited to the following parameters: task plan name, task quantity or progress, and latest completion time; Step 103: Calculate the task importance scores of all task plans and sort them from high to low, and then schedule each task plan in turn according to the sorting; According to the calculation method of the task importance score preset in step 101, the task importance scores of all task plans are calculated, and they are sorted from high to low. According to this sorting, the process proceeds to step 104 to perform scheduling calculation; Step 104: Enter the task scheduling calculation process; Here are the steps: ① The task plans involved in the scheduling calculate the revenue values ​​of each task route of all task plans according to the revenue value model of the subtasks of all task routes predicted in step 101; ② Further, the task route with the largest benefit value is selected as the task route for scheduling calculation of the task plan; ③ Enter the task scheduling operation; Step 105: Check whether the scheduling calculation of all task plans is completed; if yes, proceed to step 106; If not, proceed to step 104; Step 106: Execute all task plans and monitor all preset triggering rescheduling conditions in real time until each task plan is executed; Step 107: Whether the trigger is satisfied; if yes, proceed to step 108; If not, proceed to step 106; Step 108: The task plan enters the task scheduling operation.

2. A resource competition distributed computing scheduling method according to claim 1, characterized in that: The modeling of the revenue values ​​of the subtasks for predicting all task routes includes: The benefit value of the subtask refers to the measurable benefit or importance brought to the subject by completing the unit task quantity or unit task progress of a subtask of a task route within a unit time; The model for predicting the revenue value of subtasks of all task routes is a model established by using artificial intelligence technology to predict the revenue value of subtasks of all task routes; The steps include: ① The key factors of the benefit value of the preset subtask, that is, the factors that have a key impact on the benefit value of the subtask and can be measured; ② Obtain the historical record set big data of the benefit value of each subtask and the historical record set big data of the key factor value of the benefit value of each subtask from the management department of the relevant subject; ③ Take the key factors of the subtask's benefit value as the independent variable and the subtask's benefit value as the dependent variable; use artificial intelligence technology to find the relationship model between the dependent variable and the independent variable; use this model to input the value of the key factors of the subtask's benefit value and output the subtask's benefit value; The revenue value of the task route is the sum of the revenue values ​​of all subtasks of the task route. The calculation formula of the revenue value of the task route is: Official: ①n is the number of subtasks in the task route; ②i is the i-th subtask of the task route; ③Vi is the benefit value of the ith subtask of the task route, which is calculated according to the benefit value model of the subtasks of the task route; ④R is the revenue value of the task route; Selecting the task route with the highest benefit value to participate in the scheduling calculation will achieve resource efficiency competition and survival of the fittest in work centers.

3. The distributed computing scheduling method for resource competition according to claim 1, characterized in that: The pre-processing method of presetting each task rescheduling includes: There are two pre-processing methods for rescheduling each task: one is to modify the task plan, and the other is to reselect the task route; for each task plan, one of the above two processing methods must be preset, and when the task plan has "the number or progress that has not passed the verification", it will be processed according to this preset processing method; The output of the pre-processing method for rescheduling each task is preset to be a task route and a task plan; The specific processing is as follows: ① The task plan is modified by presetting the calculation logic and / or algorithm rules for processing the task quantity or progress, and presetting the calculation logic and / or algorithm rules for processing the latest completion time of the task, respectively performing calculations to generate a new task quantity or progress, and a new latest completion time of the task; further, the new task quantity or progress, and the new task completion time are combined with other parameters of the task plan to output a new task plan, and refresh the task importance score of the new task; further, performing task route selection; when a task has multiple task routes, selecting the task route with the largest benefit value as the task route for scheduling the task; Finally, the following can be output: task route and new task plan; ② The reselection of the task route does not change the task plan, but only changes the task route; specifically, from all the task routes available in the task plan, the task routes are sorted in descending order according to their benefit values, and the task route that follows the task route currently participating in the scheduling operation is selected: a. If the task route that follows the task exists, then output the task route that follows the task and output the current task plan at the same time; b. If the task route that follows does not exist, enter the aforementioned processing method for modifying the task plan, and finally output the task route and the new task plan.

4. The distributed computing scheduling method for resource competition according to claim 1, characterized in that: The scheduling operation process includes: Step S201: start calculating the subtasks at both ends of the task route at the same time; Step S202: Performing capability check on the work center of the subtask; Step S203: whether it is a task plan that has passed the verification; If yes, go to step S205; if no, go to step S204; Step S204: outputting the task route and the new task plan according to the pre-processing method for rescheduling each task; And respond to the reselection of the task route; Further, re-enter step S201; Step S205: whether to confirm the plan; If yes, proceed to step S206; If not, proceed to step S204; Step S206: obtaining a scheduling plan for the subtask; Scheduling parameters include but are not limited to: scheduled tasks, subtasks, work centers, scheduled quantity or progress, subtask quantity or progress, subtask duration, earliest start time, latest start time, earliest finish time, latest finish time; Step S207: Is it a task completion node? If yes, proceed to step S209; if no, proceed to step S208; Step S208: the computing terminal sends the scheduling plan of the subtask to the next computing terminal for scheduling calculation; Step S209: The current task plan scheduling ends.

5. A resource competition distributed computing scheduling method according to claim 4, characterized in that: The subtasks at the beginning and end of the task route are calculated simultaneously, including: The forward order is calculated from the starting subtask of the task route, and the reverse order is calculated from the finishing subtask of the task route. The sequential scheduling is a scheduling method that starts with the starting subtask of the task route according to the selected task route, and plans and arranges tasks in the direction of task completion according to the task route; The reverse scheduling is a scheduling method that starts with the completed subtasks of the task route according to the selected task route, and schedules tasks in the direction of starting work according to the task route. Specifically, according to the task plan and task route of the current scheduling operation, the following processing is performed: (1) Execute in sequence: ① The "scheduled quantity or progress" of the start subtask of the task plan is equal to the quantity or progress of the task plan; ② Calculate the "earliest start time" of the start subtask of the task plan; specifically, according to the calculation logic and / or algorithm rules of the earliest start time of each task's start subtask preset in step 101, according to this calculation logic and / or algorithm rules, the earliest start time of the start subtask of the task plan can be calculated; ③ The "earliest completion time" of the starting subtask of the task plan is equal to the "earliest start time" plus the "subtask duration" of the subtask; ④ The "work center, subtask duration, subtask quantity or progress" of the starting subtask of the task plan is calculated based on the calculation technology of the task route and the existing plan scheduling, using the calculation technology of the ERP system and / or APS system; (2) Execute when reversing: ① The "scheduled quantity or progress" of the completed subtasks of the task plan is equal to the quantity or progress of the task plan; ② The "latest completion time" of the completed subtask in the task plan is equal to the "latest completion time" of the task plan; ③ The "latest start time" of the completed subtask in the task plan is equal to the "latest completion time" minus the "subtask duration" of the subtask; ④ The "work center, subtask duration, subtask quantity or progress" of the completed subtask of the task plan is calculated based on the calculation technology of the task route and the existing plan scheduling, using the calculation technology of the ERP system and / or APS system; Further, enter and execute step S202.

6. The distributed computing scheduling method for resource competition according to claim 1, characterized in that: The computing terminal sends the scheduling plan of the subtask to the next computing terminal for scheduling calculation, including: Specifically, the process is as follows: (1) Execute in sequence: ① Send the scheduling information of the currently scheduled task plan obtained in S206 to the computing terminal corresponding to the subtask immediately following the current subtask according to the preset communication method and encryption method; ② Further, determine whether the subsequent subtask is an intersection subtask; if so, take the maximum value of the "earliest completion time" of all the preceding subtasks as the "earliest start time" of the intersection subtask; otherwise, directly take the "earliest completion time" of the preceding subtask as the "earliest start time" of the subsequent subtask; The intersection subtask refers to a subtask in which two or more subtasks in the same task route converge into the same subtask after completion. This subtask is the intersection subtask; that is, an intersection subtask has multiple immediate predecessor subtasks; ③ Further, the "scheduled quantity or progress" of the previous subtask is used as the "scheduled quantity or progress" of the next subtask; ④ Further, the "earliest completion time" of the next subtask is equal to the "earliest start time" of the next subtask plus the "subtask duration" of the subtask; ⑤ The "work center, subtask duration, subtask quantity or progress" of the immediately following subtask is calculated based on the task route and existing planning and scheduling technology, using the calculation technology of the ERP system and / or APS system; ⑥ Further, proceed to step S202; (2) Execute when reversing: ① Send the scheduling information of the currently scheduled task plan obtained in S206 to the computing terminal corresponding to the immediately preceding subtask of the current subtask according to the preset communication method and encryption method; ② Further, the "latest start time" of the current subtask is used as the "latest completion time" of the immediate previous subtask; ③ Further, the "scheduled quantity or progress" of the current subtask is used as the "scheduled quantity or progress" of the previous subtask; ④ Furthermore, the "latest start time" of the immediate predecessor subtask is equal to the "latest completion time" of the subtask minus the "subtask duration" of the subtask; ⑤ The "work center, subtask duration, subtask quantity or progress" of the immediate predecessor subtask is calculated based on the task route and existing planning and scheduling technology, using the calculation technology of the ERP system and / or APS system; ⑥ Further, proceed to step S202.

Images