A task planning method and system for human-machine hybrid cluster

Abstract

The present application provides a task planning method and system for a human-machine hybrid cluster, the method comprising: step 1, obtaining the capability information of personnel, the device attribute information of movable unmanned equipment and the task information; step 2, generating an initial plan; step 3, determining the consumption value of the current plan: for each possible task allocation plan, quantitative calculation is performed according to the four set optimization goals, the total consumption value corresponding to each optimization goal is obtained, and the final plan is determined through iteration; the four optimization goals include: the shortest task duration, the highest matching degree, the lowest man-machine ratio, and the task completion rate; step 4, determining the comprehensive score dominated by each optimization goal; step 5, optimizing the allocation plan using a modified version of the simulated annealing algorithm. The present application generates the best plan and multiple alternative plans to increase the flexibility and adaptability of task allocation.

Description

Technical Field

[0001] The present application relates to the technical field of hybrid cluster task allocation and scheduling, and in particular to a task planning method and system for a human-machine hybrid cluster. Background Art

[0002] With the continuous development of science and technology, mobile unmanned equipment such as ground robots, unmanned vehicles, aerial rotors and fixed-wing drones are increasingly being used in various fields. In some complex mission scenarios, such as disaster relief and material transportation, it is necessary to form a human-machine hybrid cluster with personnel and mobile unmanned equipment to perform tasks together. In order to improve the efficiency and quality of task execution, it is necessary to reasonably allocate tasks to the human-machine hybrid cluster.

[0003] Traditional task allocation methods are usually based on simple rules or a single optimization goal, and often have many limitations. For example, they only make rough allocations based on the number of personnel or equipment performance, or only allocate tasks based on the surface characteristics of the tasks, ignoring the diverse capabilities and attribute differences of the executors; some fail to fully integrate different types of equipment resources, such as ground unmanned equipment and aerial drones; some may use basic mathematical models or simple heuristic algorithms to determine task allocation, but lack comprehensive consideration of complex factors.

[0004] Insufficient consideration of the attributes and capabilities of personnel and equipment leads to unreasonable task allocation and inability to give full play to individual advantages. The optimization goal is single, and it is difficult to balance multiple factors such as time, matching degree, completion rate, and man-machine ratio. The algorithm used is prone to fall into the local optimal solution, and it is difficult to obtain the global optimal task allocation plan. Lack of flexibility and adaptability, can not be adjusted in time according to changes in tasks and environment. Summary of the invention

[0005] The present application provides a task planning method and system for a human-machine hybrid cluster, which can be used to solve the technical problem of unreasonable task allocation in traditional task allocation methods.

[0006] The present application provides a task planning method for a human-machine hybrid cluster, the method comprising:

[0007] Step 1, obtaining the capability information of personnel, the device attribute information of the mobile unmanned equipment, and the task information;

[0008] Step 2: Generate an initial solution:

[0009] A random function is used to generate the correspondence between each task and different personnel or equipment. Each task corresponds to 0, 1 or more personnel or equipment. The resources of personnel and equipment are set to be used up or not used up completely.

[0010] Step 3, determine the consumption value of the current solution: for each executable task allocation solution, perform quantitative calculations based on the four set optimization goals, obtain the total consumption value corresponding to each optimization goal, and determine the final solution through iteration; the four optimization goals include: shortest task duration, highest matching degree, lowest man-machine ratio, and task completion rate;

[0011] Step 4, determine the comprehensive score dominated by each optimization goal;

[0012] Step 5: Use a modified version of the simulated annealing algorithm to optimize the allocation plan.

[0013] The present application also provides a task planning system for a human-machine hybrid cluster, which is used to implement the solution provided in the present application. The technical details of the two are the same and will not be described one by one here. The system includes:

[0014] An information input module is used to obtain the capability information of personnel, the device attribute information of mobile unmanned equipment, and the task information;

[0015] Initial solution generation module, used to generate initial solutions:

[0016] A random function is used to generate the correspondence between each task and different personnel or equipment. Each task corresponds to 0, 1 or more personnel or equipment. The resources of personnel and equipment are set to be used up or not used up completely.

[0017] The consumption value determination module is used to determine the consumption value of the current solution: for each possible task allocation solution, quantitative calculation is performed according to the four set optimization goals, the total consumption value corresponding to each optimization goal is obtained, and the final solution is determined through iteration; the four optimization goals include: shortest task duration, highest matching degree, lowest man-machine ratio, and task completion rate;

[0018] A comprehensive score determination module is used to determine the comprehensive score dominated by each optimization objective;

[0019] The optimization module is used to optimize the allocation scheme using a modified version of the simulated annealing algorithm.

[0020] This application fully considers the detailed attributes and capabilities of personnel and mobile unmanned equipment to achieve accurate task allocation. This application sets multiple optimization goals to achieve comprehensive balance and optimization of multiple goals. This application uses an improved simulated annealing algorithm to avoid falling into local optimality and increase the possibility of finding a global optimal solution. This application generates the best solution and multiple alternative solutions to increase the flexibility and adaptability of task allocation. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 The process for forming the final allocation plan provided for this application. DETAILED DESCRIPTION

[0022] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0023] The following first introduces the embodiments of the present application in conjunction with the accompanying drawings.

[0024] Step 1, obtaining the capability information of personnel, the device attribute information of the mobile unmanned equipment, and the task information;

[0025] Capability information includes: the personnel's load capacity, the distance they can move, the average speed of movement when carrying weight and when lightly loaded;

[0026] Equipment attribute information includes the equipment's operating speed under full load and no-load conditions, load capacity, and endurance under full load and no-load conditions;

[0027] Mission information includes mission type, target type and target attributes. Mission types include exploration, transportation and rescue. Target types include exploration area, rescued personnel and materials to be transported. Target attributes include target location, rescue target weight and urgency. If the target type is exploration area, the target attributes also include the area of ​​exploration area and the complexity of regional terrain.

[0028] Step 2: Generate an initial solution:

[0029] A random function is used to generate the correspondence between each task and different personnel or equipment. Each task corresponds to 0, 1 or more personnel or equipment. The resources of personnel and equipment are set to two situations: used up or not fully used up.

[0030] Step 3: Determine the consumption value of the current solution

[0031] For each possible task allocation scheme, quantitative calculations are performed based on the four set optimization goals to obtain the total consumption value corresponding to each optimization goal, and the final scheme is determined through iteration; the four optimization goals include: shortest task duration, highest matching degree, lowest man-machine ratio, and task completion rate;

[0032] Optimization goal 1, shortest task duration:

[0033] Calculate the time it takes for the device to move from the current location to the target location and back. If it is an aerial device, estimate it based on the straight-line distance. If it is a ground device or a pedestrian, estimate it based on the straight-line distance, the moving time, and the ground environment complexity coefficient.

[0034] ;

[0035] ttravel is the time it takes for the device to travel to and from the target, D p1,p2 is the straight-line distance from the device starting point p1 to the target position p2, v 空 is the running speed of the equipment when it is unloaded, v 满 It is the operating speed of the equipment when it is fully loaded;

[0036] The time the equipment spends traveling to and from the target plus the estimated working time. If it is a rescue mission, it is the estimated rescue time. If it is an exploration mission, it is the estimated aerial exploration time or ground exploration time. If it is a transportation mission, it is the estimated loading and unloading time.

[0037] For an aerial exploration mission, if the mission name contains explore_air, obtain the area of ​​the corresponding region; then obtain the running speed of the unmanned drone assigned to perform the corresponding mission; divide the area a by the running speed v of the unmanned drone. 空 , and the aerial exploration time t of the corresponding area is obtained air :

[0038] ;

[0039] For ground exploration tasks, if the task name contains explore_ground, the complexity c of the corresponding area is obtained. area ; Next, obtain the running speed of the ground entity assigned to perform the current task when it is unloaded; multiply the area a by the complexity c area Divide by the running speed v under no-load condition 空 , and the ground exploration time of the corresponding area is obtained:

[0040] ;

[0041] For rescue and transport tasks, if the task name contains rescue or convey, obtain the type and corresponding weight of the target entity, and obtain the total load capacity of the personnel or equipment assigned to perform the task; estimate the time for rescue or loading and unloading, and add it to the time for a single entity to perform the task;

[0042] The estimated time consumption value, i.e. the first total consumption value, is the sum of the time taken by each subject to perform the task:

[0043]

[0044] C1 is the first total consumption value, m is the total number of execution entities in the allocation scheme, i represents one of the execution entities, t iIndicates the total time it takes for each execution subject to complete the task; the sum of the time it takes for each subject to complete the task indicates that the total consumption value = the sum of the total time it takes for all execution subjects to complete the task;

[0045]

[0046] where t i is the total time for a single execution subject to complete the task, t travel is the time spent on the journey to and from the target, t task It is time to carry out the task;

[0047] When the mission is transport or rescue, t task It is the time for loading and unloading supplies or rescued personnel;

[0048] When the task is to explore, t task It is the coverage search time in the target area. If it is executed by an aerial subject, the value is t air If it is executed for the ground body, the value is t ground .

[0049] Optimization goal 2: Highest matching degree: Match the execution subject’s capabilities according to the requirements of different tasks and the execution subject’s capability attributes to achieve the highest matching degree.

[0050] The target attributes required for the exploration task are the area of ​​the exploration area and the exploration type (i.e., ground-based exploration, aerial-based exploration, and aerial-based exploration). The attributes required for the execution subject include type, exploration capability, and cruising range.

[0051] First, it is set based on the rules that aerial exploration tasks are only applicable to aerial unmanned equipment, and ground exploration tasks are only applicable to staff and ground unmanned equipment. Matching is carried out according to the task requirements and the exploration device capabilities of the execution subject; the exploration capability value of each execution subject is between 1 and 10, which is a score obtained by comprehensively considering factors such as the subject's exploration device and suitable operating altitude to determine whether each execution subject has sufficient ability to complete detailed and rough exploration tasks; 1 means that only the most rough exploration can be carried out, and objects at the appropriate operating altitude (the flight equipment has its appropriate flight altitude, and the ground equipment has its exploration device to the ground) can be explored. 30 meters in level, 10 means that the most accurate exploration can be carried out, with a precision of 1 mm; the absolute value of the subject's exploration capability minus the task requirements is the efficiency consumption value of the task and subject pairing. The smaller the value, the more matched it is. If the capability is higher than the requirement, there will be a waste of capability. If the requirement is higher than the capability, the task cannot be completed satisfactorily.

[0052] Table 1 is a comparison table of exploration scale and ability value

[0053] Exploration scale 30m 10m 3m 1m 30cm 10cm 3cm 1cm 3mm 1mm Ability Value 1 2 3 4 5 6 7 8 9 10

[0054] The nature of rescue missions and transportation missions is similar, both of which involve the transportation of personnel or materials. Therefore, the main attributes that need to be included in the calculation are the weight of the target and the load capacity of the executing entity. The calculation of the matching degree is to calculate the absolute value after subtracting the target weight from the load capacity of the executing entity. If the load capacity is less than the target weight, the absolute value is multiplied by 3, and the consumption value is increased to prevent the executing entity from performing the corresponding task due to the inability to complete the transportation task.

[0055]

[0056] Where C2 is the second total consumption value, m is the total number of execution entities, i represents one of the execution entities, p i Indicates the matching degree between each execution subject and the current task in the current solution; the second total consumption value is equal to the sum of the matching degrees of all execution subjects and tasks; the matching degree is determined as follows:

[0057]

[0058] The value of the matching degree p between the subject and the task is the larger value between the first value to be taken and the second value to be taken, the first value to be taken is the difference between the subject's load capacity and the target weight; the second value to be taken is the difference between three times the target weight and the subject's load capacity; if the subject's load capacity is greater than the target weight, the matching degree = load capacity - target weight; if the subject's load capacity is less than the target weight, the monomer matching degree = (target weight - load capacity) * 3; if the subject's load capacity is equal to the target weight, the first value to be taken and the second value to be taken are both 0, and the matching degree is 0.

[0059] Optimization goal three, minimum man-machine ratio:

[0060] The man-machine ratio value represents the ratio of manpower to the total number of allocated entities in an overall allocation plan, which is the third total consumption value. The third consumption value is meaningful only when there is a surplus of execution entity resources. Otherwise, when all resources need to be allocated, the man-machine ratio value remains constant, that is, the ratio of manpower to the number of execution entities in the cluster.

[0061]

[0062] The third total consumption value C3 is equal to n in the allocation scheme human The proportion of all executing entities; n human is the number of personnel; n device is the number of devices;

[0063] Optimization goal four, task completion rate:

[0064] The completion rate indicates the status of task completion, that is, the ratio of the number of tasks that are expected to be completed to the total number of tasks. In actual situations, it is the ratio of the number of assigned tasks to the total number of tasks. The completion rate is applicable when there are many tasks and the resources that can be allocated cannot meet all task requirements at the same time.

[0065]

[0066] The fourth total consumption value C4 is equal to the number of tasks that have been assigned in the allocation plan, that is, the number that is expected to be completed t a The ratio of the total number to the total number; the total number is the number of tasks that have been assigned t a and the number of unassigned tasks t ua sum.

[0067] Step 4: Determine the comprehensive score dominated by each optimization goal:

[0068] Step 41, in order to facilitate subsequent unified processing and comparison, the four total consumption values ​​need to be standardized and mapped to the range of 0 to 1; the standardization method is as follows:

[0069] For a certain allocation scheme, the standardized scoring method for a certain optimization goal is as follows:

[0070]

[0071] The above formula is a calculation method for the standardized score of an allocation scheme on a certain optimization target, where the consumption value C i is the consumption value of the current allocation scheme on a certain optimization target, i.e., any one of C1, C2, C3, C4, and the minimum consumption value C min is the minimum consumption value among all allocation schemes recorded on the optimization target, and the maximum consumption value C max It is the maximum consumption value among all allocation schemes recorded on the optimization target;

[0072] Obtain a standardized score for each target of the allocation plan, and the higher the score, the lower the consumption value and the more perfect the allocation plan;

[0073] Step 42, for the allocation scheme, by adjusting the weights, weighted addition of the standardized scores on each objective, respectively obtain a comprehensive score dominated by each optimization objective:

[0074]

[0075] Among them, S1 is the standardized score of the dominant goal, S2, S3, and S4 are the standardized scores of the other three goals; the dominant goal is one of the four optimization goals, and the four optimization goals take turns to dominate;

[0076] Through such weighted calculation, each task allocation plan can get a comprehensive score for each goal. The score takes into account the four different optimization goals, which facilitates the overall evaluation and comparison of different plans.

[0077] Step 5: Use a modified version of the simulated annealing algorithm to optimize the allocation plan.

[0078] For each optimization target solution, the optimized allocation solution is iteratively optimized using the adjusted simulated annealing algorithm. In each iteration, the following steps are performed:

[0079] Step 51, determining the initial temperature, heating threshold, reheating factor, cooling rate, and number of iterations of the simulated annealing algorithm;

[0080] Step 52, generate a neighbor solution: generate a new neighbor solution according to the current solution; if there is an unused execution subject, randomly select one of the unused execution subjects and assign it to a randomly selected task; if all execution subjects have been used, randomly select two tasks and exchange their execution subjects;

[0081] Step 53, determining the comprehensive score change: using the difference between the comprehensive score of the neighbor solution and the comprehensive score of the current solution;

[0082] Step 54, accepting the neighbor solution: decide whether to accept the neighbor solution as the new current solution according to the acceptance probability; the acceptance probability is calculated based on the difference in comprehensive scores and the current temperature; in the actual code implementation process, determine whether the acceptance probability is greater than a random value between 0 and 1, if it is greater than or equal to it, accept it, if it is less than it, do not accept it:

[0083]

[0084] The above formula represents the acceptance probability p accept The determination method is that if the comprehensive score S of the neighbor solution neighbor Greater than the comprehensive score S of the current solution current , then the acceptance probability is 1, and it is directly accepted, replacing the current solution with the neighbor solution; if the comprehensive score of the neighbor solution S neighbor Lower than or equal to the overall score S of the current solution current , then the acceptance probability is 1 minus the ratio of the difference between the current solution and the neighbor solution score to the current temperature T;

[0085] Step 55, updating the optimal solution: updating the optimal solution and its consumption value and score according to the above acceptance probability;

[0086] Step 56, cooling: reducing the current temperature according to a preset cooling rate;

[0087] Step 57, reheating: whenever the reheating threshold is reached, the current temperature is increased according to the reheating factor to increase the possibility of jumping out of the local optimum;

[0088] Step 58, termination condition: when one of the three conditions of reaching the preset maximum number of iterations, the temperature being lower than the termination threshold, and the execution time limit is reached, the iteration process is terminated;

[0089] Step 59, output result: finally output the optimized task allocation plan and its corresponding minimum cost;

[0090] When the algorithm detects that it is trapped in a local optimum, step 57 is executed, and step 57 is executed at most twice. Step 57 adjusts the parameters of the algorithm, just like increasing the "vitality" of the search, so that it has the opportunity to jump out of the local optimum and continue to explore a better solution. In addition, the number of "reheating" is strictly limited to two times to control the computational cost and avoid excessive search. In the algorithm, linear functions are used to calculate probabilities, which speeds up the calculation, and because there is a reheating step, the possibility of premature convergence and falling into a local optimal solution due to the linear function is reduced.

[0091] The above-described embodiments of the present application do not constitute a limitation on the protection scope of the present application.

Claims

1. A task planning method for a human-machine hybrid cluster, characterized in that: The method comprises: Step 1, obtaining the capability information of personnel, the device attribute information of the mobile unmanned equipment, and the task information; Step 2: Generate an initial solution: A random function is used to generate the correspondence between each task and different personnel or equipment. Each task corresponds to 0, 1 or more personnel or equipment. The resources of personnel and equipment are set to be used up or not used up completely. Step 3, determine the consumption value of the current solution: for each executable task allocation solution, perform quantitative calculations according to the four set optimization goals, obtain the total consumption value corresponding to each optimization goal, and determine the final solution through iteration; the four optimization goals include: shortest task duration, highest matching degree, lowest man-machine ratio, and task completion rate; the four optimization goals correspond to the first total consumption value, the second total consumption value, the third total consumption value, and the fourth total consumption value respectively; Among them, the first total consumption value corresponds to the total consumption value of the shortest task duration, which is equal to the sum of the time taken by each subject to perform the task; The second total consumption value corresponds to the total consumption value of the highest matching degree, which is equal to the sum of the matching degrees of all execution entities and tasks; The third total consumption value corresponds to the total consumption value of the minimum man-machine ratio, which is equal to the ratio of personnel to the total allocated subjects in an overall allocation plan; The fourth total consumption value corresponds to the total consumption value of the task completion rate, which is equal to the ratio of the number of tasks that have been assigned in the allocation plan, that is, the number that is expected to be completed, to the total number of tasks; Step 4, determine the comprehensive score dominated by each optimization goal; Step 5, using a modified version of the simulated annealing algorithm to optimize the allocation scheme; Step 5: Use a modified version of the simulated annealing algorithm to optimize the allocation scheme, including: Step 51, determining the initial temperature, heating threshold, reheating factor, cooling rate, and number of iterations of the simulated annealing algorithm; Step 52, generate a neighbor solution: generate a new neighbor solution according to the current solution; if there is an unused execution subject, randomly select one of the unused execution subjects and assign it to a randomly selected task; if all execution subjects have been used, randomly select two tasks and exchange their execution subjects; Step 53, determining the comprehensive score change: using the difference between the comprehensive score of the neighbor solution and the comprehensive score of the current solution; Step 54, accepting the neighbor solution: decide whether to accept the neighbor solution as the new current solution based on the acceptance probability; the acceptance probability is calculated based on the difference in comprehensive scores and the current temperature; determine whether the acceptance probability is greater than a random value between 0 and 1, if it is greater than or equal to it, accept it, if it is less than it, do not accept it: The above formula represents the acceptance probability p accept The method of determining if the comprehensive score S of the neighbor solution neighbor Greater than the comprehensive score S of the current solution current , then the acceptance probability is 1, and it is directly accepted, replacing the current solution with the neighbor solution; if the comprehensive score of the neighbor solution S neighbor Lower than or equal to the overall score S of the current solution current , then the acceptance probability is 1 minus the ratio of the difference between the current solution and the neighbor solution score to the current temperature T; Step 55, update the optimal solution: update the optimal solution, consumption value and score according to the acceptance probability; Step 56, cooling: reducing the current temperature according to a preset cooling rate; Step 57, reheating: whenever the reheating threshold is reached, the current temperature is increased according to the reheating factor to increase the possibility of jumping out of the local optimum; Step 58, termination condition: when one of the three conditions of reaching the preset maximum number of iterations, the temperature being lower than the termination threshold, and the execution time limit is reached, the iteration process is terminated; Step 59, output result: finally output the optimized task allocation plan and the corresponding minimum cost; When the algorithm detects that it is trapped in a local optimum, step 57 is executed, and step 57 is executed at most twice.

2. The method according to claim 1, characterized in that: In step 1, the capability information includes: the load capacity of the personnel, the distance they can move, the weight they carry, and the average speed of movement when they are lightly loaded; Equipment attribute information includes the equipment's operating speed under full load and no-load conditions, load capacity, and endurance under full load and no-load conditions; Mission information includes mission type, target type and target attributes; mission types include exploration, transportation and rescue; target types include exploration area, rescued personnel and materials to be transported; target attributes include target location and rescue target weight. If the target type is an exploration area, the target attributes also include the area of ​​the exploration area and the complexity of the regional terrain.

3. The method according to claim 1, characterized in that Optimization goal 1, the shortest task duration is determined by the following method: Calculate the time it takes for the device to move from the current location to the target location and back. If it is an aerial device, estimate it based on the straight-line distance. If it is a ground device or a pedestrian, estimate it based on the straight-line distance, the moving time, and the ground environment complexity coefficient. t travel is the time it takes for the device to travel to and from the target, D p1,p2 is the straight-line distance from the device starting point p1 to the target position p2, v 空 is the running speed of the equipment when it is unloaded, v 满 It is the operating speed of the equipment when it is fully loaded; The time the equipment spends traveling to and from the target plus the estimated working time. If it is a rescue mission, it is the estimated rescue time. If it is an exploration mission, it is the estimated aerial exploration time or ground exploration time. If it is a transportation mission, it is the estimated loading and unloading time. For an aerial exploration mission, if the mission name contains explore_air, obtain the area of ​​the corresponding region; then obtain the running speed of the unmanned drone assigned to perform the corresponding mission; divide the area a by the running speed v of the unmanned drone. 空 , and the aerial exploration time t of the corresponding area is obtained air : For ground exploration tasks, if the task name contains explore_ground, the complexity c of the corresponding area is obtained. area ; Next, obtain the running speed of the ground entity assigned to perform the current task when it is unloaded; multiply the area a by the complexity c area Divide by the running speed v under no-load condition 空 , and the ground exploration time of the corresponding area is obtained: For rescue and transport tasks, if the task name contains rescue or convey, obtain the type and corresponding weight of the target entity, and obtain the total load capacity of the personnel or equipment assigned to perform the task; estimate the time for rescue or loading and unloading, and add it to the time for a single entity to perform the task; The estimated time consumption value, i.e. the first total consumption value, is the sum of the time taken by each subject to perform the task: C1 is the first total consumption value, m is the total number of execution entities in the allocation scheme, i represents one of the execution entities, t i Indicates the total time it takes for each execution subject to complete the task; the sum of the time it takes for each subject to complete the task indicates the total consumption value = the sum of the total time it takes for all execution subjects to complete the task; t i =t travel +t task ; where t i is the total time for a single execution subject to complete the task, t travel is the time spent on the journey to and from the target, t task It is time to carry out the task; When the mission is transport or rescue, t task It is the time for loading and unloading supplies or rescued personnel; When the task is to explore, t task It is the coverage search time in the target area. If it is executed by an aerial subject, the value is t air If it is executed on the ground, the value is t ground .

4. The method according to claim 1, characterized in that: Optimization goal 2, the highest matching degree is determined by the following method: Match the execution subject's capabilities according to different tasks and their attributes to achieve the highest matching degree; The rules stipulate that aerial exploration missions are only applicable to aerial unmanned equipment, and ground exploration missions are only applicable to staff and ground unmanned equipment. They are matched according to the mission requirements and the exploration device capabilities of the execution subject. The exploration capability value of each execution subject is between 1 and 10; 1 means that only the most rough exploration can be carried out, and objects of 30 meters can be explored at a suitable operating altitude; 10 means that the most precise exploration can be carried out, with an accuracy of 1 mm. Where C2 is the second total consumption value, m is the total number of execution entities, i represents one of the execution entities, p i Indicates the matching degree between each execution subject in the current plan and the current task; The second total consumption value is equal to the sum of the matching degrees of all execution entities and tasks, where the matching degree is determined as follows: The value of the matching degree p between the execution subject and the task is the larger value between the first value to be taken and the second value to be taken, the first value to be taken is the difference between the load capacity of the execution subject and the target weight; the second value to be taken is the difference between three times the target weight and the load capacity of the subject; if the load capacity of the subject is greater than the target weight, the matching degree = load capacity - target weight; if the load capacity of the subject is less than the target weight, the matching degree = (target weight - load capacity) * 3; if the load capacity of the subject is equal to the target weight, the first value to be taken and the second value to be taken are both 0, and the matching degree is 0.

5. The method according to claim 1, characterized in that: Optimization goal three, the minimum man-machine ratio, is determined by the following method: The man-machine ratio value represents the ratio of manpower to the total allocated entities in an overall allocation plan, which is the third total consumption value; The third total consumption value C3 is equal to n in the allocation scheme human The proportion of all executing entities; n human is the number of personnel; n device is the number of devices.

6. The method according to claim 1, characterized in that Optimization goal four, task completion rate, is determined by the following method: The completion rate is the ratio of the number of tasks that are expected to be completed to the total number of tasks, which is equal to the ratio of the number of tasks that have been assigned to the total number of tasks; The fourth total consumption value C4 is equal to the number of tasks that have been assigned in the allocation plan, that is, the number that is expected to be completed t a The ratio of the total number to the total number; the total number is the number of tasks that have been assigned t a and the number of unassigned tasks t ua sum.

7. The method according to claim 1, characterized in that Step 4: Determine the comprehensive score dominated by each optimization goal, including: Step 41, normalize the four total consumption values ​​and map them to a range from 0 to 1; the normalization method is as follows: For a certain allocation scheme, the standardized scoring method for a certain optimization goal is as follows: The above formula is a calculation method for the standardized score of an allocation scheme on a certain optimization target, where the consumption value C i is the consumption value of the current allocation scheme on a certain optimization target, and the minimum consumption value C min is the minimum consumption value among all allocation schemes recorded on the optimization target, and the maximum consumption value C max It is the maximum consumption value among all allocation schemes recorded on the optimization target; Obtain a standardized score for each target of the allocation plan, and the higher the score, the lower the consumption value and the more perfect the allocation plan; Step 42, for the allocation scheme, by adjusting the weights, weighted addition of the standardized scores on each objective, respectively obtain a comprehensive score dominated by each optimization objective: S general =0.7*S1+0.1*S2+0.1*S3+0.1*S4; Among them, S1 is the standardized score of the dominant target, and S2, S3, and S4 are the standardized scores of the other three targets. The dominant target is one of the four optimization targets, and the four optimization targets take turns to dominate.

8. A task planning system for a human-machine hybrid cluster, characterized in that: The system comprises: An information input module is used to obtain the capability information of personnel, the device attribute information of mobile unmanned equipment, and the task information; Initial solution generation module, used to generate initial solutions: A random function is used to generate the correspondence between each task and different personnel or equipment. Each task corresponds to 0, 1 or more personnel or equipment. The resources of personnel and equipment are set to be used up or not used up completely. The consumption value determination module is used to determine the consumption value of the current plan: for each executable task allocation plan, quantitative calculation is performed according to the four set optimization goals, the total consumption value corresponding to each optimization goal is obtained, and the final plan is determined through iteration; the four optimization goals include: shortest task duration, highest matching degree, lowest man-machine ratio, and task completion rate; the four optimization goals correspond to the first total consumption value, the second total consumption value, the third total consumption value, and the fourth total consumption value in turn; Among them, the first total consumption value corresponds to the total consumption value of the shortest task duration, which is equal to the sum of the time taken by each subject to perform the task; The second total consumption value corresponds to the total consumption value of the highest matching degree, which is equal to the sum of the matching degrees of all execution entities and tasks; The third total consumption value corresponds to the total consumption value of the minimum man-machine ratio, which is equal to the ratio of personnel to the total allocated subjects in an overall allocation plan; The fourth total consumption value corresponds to the total consumption value of the task completion rate, which is equal to the ratio of the number of tasks that have been assigned in the allocation plan, that is, the number that is expected to be completed, to the total number of tasks; A comprehensive score determination module is used to determine the comprehensive score dominated by each optimization objective; The optimization module is used to optimize the allocation scheme using a modified version of the simulated annealing algorithm; The optimization module is specifically used to determine the initial temperature, heating threshold, reheating factor, cooling rate, and number of iterations of the simulated annealing algorithm; Generate neighbor solution: Generate a new neighbor solution based on the current solution; if there is an unused execution subject, randomly select one of the unused execution subjects and assign it to a randomly selected task; if all execution subjects have been used, randomly select two tasks and exchange their execution subjects; Determine the change in comprehensive score: use the difference between the comprehensive score of the neighbor solution and the comprehensive score of the current solution; Accept neighbor solution: Decide whether to accept the neighbor solution as the new current solution based on the acceptance probability; the acceptance probability is calculated based on the difference in comprehensive scores and the current temperature; determine whether the acceptance probability is greater than a random value between 0 and 1, if it is greater than or equal to it, accept it, if it is less than it, do not accept it: The above formula represents the acceptance probability p accept The method of determining if the comprehensive score S of the neighbor solution neighbor Greater than the comprehensive score S of the current solution current , then the acceptance probability is 1, and it is directly accepted, replacing the current solution with the neighbor solution; if the comprehensive score of the neighbor solution S neighbor Lower than or equal to the overall score S of the current solution current , then the acceptance probability is 1 minus the ratio of the difference between the current solution and the neighbor solution score to the current temperature T; Update the optimal solution: Update the optimal solution, consumption value and score according to the acceptance probability; Cooling: lower the current temperature according to the preset cooling rate; Reheating: Whenever the reheating threshold is reached, the current temperature is increased according to the reheating factor to increase the possibility of escaping the local optimum; Termination condition: The iteration process ends when one of the three conditions is reached: the preset maximum number of iterations, the temperature is lower than the termination threshold, or the execution time limit; Output results: The final output is the optimized task allocation plan and the corresponding minimum cost; When the algorithm detects that it is trapped in a local optimum, reheating is performed, and reheating is performed at most twice.

9. The system according to claim 8, characterized in that Capability information includes: the personnel's load capacity, the distance they can move, the average speed of movement when carrying weight and when lightly loaded; Equipment attribute information includes the equipment's operating speed under full load and no-load conditions, load capacity, and endurance under full load and no-load conditions; Mission information includes mission type, target type and target attributes; mission types include exploration, transportation and rescue; target types include exploration area, rescued personnel and materials to be transported; target attributes include target location, rescue target weight and urgency. If the target type is exploration area, the target attributes also include the area of ​​exploration area and the complexity of regional terrain.

Images