Dynamic task allocation method and system for multiple unmanned ships based on improved CBBA algorithm

By introducing priority selection strategies, distance reward and punishment factors and time window constraints into the CBBA algorithm, we construct the priority selection cluster and cost function of unmanned ships, which solves the problem of high path cost and low task completion in task redistribution. It realizes more efficient task allocation and real-time response capabilities.

CN116911535BInactive Publication Date: 2024-11-22HARBIN INST OF TECH AT WEIHAI
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202310791477.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-06-30
Publication Date
2024-11-22
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

Traditional CBBA algorithms have high path cost and low task completion in the calculation of task reassignment schemes, and cannot effectively deal with the disappearance of unmanned ship mission capabilities and dynamic environmental changes.

Method used

Introduce priority selection strategies, distance reward and punishment factors and time window constraints. By building priority selection clusters and cost functions of unmanned ships, improve the speed of algorithm calculation, ensure the rationality and real-time nature of task allocation plans, and use bidding to resolve conflicts to ensure tasks. efficiency and accuracy of allocation.

Benefits of technology

It improves the task completion degree, shortens the total working time of the unmanned ship, reduces the cost of paths, responds to new tasks and emergencies in real time, and verifies the effectiveness of the improved algorithm.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116911535B_ABST
    Figure CN116911535B_ABST
Patent Text Reader

Abstract

The present invention provides a method and system for dynamic task allocation of multiple unmanned ships based on an improved CBBA algorithm, and belongs to the technical field of dynamic task allocation of multiple unmanned ships. In order to solve the problem that the traditional CBBA algorithm has high path cost indicators and low task completion in the calculation of task redistribution schemes, and the task capability of the unmanned ship disappears for new tasks or emergencies. By constructing an unmanned ship priority selection cluster, the calculation speed of the algorithm is improved, the algorithm converges quickly, and the indicator variables of the distance reward and punishment factor and the judgment time window constraint are introduced. The task package is constructed according to the cost function, so that the task allocation scheme tends to let the unmanned ship perform tasks with a shorter distance, so that each individual with task execution capability in the environment can construct a task package by itself with the goal of maximizing its own benefits and minimizing losses; adding time window constraints, excluding tasks that do not meet the requirements and retaining tasks that meet the requirements, and then performing conflict resolution, and finally judging whether a consensus is reached between the unmanned ships.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of dynamic task allocation for multiple unmanned ships, and in particular to a method and system for dynamic task allocation for multiple unmanned ships based on an improved CBBA algorithm. Background Art

[0002] The ocean is rich in natural resources, and autonomous underwater robots are one of the important tools for human resource exploration and utilization, scientific exploration and research. With the development and progress of unmanned ship technology, the increasing demand for marine resource exploration and development, and people's desire for a more comprehensive and profound understanding of the ocean, the work undertaken by unmanned ships is becoming more and more complex and the workload is increasing. Due to the limitations of their own detection capabilities and energy, single unmanned ships can no longer adapt to new mission requirements, so multi-unmanned ships, as a new research direction, have attracted the attention of more and more scholars. The autonomous allocation technology of unmanned ship tasks can realize the detection of multiple targets in vast sea areas.

[0003] In practical applications, the traditional CBBA algorithm is generally used for task allocation. After the unmanned ship system obtains the static task allocation plan issued by the command center, it will start working according to the established plan. However, the actual situation often cannot remain unchanged for a long time. New tasks may appear during the operation of the unmanned ship system, and the task capability may disappear due to emergencies in the unmanned ship. At this time, in order to better adapt to the dynamic environment, it is necessary to dynamically adjust the task allocation plan and reallocate the tasks to be executed in the current environment. Considering the high path cost index and low task completion rate of the traditional CBBA algorithm in the calculation of the task reallocation plan, an improved CBBA algorithm is proposed, and the performance of the improved algorithm is verified through simulation experiments. Summary of the invention

[0004] The technical problems to be solved by the present invention are:

[0005] In order to solve the problems of high path cost index and low task completion in the calculation of task reallocation scheme in the traditional CBBA algorithm, and the loss of unmanned ship mission capability due to new tasks or emergencies.

[0006] The present invention adopts the following technical solutions to solve the above technical problems:

[0007] The present invention provides a method for dynamic task allocation of multiple unmanned ships based on an improved CBBA algorithm, comprising the following steps:

[0008] Step 1: Read mission information, unmanned ship information and environment information;

[0009] Step 2: Introduce a priority selection strategy and build an unmanned ship priority selection cluster to improve the algorithm's calculation speed and make the algorithm converge quickly;

[0010] Step 3: Introduce distance reward and penalty factors and indicator variables for judging time window constraints, and construct task packages based on the cost function to make the task allocation plan tend to allow the unmanned ship to perform tasks with shorter distances and be more adaptable to task environments with time window constraints;

[0011] Step 4: After the task package is constructed according to step 3, the task package is updated. After the update, the judgment process of the time window constraint is added. The task execution order is used as the judgment order. The tasks in the task package are judged in turn whether they meet the time window constraint. If not, the task is removed from the task sequence. The remaining tasks are judged whether they meet the time window constraint until all tasks in the task package meet their constraint conditions.

[0012] Step 5: After completing the task package update, resolve conflicts for the task packages that meet the time window constraints, and assign the tasks to the unmanned ships with the lowest cost and the highest benefit through bidding;

[0013] Step 6: After the conflict is resolved, determine whether the unmanned ships have reached a consensus. If so, output the final task allocation plan. If not, repeat steps 3 to 6 until a consensus is reached and the final task allocation plan is output.

[0014] Furthermore, in step 2, the method for constructing the unmanned ship priority selection cluster is as follows: Assume that the i-th cluster contains inum i Tasks, starting from unmanned ship No. 1, calculate its distance to all other task points in the environment, divide the inum1 task with the greatest benefit into cluster No. 1, then remove the tasks in cluster No. 1 from the environment, continue to calculate the cluster of unmanned ship No. 2, and so on, until the cluster information of the unmanned ships in the environment is confirmed to be completed.

[0015] Furthermore, a distance reward and penalty factor φ is introduced into the cost function. The distance reward and penalty factor φ is shown in formula (1):

[0016]

[0017] Among them, pxy j Indicates the mission location that the unmanned ship is currently performing; pxy ja Indicates the location of the mission performed by the unmanned ship at the last moment; pxy jn Indicates the location of the mission that the unmanned ship is expected to perform next;

[0018] When φ≤1, the probability of this situation occurring can be increased by reducing the cost; when φ>1, the probability of this situation occurring can be reduced by increasing the cost.

[0019] Furthermore, an indicator variable for judging the time window constraint is introduced into the cost function, as shown in formula (2):

[0020]

[0021] Among them, tst j Indicates the time when task number j starts execution; tet j Indicates the estimated completion time of task number j; trans(T j ) represents the time window interval that the task numbered j needs to meet;

[0022] Furthermore, after introducing the indicator variables of distance reward and penalty factors and time window constraints, it is necessary to update the task package according to the cost consumed by different unmanned ships to perform the task. The unmanned ship with the lowest cost can add the task to the task package. For example, the unmanned ship numbered i joins the task numbered j, and the cost Co consumed is ij The calculation method of is shown in formula (3),

[0023]

[0024] Among them, ω1 and ω2 are inertia weight factors, which are used to adjust the importance of the two costs; t represents the expected joining time of the task numbered j, and t-1 represents the last moment of the unmanned ship; Represents the position vector of the unmanned ship numbered i at time t; represents the position vector of task number j at time t; tc ij It represents the time required for the unmanned ship numbered i to perform the task numbered j.

[0025] Furthermore, in step five, the method for conflict resolution is that if a conflict is detected, the value of each task package to different unmanned ships is calculated using a cost function through communication between the unmanned ships that cause the conflict. Then, each task is used as the auctioneer and all the unmanned ships are used as auctioneers. All tasks are auctioned separately in turn, and the tasks are assigned to the unmanned ships with the lowest cost and the highest benefit by bidding.

[0026] Furthermore, after the task is resolved, each unmanned ship needs to unify the winner list and the winner bid. The winner list is the current solution, and the winner bid is the basis for obtaining such a solution.

[0027] Furthermore, after unifying the winner list and the winner's bid, the task volume information and time consumption information performed by each unmanned ship in the current system are calculated. If the task execution volume of a certain ship exceeds the average task volume of the unmanned ship and the time consumption information of the ship is equal to the system time consumption information, the part of its task package that exceeds the average task volume is removed, and the removed tasks are returned to the task environment.

[0028] Furthermore, in step six, the method for judging whether a consensus is reached between the unmanned ships is to check whether there are any changes in the winner list and the winner's bid of each unmanned ship after another communication cycle. If there is no change, the current task allocation plan is output as the final task allocation plan; if there is a change, it is necessary to re-construct the task, auction the task, optimize the time window, and judge whether a consensus is reached again.

[0029] A multi-unmanned ship dynamic task allocation system based on an improved CBBA algorithm, the system has a program module corresponding to the above steps, and executes the steps in the multi-unmanned ship dynamic task allocation method based on the improved CBBA algorithm during operation.

[0030] Compared with the prior art, the present invention has the following beneficial effects:

[0031] The present invention provides a method and system for dynamic task allocation of multiple unmanned ships based on an improved CBBA algorithm. The method improves the calculation speed of the algorithm by constructing an unmanned ship priority selection cluster, makes the algorithm converge quickly, reduces the time consumed in the algorithm conflict resolution process, introduces distance reward and punishment factors and indicator variables for judging time window constraints, and constructs task packages according to cost functions, so that the obtained task allocation scheme tends to allow unmanned ships to perform tasks at a relatively close distance, so that each individual with task execution capability in the environment can construct a task package by itself with the goal of maximizing its own benefits and minimizing its own losses; adds time window constraints, excludes tasks that do not meet the requirements and retains tasks that meet the requirements, then performs conflict resolution, and finally judges whether a consensus is reached among the unmanned ships;

[0032] The traditional CBBA algorithm and the improved CBBA algorithm of the present invention are compared under the same environment. The solution obtained by the traditional CBBA algorithm allows the unmanned ship to perform tasks that are too far away, the driving time is too long, and the distribution of tasks is not reasonable enough. After the improved CBBA is added to the preferred task cluster, the unmanned ship can better select the appropriate solution, and the shortened voyage further shortens the total working time of the unmanned ship, which is convenient for meeting the time window of possible tasks in the future.

[0033] In summary, compared with the traditional CBBA algorithm, the improved CBBA algorithm improves the task completion rate by 36.4%, reduces the average waiting time of the unmanned ship by 46.30%, reduces the path cost, ensures the task completion, and can respond to new tasks or emergencies in real time, thereby verifying the effectiveness of the improved CBBA algorithm of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 It is a flow chart of a method for dynamic task allocation of multiple unmanned ships based on an improved CBBA algorithm in an embodiment of the present invention;

[0035] Figure 2 A process diagram of cluster division in an embodiment of the present invention;

[0036] Figure 3 A schematic diagram of a situation in which a task that does not meet the time window constraint is removed in an embodiment of the present invention;

[0037] Figure 4 It is a planar schematic diagram of a task allocation scheme obtained by using a traditional CBBA algorithm and an improved CBBA algorithm in a smaller scale environment in an embodiment of the present invention;

[0038] Figure 5 A three-dimensional schematic diagram of a solution obtained by using a traditional CBBA algorithm in an embodiment of the present invention;

[0039] Figure 6 A three-dimensional schematic diagram of a solution obtained by using an improved CBBA algorithm in an embodiment of the present invention;

[0040] Figure 7 It is a schematic diagram of the working state of an unmanned boat in an exemplary environment according to an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In the description of the present invention, it should be noted that the terms "first", "second", and "third" mentioned in the embodiments of the present invention are only used for descriptive purposes and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined as "first", "second", and "third" may explicitly or implicitly include one or more of the features.

[0042] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0043] Specific implementation plan 1: Combine Figures 1 to 7 As shown, the present invention provides a method for dynamic task allocation of multiple unmanned ships based on an improved CBBA algorithm, comprising the following steps:

[0044] Step 1: Read mission information, unmanned ship information and environment information;

[0045] Step 2: Build an unmanned ship priority cluster, including:

[0046] Assume that the i-th cluster contains inum i Tasks, starting from unmanned ship No. 1, calculate the distance between it and all other task points in the environment, divide the inum1 task with the largest benefit into cluster No. 1, then remove the tasks in cluster No. 1 from the environment, continue to calculate the cluster of unmanned ship No. 2, and so on, until the cluster information of unmanned ships in the environment is confirmed and completed; the specific process of cluster division is as follows Figure 2 As shown;

[0047] This step is to reduce the conflicts between unmanned ships during the construction of the initial task package. By introducing a priority selection strategy, the algorithm's calculation speed is improved, allowing the algorithm to converge quickly. It can be understood that before the initial task package is constructed, the tasks to be executed in the environment are first divided into Unum clusters that can correspond one-to-one with the unmanned ships. When the unmanned ship constructs its corresponding task package, it will give priority to selecting tasks from the cluster. When all the tasks in the cluster are selected, the unmanned ship can select the remaining tasks according to its own execution capabilities and environmental constraints.

[0048] When constructing a cluster, we first need to ensure that the number of tasks in each cluster is approximately the same based on the principle of average distribution. This is also to make the resource allocation of the unmanned ship system as balanced as possible and minimize the time consumed by the algorithm conflict resolution process.

[0049] Step 3: Introduce the distance reward and punishment factor and the indicator variable of the judgment time window constraint, and construct the task package according to the cost function. Specifically,

[0050] In order to avoid the situation where unmanned ships cross a large area in practical applications and adapt to solving the task environment with time window constraints, a distance reward and penalty factor φ is introduced into the cost function. The distance reward and penalty factor φ is shown in formula (1):

[0051]

[0052] Among them, pxy j Indicates the mission location that the unmanned ship is currently performing; pxy ja Indicates the location of the mission performed by the unmanned ship at the last moment; pxy jn Indicates the location of the mission that the unmanned ship is expected to perform next;

[0053] The purpose of introducing the distance reward and penalty factor φ is to make the obtained task allocation plan tend to let the unmanned ship perform tasks with shorter distances. The distance reward and penalty factor φ is the ratio of the next expected range of the task currently performed by the unmanned ship to the previous confirmed range. The larger the φ is, the larger the range the unmanned ship will cross, and the greater the cost of performing the task. When φ≤1, it means that the unmanned ship has been crossing within a small range, and the probability of this situation can be increased by reducing the cost; when φ>1, it means that the unmanned ship will cross a large range next, so the probability of this situation can be reduced by increasing the cost;

[0054] In addition, in order to better adapt to the task environment with time window constraints, an indicator variable for judging the time window constraint is introduced, as shown in formula (2):

[0055]

[0056] Among them, tst j Indicates the time when task number j starts execution; tet j It indicates the estimated completion time of task number j, trans(T j ) represents the time window interval that the task numbered j needs to meet;

[0057] After introducing the indicator variables of distance reward and penalty factors and time window constraints, the task package needs to be updated according to the cost consumed by different unmanned ships to perform the task. The unmanned ship with the lowest cost can add the task to the task package. For example, the unmanned ship numbered i joins the task numbered j, and the cost Co consumed ij The calculation method of is shown in formula (3),

[0058]

[0059] Among them, ω1 and ω2 are inertia weight factors, which are used to adjust the importance of the two costs; t represents the expected joining time of the task numbered j, and t-1 represents the last moment of the unmanned ship; Represents the position vector of the unmanned ship numbered i at time t; represents the position vector of task number j at time t; tc ij It represents the time required for the unmanned ship numbered i to perform the task numbered j;

[0060] It can be understood that, based on the environmental information that has been read, the algorithm is based on the greedy principle, allowing each individual with task execution ability in the environment to build a desired task package set TA by itself with the goal of maximizing its own benefits and minimizing its losses. iDuring this phase, the unmanned ship will continue to add new tasks to its own mission package collection and build an execution sequence until the length limit of the mission package is reached.

[0061] Assume that there are Tnnum tasks to be executed in the new environment. Initially, the task package TA i , winner set W i And the winner cost function table E i are all empty sets, and the task package is TA i is a one-dimensional matrix, the length of which represents the total number of tasks in the mission plan obtained by the unmanned ship; the winner set W i is a 1×Tnnum matrix, the elements in the matrix are the serial numbers of the unmanned ships, indicating the distribution of tasks in the current solution; the winner cost function table E i It is used to represent the cost that the unmanned ship needs to consume to perform the tasks in its plan in the current allocation plan;

[0062] Step 4: After the task package is constructed according to step 3, the task package is updated. After the update, the judgment process of the time window constraint is added. The task execution order is used as the judgment order. The tasks in the task package are judged in turn whether they meet the time window constraint. If not, the task is removed from the task sequence. The remaining tasks are judged whether they meet the time window constraint until all tasks in the task package meet their constraint conditions.

[0063] When some tasks in the environment must be completed within a specific time interval, tasks need to be reallocated according to the time window constraints; combined with Figure 3 As shown, t1, t2 and t3 are the task intervals of the three tasks respectively, among which the task interval of t1 does not exceed the deadline of task t1, and the deadline of task t2 is within the task interval of t2, so task t2 is removed, and tasks t1 and t3 are left. At this time, the task interval of t3 does not exceed the deadline of task t3, and finally tasks t1 and t3 are retained;

[0064] Step 5: After completing the task package update, resolve conflicts for the task packages that meet the time window constraints. Specifically,

[0065] If a conflict is detected, the communication between the unmanned ships that caused the conflict is used to calculate the value of each task package to different unmanned ships using the cost function. Then, each task is used as the auctioneer and all the unmanned ships are used as the auctioneers. All tasks are auctioned in turn, and the tasks are assigned to the unmanned ships with the lowest cost and the highest benefit by bidding. Each unmanned ship is made to unify the winner list and the winner bid. The winner list is the current solution, and the winner bid is the basis for obtaining this solution.

[0066] In addition, the calculation of the task volume and time consumption information of each unmanned ship in the current system is added. If the task volume of a ship exceeds the average task volume of the unmanned ship and the time consumption information of the ship is equal to the system time consumption information, the part of its task package that exceeds the average task volume will be removed, and the removed tasks will be returned to the task environment;

[0067] In the task package construction stage, each unmanned ship will communicate with the unmanned ship system cluster only when a new task is added to its own task package. Due to the difference in computing time, there is a possibility that tasks are assigned multiple times. Therefore, the goal of this stage is to eliminate the conflict factors in the current task allocation plan of each unmanned ship in the unmanned ship cluster.

[0068] Step 6: After the conflict is resolved, determine whether the unmanned ships have reached a consensus. If so, output the final task allocation plan. If not, repeat steps 3 to 6 until a consensus is reached and the final task allocation plan is output;

[0069] The method to determine whether a consensus has been reached between the unmanned ships is to check whether the winner list and the winner's bid of each unmanned ship have changed after another communication cycle. If there is no change, the current task allocation plan is output as the final task allocation plan; if there is a change, it is necessary to re-construct the task, auction the task, optimize the time window, and determine whether a consensus has been reached again.

[0070] The other combinations and connection relationships of this embodiment are the same as those of the first embodiment.

[0071] Example

[0072] System simulation environment: AMD Ryzen 7 4800H with Radeon Graphics 2.90 GHz, PC with 16 GB memory, Windows 10 operating system, MATLAB R2018a.

[0073] A two-dimensional map model is used, and the coordinate system is a plane coordinate system. The speed of each unmanned ship is different, but the completion time for each task is the same. For the convenience of calculation, the path consumption problem that may exist during the execution of the unmanned ship is not considered. The goal is to find a solution that makes the overall performance of the system the best, that is, the solution with the minimum total system time consumption and total path consumption. Detailed environmental information is shown in Table 1.

[0074] Table 1 Example environment information table

[0075]

[0076]

[0077] In the example environment, 5 unmanned ships and 30 tasks are set, and the time for the unmanned ship to complete each task is 2 minutes. Table 1 shows the time window interval required for each task. The starting point of the interval is the time when the task appears, and the end point is the time when the task cannot be executed. The traditional CBBA algorithm and the improved CBBA algorithm of the present invention are used to solve the dynamic task allocation scheme for the current environment.

[0078] The schematic diagram of the scheme obtained by the traditional CBBA algorithm and the improved CBBA algorithm of the present invention is shown in FIG. Figure 4 As shown in Figure 2, in order to facilitate the analysis of the satisfaction of the time window conditions, Figure 4 The timeline is introduced on the basis of , and the time window required for each task is displayed in the figure, such as Figure 5 and Figure 6 As shown,

[0079] The task allocation scheme solved by the traditional CBBA algorithm and the improved CBBA algorithm of the present invention in the current environment, as well as the task completion amount, total time cost and total path cost information of the system under the scheme are all recorded in Table 2. Table 2 is as follows:

[0080] Table 2 Information table of task allocation scheme obtained by algorithm

[0081]

[0082] As shown in Table 2, the improved CBBA algorithm has a larger task completion volume than the traditional CBBA algorithm, while the time cost and path cost are smaller. This situation shows that the solution obtained by the traditional CBBA algorithm has the unmanned boat perform tasks that are too far away, the driving time is too long, and the task allocation is not reasonable. The large amount of time consumed during the driving process causes the subsequent tasks with time window constraints to be unable to be completed within the specified time, which also leads to low task completion volume.

[0083] Since the environment is very dynamic and changeable, and all tasks in the environment are subject to time window constraints, the unmanned ship will have three states during the operation of the system, namely waiting, working and sailing. The working state diagram of the unmanned ship under the traditional CBBA algorithm is shown in the figure. Figure 7 As shown in (a), the working state diagram of the unmanned ship under the scheme obtained by the improved CBBA algorithm is as follows Figure 7 (b)

[0084] Combination Figure 7As shown in the figure, under the scheme obtained by the traditional CBBA algorithm, the average waiting time of the unmanned ship is 5.7847 minutes. Under the scheme obtained by the improved CBBA algorithm, the average waiting time of the unmanned ship is 3.1066 minutes. The waiting time of the unmanned ship in the scheme obtained by the improved CBBA algorithm is shorter. The emergence of this situation shows that after the improved CBBA joins the preferred task cluster, the unmanned ship can better select the appropriate scheme, and the shortened voyage further shortens the total working time of the unmanned ship, which is convenient for the subsequent meeting of the time window of possible tasks in the future.

[0085] Comprehensive analysis shows that under the current environment, the improved CBBA algorithm improves the task completion rate by 36.4% and reduces the average waiting time of the unmanned ship by 46.30% compared with the traditional CBBA algorithm, which verifies the effectiveness of the improved CBBA.

[0086] Although the present invention is disclosed as above, the protection scope of the present invention is not limited thereto. Those skilled in the art may make various changes and modifications without departing from the spirit and scope of the present invention, and these changes and modifications will fall within the protection scope of the present invention.

Claims

1. A method for dynamic task allocation of multiple unmanned ships based on an improved CBBA algorithm, characterized in that: The following steps are involved: Step 1: Read mission information, unmanned ship information and environment information; Step 2: Introduce a priority selection strategy and build an unmanned ship priority selection cluster to improve the algorithm's calculation speed and make the algorithm converge quickly; Step 3: Introduce distance reward and penalty factors and indicator variables for judging time window constraints, and construct task packages based on the cost function to make the task allocation plan tend to allow unmanned ships to perform tasks with shorter distances and be more adaptable to task environments with time window constraints, including: The distance reward and penalty factor φ is introduced into the cost function. The distance reward and penalty factor φ is shown in formula (1): Among them, pxy j Indicates the mission location that the unmanned ship is currently performing; Indicates the location of the mission performed by the unmanned ship at the last moment; Indicates the location of the mission that the unmanned ship is expected to perform next; When φ≤1, the probability of this situation can be increased by reducing the cost; when φ>1, the probability of this situation can be reduced by increasing the cost; Introduce the indicator variable of the time window constraint into the cost function, as shown in formula (2): Among them, tst j Indicates the time when task number j starts execution; tet j Indicates the estimated completion time of task numbered j; After introducing the indicator variables of distance reward and penalty factors and time window constraints, the task package needs to be updated according to the cost consumed by different unmanned ships to perform the task. The unmanned ship with the lowest cost can add the task to the task package. For example, the unmanned ship numbered i joins the task numbered j, and the cost Co consumed ij The calculation method of is shown in formula (3), Among them, ω1 and ω2 are inertia weight factors, which are used to adjust the importance of the two costs; t represents the expected joining time of the task numbered j, and t-1 represents the last moment of the unmanned ship; Represents the position vector of the unmanned ship numbered i at time t; represents the position vector of task number j at time t; tc ij It represents the time required for the unmanned ship numbered i to perform the task numbered j; Step 4: After the task package is constructed according to step 3, the task package is updated. After the update, the judgment process of the time window constraint is added. The task execution order is used as the judgment order. The tasks in the task package are judged in turn whether they meet the time window constraint. If not, the task is removed from the task sequence. The remaining tasks are judged whether they meet the time window constraint until all tasks in the task package meet their constraint conditions. Step 5: After completing the task package update, resolve conflicts for the task packages that meet the time window constraints, and assign the tasks to the unmanned ship with the lowest cost and the highest benefit through bidding, including: The conflict resolution method is that if a conflict is detected, the communication between the unmanned ships that caused the conflict is used to calculate the value of each task package to different unmanned ships using the cost function, and then each task is used as the auctioneer, and all unmanned ships are used as auctioneers. All tasks are auctioned separately in turn, and the tasks are assigned to the unmanned ships with the lowest cost and the highest benefit in a bidding manner; Step 6: After the conflict is resolved, determine whether the unmanned ships have reached a consensus. If so, output the final task allocation plan. If not, repeat steps 3 to 6 until a consensus is reached and the final task allocation plan is output.

2. The method for dynamic task allocation of multiple unmanned ships based on the improved CBBA algorithm according to claim 1 is characterized in that: In step 2, the method of constructing the unmanned ship priority selection cluster is as follows: Assume that the i-th cluster contains inum i Tasks, starting from unmanned ship No. 1, calculate its distance to all other task points in the environment, divide the inum1 task with the greatest benefit into cluster No. 1, then remove the tasks in cluster No. 1 from the environment, continue to calculate the cluster of unmanned ship No. 2, and so on, until the cluster information of the unmanned ships in the environment is confirmed to be completed.

3. The method for dynamic task allocation of multiple unmanned ships based on the improved CBBA algorithm according to claim 2 is characterized in that: After the task is resolved, each unmanned ship needs to unify the winner list and the winner bid. The winner list is the current solution, and the winner bid is the basis for obtaining this solution.

4. The method for dynamic task allocation of multiple unmanned ships based on the improved CBBA algorithm according to claim 3 is characterized by: After unifying the winner list and the winner's bid, the task volume information and time consumption information of each unmanned ship in the current system are calculated. If the task execution volume of a certain ship exceeds the average task volume of the unmanned ship and the time consumption information of the ship is equal to the system time consumption information, the part of its task package that exceeds the average task volume is removed, and the removed tasks are returned to the task environment.

5. The method for dynamic task allocation of multiple unmanned ships based on the improved CBBA algorithm according to claim 4 is characterized in that: In step six, the method for judging whether a consensus has been reached between the unmanned ships is to check whether there are any changes in the winner list and the winner's bid of each unmanned ship after another communication cycle. If there is no change, the current task allocation plan is output as the final task allocation plan; if there is a change, it is necessary to re-construct the task, auction the task, optimize the time window, and judge whether a consensus has been reached again.

6. A multi-unmanned ship dynamic task allocation system based on an improved CBBA algorithm, characterized by: The system has a program module corresponding to the steps of any one of claims 1 to 5, and executes the steps of the above-mentioned multi-unmanned ship dynamic task allocation method based on the improved CBBA algorithm when running.