UUV cluster dynamic task planning method based on multi-target genetic algorithm, program, equipment and storage medium

Abstract

The invention discloses a UUV cluster dynamic task planning method based on a multi-target genetic algorithm, a program, equipment and a storage medium, and belongs to the field of underwater multi-UUV cooperative detection of multiple targets. According to the method, firstly, for path planning of regional task points, chromosome representation is completed through sequential coding, chromosomes are selected and ranked randomly, and the chromosome with the highest fitness value is selected from each group; then, in each group of the current population, selecting an optimal parent individual to carry out crossover and mutation operations so as to improve the quality of offspring individuals, and carrying out updating to obtain a next-generation population; and finally, obtaining a plurality of optimal individuals, and outputting a task planning path result. Dividing the search area into a plurality of task sub-areas, and completing the dynamic task planning of the UUV cluster according to the target distribution condition of the sub-areas and the condition of each UUV. According to the method, the search strategy can be adjusted in real time according to the real-time detection result and the environment change, so that the target distribution non-uniformity and the cluster efficiency difference are effectively relieved.

Description

Technical Field

[0001] The present invention relates to the field of underwater multi-UUV collaborative detection of multiple targets, and in particular to a UUV cluster dynamic task planning method, program, device and storage medium based on a multi-objective genetic algorithm. Background Art

[0002] The waters near ports are characterized by a relatively concentrated distribution of facilities and a complex local environment. Unmanned underwater vehicles (UUVs), with their small size, strong autonomous operation capabilities, and high concealment, offer significant potential for application and development in search and detection missions near ports. Especially in the complex hydrological conditions surrounding ports, UUVs can efficiently and effectively conduct complex regional search and detection missions, providing critical information support. However, in scenarios involving distributed operational systems, the inherent operational limitations of a single UUV often make it difficult to meet the demands of large-scale search missions. Consequently, current research has delved into the technical scope of multi-UUV systems. Multi-UUV collaborative control is a core component of underwater unmanned system research. Through the collaborative cooperation of multiple intelligent agents, a range of complex tasks can be successfully accomplished. Furthermore, to address the heterogeneous distribution of underwater targets, this study designs an optimal dynamic mission planning strategy for multiple UUVs and explores the operational model of collaborative operations. Furthermore, the study of multi-UUV collaborative search strategies has important theoretical value for accelerating the practical application of UUVs.

[0003] In solving the multi-UUV cooperative task planning problem, ensuring that the algorithm has excellent convergence performance and can generate higher quality solution set has always been the core issue in the research field. To solve this challenge, the literature (Task allocation of multiple autonomous underwater vehicle system based on multiobjective optimization. IEEE International Conference on Mechatronics and Automation. IEEE, 2016, pp. 2512-2517) studies the single ant colony system single-objective optimization and the interaction between different ant colony systems to obtain the Pareto solution set of the multi-objective optimization problem, providing a better solution for UUV ant colony reconnaissance task planning; the literature (Multi-criteria optimal routing for last-mile parcel delivery. IEEE International Conference on Smart Mobility (SM). IEEE, 2022, pp. 21-26) proposes an adaptive genetic algorithm (AGA) based on clustering, uniform cost search, greed and bionic algorithm.Compared with the path planning strategy of traditional genetic algorithms that only rely on a single criterion objective function, the adaptive genetic algorithm proposed by the institute can comprehensively consider multiple complex factors, including travel time, road capacity, terrain elevation changes, the number of traffic lights, and the number of turns, so as to find a better and more comprehensive driving path; the literature (A dynamic task allocation algorithm for heterogeneous uuv swarms. Sensors, vol. 22, no. 6, p. 2122, 2022) proposed a dynamic extended consensus bundle algorithm (DECBBA) based on consistency constraints to solve the UUV swarm task planning problem under communication constraints with high efficiency and good performance; the literature (Dynamic task allocation algorithm for moving targets interception. IEEE Transactions on Systems, Man, and Cybernetics: Systems, vol. 53, no. 2, pp. 776–788, 2022) studied the dynamic task allocation problem under communication and speed constraints, and proposed an improved target evaluation metric to solve the invalid selection problem when selecting the k best participants; the literature (Research on dynamic task allocation for multiple unmanned A distributed immune multi-agent algorithm (DIMAA) based on an immune multi-agent network framework was proposed to address the distributed task allocation problem of multiple unmanned aerial vehicles (UAVs) performing search and rescue missions. In summary, the complexity and diversity of the task environment often lead to cluster task decision-making and planning issues in swarm collaborative tasks. For example, the poor quality of the global solution, the conflict between randomness and convergence, and the algorithm's weak ability to escape from local optimal solutions significantly affect the algorithm's global optimization capabilities and inherent convergence performance. Furthermore, most of these studies focus on static task planning and fail to consider the kinematic and sonar detection characteristics of the UUV nodes themselves. Furthermore, due to these UAV characteristics, individuals in a swarm may not start or end their search process at the same location, meaning that most existing methods are unsuitable for practical underwater search and detection missions.

[0004] To this end, the study designed a search and detection sub-area division strategy based on the UUV's 3-DOF model and forward-looking sonar model. Taking into account the various constraints that exist during the UUV cluster's mission execution, the MGA and elite MGA algorithms were proposed to solve the multi-objective optimization problem of UUV clusters in complex underwater environments. Furthermore, when the UUV cluster performs a dynamic collaborative search mission, when the number of detected targets in a certain area is higher than the average level, the MDC-MGA algorithm can adaptively reallocate the mission targets in that area to each UUV node, achieving dynamic scheduling of each UUV, reducing the adverse effects of target distribution non-uniformity and differences in the UUV cluster's target detection efficiency, and greatly improving the UUV cluster's regional search and target detection efficiency. Summary of the Invention

[0005] This paper proposes a UUV cluster dynamic task planning method based on multi-objective genetic algorithm (MDC-MGA), which aims to address the problems of limited number of UUVs, communication range and energy consumption of UUV clusters in multi-UUV area detection environments, and the difficulty of existing algorithms to adjust task decisions according to the real-time location and working status of UUV clusters.

[0006] The present invention provides a UUV cluster dynamic task planning method based on a multi-objective genetic algorithm, comprising the following steps:

[0007] Path planning is performed for regional task points based on a multi-objective genetic algorithm; the multi-objective genetic algorithm outputs the optimal individual with the highest fitness value after iteration to obtain the task planning path;

[0008] A UUV cluster dynamic task planning method is used to search and detect the area, and the search area is divided into multiple task sub-areas. The dynamic task planning of the UUV cluster is completed according to the target distribution in the sub-area and the conditions of each UUV. The UUV cluster dynamic task planning method is subject to local locking, kinematics, communication distance, endurance, total path length, and time consumption constraints when allocating task planning paths.

[0009] Furthermore, a multi-objective genetic algorithm is used for path planning. The specific process is as follows:

[0010] Step 1.1: Encode the path planning of the regional task points in sequence to generate an initial population containing K chromosomes;

[0011] X k =(x1,x2,…,x z )1≤k≤K

[0012] Where X is the total search path of each UUV; K is the number of chromosomes; z is the length of each chromosome, which also depends on the number of total mission points;

[0013] Step 1.2: Calculate the fitness value of each chromosome individual in each group of the population, and select the chromosome with the highest fitness value from each group as the optimal individual;

[0014] Step 1.3: Perform color crossover and chromosome mutation on each group of optimal individuals according to the crossover probability and mutation probability to generate new chromosomes, and then update the population to obtain the next generation population. Return to step 1.1;

[0015] Step 1.4: Repeat steps 1.1 to 1.3 above until the maximum number of iterations is reached, obtain the optimal individual with the highest fitness value in each group, and output the task planning path result.

[0016] Furthermore, the fitness function of the chromosome individual is:

[0017]

[0018] in, is the total length of collision-free paths of all K UUVs under the scheme corresponding to chromosome X; K·d task_point is the task point spacing; T j J is the time consumed by the UUV to reach the predicted allocation area from the current position and perform target search, which is the larger value of the UUV movement time and the turning time; max Take the maximum penalty result in history; α, β, λ are weight parameters, and the multi-objective problem is transformed into a single-objective problem by adjusting the weight parameters; J(X) is the constraint condition of the chromosome; T max The maximum penalty time in the history of UUV mission execution;

[0019] The constraints of the chromosome are designed based on the performance parameters of the UUV cluster and the mission requirements. The constraints are as follows:

[0020] J=c x f x +c T T+c v,a (f v +f a )+c t f t +c s f s +c ξ ζ

[0021] Among them, c x 、c T 、c v,a 、c t 、c s 、c ξ is a weight parameter; ζ is a "slack" variable used to measure the degree to which the current solution exceeds the threshold on various hard constraints;

[0022] The relative position f between the UUV and the target sub-area x for:

[0023]

[0024] Among them, x (m) is the center position of the mth sub-region, q represents the current position of the UUV, and n is the number of sub-regions;

[0025] The time consumption T of the UUV from the current position to the predicted assigned area and performing target search is:

[0026]

[0027] Among them, {P j ,ξ j} is the position of the next movement point of UUV and heading angle leagth(P(·)) is the length of the collision-free path; v max and are the maximum speed and yaw rate respectively; the time consumption T is calculated by calculating the navigation time and the UUV turning time, and the larger value of the two times represents the total time loss in the execution of the task;

[0028] The control penalty is:

[0029]

[0030] The UUV kinematic constraints are:

[0031]

[0032] Among them, f v 、f a These are the velocity constraints and acceleration kinematic constraints added respectively; is the speed of the UUV at the next moving point; is the acceleration of the next moving point of the UUV; is the angular velocity of the next moving point of the UUV; is the angular acceleration of the next moving point of the UUV; a max is the maximum acceleration;

[0033] Communication distance constraint f for multiple UUVs in the process of regional collaborative search t for:

[0034]

[0035] Among them, x c,j,μ and x c,i,μare the current positions of the j-th UUV and the i-th UUV respectively;

[0036] The endurance constraint of UUV during mission execution s for:

[0037]

[0038] Among them, S max is the maximum range of UUV, S c,i,μ The current distance traveled by the UUV.

[0039] Furthermore, in step 1.3, for each group of optimal individuals, the crossover probability of the optimal individual is calculated based on the difference between the individual fitness and the optimal / average fitness of the group. and mutation probability

[0040]

[0041] Among them, F k is the fitness value of the i-th individual in the current generation; F best is the maximum fitness of all individuals in the current generation, F avg is the arithmetic mean of all individual fitness in the current generation; P c,min With P m,min is the lower limit of the adaptive probability; k c and k m They represent two constant parameters, which are used to adjust the weights in the task.

[0042] Furthermore, the UUV cluster dynamic task planning method is used to search and detect the area, and the specific process is as follows:

[0043] Step 2.1: Divide the search area into multiple sub-areas;

[0044] Step 2.2: The UUV executes the partition search strategy and assigns the obtained mission planning path to each UUV node;

[0045] Step 2.3: Execute the sub-area obstacle avoidance process; if the UUV finds an obstacle, re-update the mission planning path based on the current UUV real-time position; if the UUV does not find an obstacle, execute the sub-area target search process;

[0046] Step 2.4: Execute the sub-area target search process; the UUV performs the search task along the current mission planning path. When the number of mission point targets identified by the UUV in a sub-area exceeds that of other sub-areas or the distance between adjacent mission points is smaller than the distance between adjacent mission points in other areas, the UUV cluster will synchronously complete mutual communication and mark the key area. The UUV closest to the key area will also enter the key area. Multiple UUVs in the key area will search for mission points according to the newly updated mission planning path.

[0047] Step 2.5: After the UUV completes the search of the sub-area according to the mission planning path; if more than 15% of the mission points in the current sub-area have not been searched, the mission planning path is updated again and steps 2.3 to 2.4 are repeated until less than 15% of the mission points have not been searched, and the target search process ends.

[0048] Furthermore, when the UUV executes the partition search strategy, it first calculates the search cost of each task sub-region and determines whether the sub-region with the minimum search cost is preset as locked, that is:

[0049]

[0050] V i =min{f x ,LockArea(Sub n um)=0}

[0051] Among them, LockArea(Sub n um) is the sub-area Sub n um's locking flag. If the flag is equal to 0, it means that the sub-area has not been locked as a mission area by any UUV; otherwise, the sub-area has been locked by a UUV; V i Represents the cost of the sub-region Subnum, considering the sub-region closest to the current UUV node and not locked as the search task sub-region that the UUV needs to perform.

[0052] Furthermore, when the UUV executes the partition search strategy, there are key constraints. Through dynamic local optimization of the constraints, real-time task allocation for the current sub-region is achieved;

[0053] The constraints are:

[0054] J=c x f x +c T T+c v,a (f v +f a )+c t f t +c s fs +c ξ ζ

[0055] The present invention also provides a computer device / equipment / system, comprising a memory, a processor, and a computer program stored in the memory, wherein when the processor executes the computer program, the steps of the UUV cluster dynamic task planning method based on a multi-objective genetic algorithm are implemented.

[0056] The present invention also provides a computer-readable storage medium having a computer program / instruction stored thereon, which, when executed by a processor, implements the steps of any of the above-mentioned methods for dynamic task planning of UUV clusters based on a multi-objective genetic algorithm.

[0057] The present invention also provides a computer program product, comprising a computer program / instruction, which, when executed by a processor, implements the steps of any of the above-mentioned methods for dynamic task planning of UUV clusters based on a multi-objective genetic algorithm.

[0058] The beneficial effects of the present invention are:

[0059] (1) In the existing technology, a global path optimization is usually performed before the mission begins, and the path remains fixed during the execution process; however, this method realizes the dynamic scheduling of each UUV node through the MDC-MGA framework, and relies on the multi-objective genetic operator to continuously recalculate and publish the latest mission route of each UUV in each generation, so that the system can adjust the search strategy in real time according to the real-time detection results and environmental changes, thereby effectively alleviating the target distribution non-uniformity and cluster efficiency differences.

[0060] (2) Compared with traditional allocation that only focuses on a single indicator (such as the shortest path or the least time), the method of the present invention incorporates "total path length, maximum task duration and constraint violation penalty" into a unified fitness and balances them through weight parameters to achieve true multi-objective collaborative optimization, taking into account both task efficiency and practical feasibility.

[0061] (3) The method of the present invention is different from the conventional genetic algorithm. The present invention adopts the MGA process of "random grouping + sequential encoding + breakpoint moving crossover + elite reintroduction": random grouping: to ensure the diversity of the group and avoid premature convergence; breakpoint position moving crossover: to expand the feasible solution space while retaining the continuity of the task point; through the parallel injection of historical elite individuals, it can both accelerate convergence and prevent falling into the local optimum.

[0062] (4) The method of the present application has fewer constraints than traditional methods, and the weight of the offline penalty function is simply reduced; the present application calculates the following multi-dimensional constraints in real time and dynamically accumulates them through the relaxation variable ζ: kinematic constraints (maximum speed, acceleration), communication distance constraints (effective range of underwater communication link), endurance constraints (remaining range and distance traveled), local locking constraints (prevent repeated allocation of sub-regions). Through the online penalty model, local dynamic optimization and global integrated scheduling of constraints are realized, ensuring that the planning result is efficient and safe and feasible. BRIEF DESCRIPTION OF DRAWINGS

[0063] Figure 1 is a schematic diagram of chromosome selection and grouping;

[0064] Figure 2 is a schematic diagram of chromosome crossing;

[0065] Figure 3 is a schematic diagram of chromosome mutation;

[0066] Figure 4 is a schematic diagram of chromosome crossing mutation;

[0067] Figure 5 is a pseudo-code schematic diagram of the MGA module of the present application;

[0068] Figure 6 is a flow chart of the MDC-MGA task planning of the present application;

[0069] Figure 7 is a global map of task search area division of the present application;

[0070] Figure 8 is a trajectory diagram of the task planning of the present application, wherein figure (a) is a trajectory diagram of the first stage task planning of the UUV cluster; figure (b) is a trajectory diagram of the second stage task planning of the UUV cluster;

[0071] Figure 9 is the first stage task planning of the UUV cluster of the present application; wherein figure (a) is a task point diagram of the first stage task planning of the UUV cluster, figure (b) is a path diagram of the first stage task planning of the UUV cluster, and figure (c) is an optimal decision convergence diagram of the first stage task planning of the UUV cluster;

[0072] Figure 10 is the second stage task planning of the UUV cluster of the present application; wherein figure (a) is a task point diagram of the second stage task planning of the UUV cluster, figure (b) is a path diagram of the second stage task planning of the UUV cluster, and figure (c) is an optimal decision convergence diagram of the second stage task planning of the UUV cluster. Figure 11 is a static target search process of different methods. DETAILED DESCRIPTION

[0073] The present invention will be further described below with reference to the accompanying drawings.

[0074] The present invention discloses a UUV cluster dynamic mission planning method based on a multi-objective genetic algorithm, comprising:

[0075] Step 1: Use the MGA collaborative area search method to plan the task point path;

[0076] Step 1.1: Set up the encoding and decoding scheme. This study uses an iterative counter and sequential encoding scheme to encode the unlocked areas within the port detection area and randomly generates an initial population of K chromosomes.

[0077] Chromosome representation: For path planning of regional task points, sequential coding is used to complete chromosome representation, also known as natural number coding;

[0078] X k =(x1,x2,…,x z )1≤k≤K

[0079] According to the number of mission points, X represents the total search path of each UUV, K represents the number of chromosomes, and the length of each chromosome is set to z, which also depends on the number of total mission points.

[0080] Chromosome selection and sorting: For each generation, the individuals in the population are divided into l groups, and each chromosome is randomly sorted into K individuals. The number of breakpoints is defined as n break_point =l-1, where n break_point ≤(z-1). Without considering the arrangement between groups, the size of the solution space is In order to simplify the description of the proposed MGA, the length of each chromosome is set to n = 8, the number of chromosomes is set to K = 9, and the individuals are divided into l = 3 groups. Figure 1 As shown in , the core idea is to select the chromosome with the highest fitness value from each group. However, unlike directly screening out the l best individuals with the highest fitness value from the contemporary population, this random grouping method ensures population diversity.

[0081] Step 1.2: Calculate the fitness of each individual in the population and record the best individual in the current population. The fitness function is the only indicator of whether an individual can survive and provides a screening basis for the subsequent individual selection algorithm.

[0082] Step 1.3: Randomly divide the population generated in the previous step into l groups. Select the best individual in each group, that is, the one with the highest fitness. Individuals with higher fitness are more likely to pass on their genes to the next generation, thus continuously eliminating individuals with lower fitness. At the same time, selecting the best individuals in each group also ensures genetic diversity.

[0083] Step 1.4: Genetic manipulation is performed on each group of selected individuals. Various crossover, mutation, and crossover-mutation operations are performed on the genes of the selected chromosomes with certain crossover and mutation probabilities to generate new chromosomes. Genetic manipulations maximize the preservation of excellent genes while increasing genetic diversity through mutation, thereby improving the probability of finding the optimal solution.

[0084] Changing the breakpoint position to perform chromosome crossover; directly applying the single-objective crossover algorithm to multi-objective problems will produce many infeasible solutions. To address this phenomenon, the following methods are usually used to further address it:

[0085] discard unworkable solutions;

[0086] Use penalty functions to reduce the fitness of infeasible solutions;

[0087] Construct the matching operator so that only feasible solutions are generated;

[0088] Turn an unworkable solution into a workable solution.

[0089] However, these solutions may result in the loss of multiple populations in the current generation or the loss of continuity information between task points. Figure 2 As shown in the figure, the study proposes a method to change the breakpoint position for crossover operation.

[0090] In Table 1, the changes in routes R2 and R3 reflect the reak_Point2 Population evolution after position change.

[0091] Table 1 Updated UUV paths after cross operation

[0092]

[0093] Traditional methods can consume significant time searching the solution space for non-global optimal solutions. This algorithm preserves the continuity of information between task points, thereby expanding the solution space. Furthermore, by altering breakpoints to ensure that all subsequent individuals are feasible, the algorithm can find the global optimal solution with greater probability and efficiency.

[0094] Select the best parent individual and perform crossover and mutation operations; since the study uses sequential coding, the legitimacy of the mutated chromosome is guaranteed. Figure 3 As shown, chromosome mutations all occur in the form of switching, sliding, and flipping.

[0095] In Table 2, the changes in the R1-R3 route reflect the population evolution after the mutation operation.

[0096] Table 2 Updated UUV paths after performing mutation operations

[0097]

[0098] In addition, if Figure 4 As shown, the study also applied a variety of crossover mutation evolution mechanisms including exchange, sliding, flipping and breakpoint selection.

[0099] In Table 3, the changes in the R1-R3 route reflect the population evolution after crossover and mutation operations.

[0100] Table 3 Updated UUV paths after crossover and mutation operations

[0101]

[0102] In each group of the current population, a fitness value selection mechanism is introduced to select the best parent individuals, and crossover and mutation operations are performed to improve the quality of offspring individuals.

[0103] Step 1.5: Determine whether the algorithm has terminated based on the convergence criteria. Researchers will select the maximum number of iterations as the stopping criterion. If the termination criteria have not been met, update the iteration counter and return to Step 1.1 to continue iterative evolution. Otherwise, find the individual with the highest fitness among all recorded optimal individuals and return it as the global optimal solution, ending the program.

[0104] Using elite MGA full space parallel search: The idea of ​​elitism is that the best individuals should be retained for the next generation. In MGA, there are two steps to ensure elitism:

[0105] Elite individuals in a group;

[0106] Storing historical elite individuals and reintroducing them into the population in later generations;

[0107] Elite MGA uses elitism to perform parallel search of the entire space while focusing on the high-performance part, which can improve efficiency while avoiding falling into local minima.

[0108] Step 2: UUV cluster MDC-MGA dynamic collaborative mission planning; apply the MDC-MGA mission planning method to the multi-UUV collaborative area search and detection framework, such as Figure 6 shown.

[0109] Step 2.1: Divide the search area into multiple sub-areas based on the actual underwater environment. At the same time, set the key search area based on expert experience, which means that the distance between adjacent task points in the key area is smaller than that in other areas.

[0110] Step 2.2: Initialize the MGA module and publish the mission path to each UUV.

[0111] Step 2.3: Obstacle avoidance and collaborative target search. In this process, obstacle avoidance takes precedence over target search. When the obstacle avoidance process is activated, the UUV search process will be suspended until the obstacle avoidance process is completed.

[0112] Step 2.4: All UUVs will perform the search mission through different mission points according to the given path. However, there are two cases where the planned path will be re-updated, and the MGA module will be activated to re-issue the new mission path to each UUV separately:

[0113] Case 1: If the target found in one area is much higher than in other areas, the MGA module will re-issue the mission path to each UUV and assign a group of UUVs. In order to find the target in the area in the shortest time, the path of the group of UUVs will start from a position close to the area.

[0114] Case 2: If one of the UUVs has completed all mission points in its path and the remaining mission points ratio is higher than 15%.

[0115] Step 2.5: When the proportion of unexplored task points is determined to be less than 15%, the search is terminated.

[0116] Example 1

[0117] A UUV swarm dynamic mission planning method based on a multi-objective genetic algorithm includes:

[0118] The MGA collaborative region search method based on the multi-objective genetic algorithm calculates the fitness of each individual in the population by setting and decoding schemes, and records the best individual in the current population. Genetic operations are performed on each group of selected individuals. Various crossover, mutation, and crossover-mutation operations are performed on the genes of the selected chromosomes with certain crossover and mutation probabilities to generate new chromosomes. Finally, the individual with the highest fitness value is found among all the recorded best individuals and returned as the global optimal solution. Specifically:

[0119] (1) For the path planning of regional task points, sequential coding is used to complete chromosome representation;

[0120] (2) Chromosomes are selected and randomly sorted, and the chromosome with the highest fitness value is selected from each group. However, unlike directly screening the best individual with the highest fitness value from the contemporary population, this random grouping method also ensures population diversity;

[0121] The fitness function of the chromosome individual is:

[0122]

[0123] in, is the total length of collision-free paths of all K UUVs under the scheme corresponding to chromosome X; K·d task_point is the task point spacing; T j J is the time consumed by the UUV to reach the predicted allocation area from the current position and perform target search, which is the larger value of the UUV movement time and the turning time; max Take the maximum penalty result in history; α, β, λ are weight parameters, and the multi-objective problem is transformed into a single-objective problem by adjusting the weight parameters; J(X) is the constraint condition of the chromosome; T max is the historical maximum penalty time in the execution of the UUV mission; α, β, and λ are pre-set weight parameters. By selecting the values ​​of these weight parameters, the multi-objective problem can be transformed into a single-objective optimization, which not only retains the flexibility of multi-objective optimization but also facilitates the direct ranking and selection of individuals in the genetic operator.

[0124] The constraints of the chromosome are designed based on the performance parameters of the UUV cluster and the mission requirements. The constraints are as follows:

[0125] J=c x f x +c T T+c v,a (f v +f a )+c t f t +c s f s +c ξ ζ

[0126] Among them, c x 、c T 、c v,a 、c t 、c s 、c ξ is a weight parameter; ζ is a "slack" variable used to measure the degree to which the current solution exceeds the threshold on various hard constraints;

[0127] The relative position f between the UUV and the target sub-area x for:

[0128]

[0129] Among them, x (m) is the center position of the mth sub-region, q represents the current position of the UUV, and n is the number of sub-regions;

[0130] The time consumption T of the UUV from the current position to the predicted assigned area and performing target search is:

[0131]

[0132] Among them, {P j ,ξ j} is the position of the next movement point of UUV and heading angle leagth(P(·)) is the length of the collision-free path; v max and are the maximum speed and yaw rate respectively; the time consumption T is calculated by calculating the navigation time and the UUV turning time, and the larger value of the two times represents the total time loss in the execution of the task;

[0133] Use the set threshold size to constrain the control penalty:

[0134]

[0135] The UUV kinematic constraints are:

[0136]

[0137] Among them, f v 、f a These are the velocity constraints and acceleration kinematic constraints added respectively; is the speed of the UUV at the next moving point; is the acceleration of the next moving point of the UUV; is the angular velocity of the next moving point of the UUV; is the angular acceleration of the next moving point of the UUV; a max is the maximum acceleration; by adding the velocity and acceleration kinematic constraints f v 、f a , ensuring that the motion state of the UUV during the actual search process is within a reasonable range, and preventing the occurrence of out-of-limit and unfeasible motion.

[0138] Communication distance constraint f for multiple UUVs in the process of regional collaborative search t for:

[0139]

[0140] Among them, x c,j,μ and x c,i,μ are the current positions of the j-th UUV and the i-th UUV respectively;

[0141] The endurance constraint of UUV during mission execution s for:

[0142]

[0143] Among them, S maxSmax is the maximum range of the UUV. c,i,μ Scur is the current distance traveled by the UUV.

[0144] (3) In each group of the current population, a fitness value selection mechanism is introduced to select the optimal parent individual, and crossover and mutation operations are performed according to the crossover probability and mutation probability to improve the quality of the offspring individuals. In addition, due to the use of sequential coding, the legality of the mutated chromosomes can be further ensured;

[0145] For each group of optimal individuals, the crossover probability of the optimal individual is calculated according to the difference between the individual fitness and the optimal / average fitness of the population and the mutation probability

[0146]

[0147] where Fk is the fitness value of the kth individual in the current generation; Fmax is the maximum fitness of all individuals in the current generation, Fmean is the arithmetic mean of the fitness of all individuals in the current generation; Pmin and Pmax are the lower and upper limits of the adaptive probability; k1 and k2 are two constant parameters representing the weights in the task, respectively. k best avg c,min m,min c m

[0148] (5) Until the maximum number of iterations, the optimal individual with the highest fitness value in each group is obtained, and the task planning path result is output.

[0149] According to the actual underwater environment, the search area is divided into multiple task sub-regions, and the dynamic task planning of the UUV cluster is completed according to the target distribution of the sub-regions and the task load of each UUV. Specifically:

[0150] (1) According to the actual underwater environment, the search area is divided into n sub-regions. At the same time, according to the expert experience, the key search area is set, that is, the distance between adjacent task points in the key area is less than that in other areas.

[0151] (2) The UUV executes the partition search strategy, and assigns the obtained task planning path to each UUV node;

[0152] ​​​​​​​(3) During the target search and obstacle avoidance phase, the core strategy is to prioritize the avoidance process. Once the system detects a collision risk or an unknown obstacle in the waterway, the UUV will immediately interrupt the current search mission and enter the obstacle recognition and avoidance mode, accurately avoiding obstacles through real-time path replanning and motion control. After the avoidance is completed and the navigation safety status is restored, the system automatically switches back to the collaborative target search module to continue the coverage search and data collection. This design not only ensures the absolute priority of navigation safety, but also avoids functional conflicts between the avoidance and search modules, thereby achieving efficient and continuous underwater autonomous operations.

[0153] (4) Each UUV will perform the search mission by crossing the mission points in turn along the predefined route. When the system identifies that the number of targets in a certain sub-area far exceeds that in other sub-areas, the UUV cluster will synchronize and communicate with each other and mark the key area. At the same time, the system will group the first UUV that detects the key area with the other UUV that is currently closest to the key area into a collaborative search team, and re-plan the mission based on the current real-time status of each UUV to achieve rapid coverage of the key area and accurate search of targets. In addition, during the regional search process, if any UUV completes the assigned route and more than 15% of the mission points in the area are still not executed, the optimized path will be re-issued to each UUV, and the sub-area with the highest probability of residual targets will be searched first, so as to achieve precise focus on the key area and efficient and comprehensive search.

[0154] When the UUV executes the partition search strategy, it first calculates the search cost of each task sub-region and determines whether the sub-region with the minimum search cost is preset as locked, that is:

[0155]

[0156] V i =min{f x ,LockArea(Sub n um)=0}

[0157] Among them, LockArea(Sub n um) is the sub-area Sub n The lock flag of um, if the flag is equal to 0, it means that the sub-area has not been locked as a mission area by any UUV; otherwise, the sub-area has been locked by a UUV. i Indicates sub-area Sub n um, the sub-area that is closest to the current UUV node and is not locked is given priority as the search task sub-area that the UUV needs to perform.

[0158] After the task sub-area is assigned, the UUV number to which the task belongs must be recorded to prevent the task sub-area from being repeatedly assigned to multiple UUVs, which in turn causes the UUV cluster to repeatedly search the sub-area, namely:

[0159] Task(Sub n um)=i,i∈{1,2,...,m}

[0160] The above formula represents the sub-area Sub n um is assigned to the UUV numbered i. If the sub-area pre-locked by the current UUV has not been locked in advance by other UUV nodes, the sub-area is directly identified as the mission sub-area that the current UUV will execute and target search is performed. Otherwise, other non-locked mission sub-areas are reconsidered.

[0161] In addition, most traditional optimization methods are based on global path optimization designed offline in unknown environments, and the constraints considered are relatively simple. Therefore, this study considers the various key constraints that exist when the UUV cluster performs regional search tasks. Through dynamic local optimization of the constraints, it is finally guaranteed that the optimal one-step solution for the current sub-region task allocation can be obtained in real time. The constraints are:

[0162] J=c x f x +c T T+c v,a (f v +f a )+c t f t +c s f s +c ξ ζ

[0163] Among them, c x 、c T 、c v,a 、c t 、c s 、c ξ is a weight parameter. ζ is a "slack" variable used to measure the degree to which the current solution exceeds the threshold for various hard constraints. When certain sub-constraints are violated, ζ accumulates the violations, assigns weights, and then adds them to the overall constraint function. This strictly suppresses infeasible solutions during the optimization process, ensuring that the ultimately selected path is both efficient and meets all safety and feasibility requirements.

[0164] Example 2

[0165] Simulation conditions and contents:

[0166] The research simulation environment is a sea area of ​​a certain depth, with an area of ​​2500m×2500m (see Figure 7 First, the entire search and detection mission area is decomposed into 36 first-level mission areas, and each first-level mission area is decomposed into 4 second-level sub-areas, where the mission point spacing d of the second-level sub-area is 1 / 4 of the first-level sub-area. break_point = 207.5m. Secondly, under limited communication conditions, static targets (the number of static targets was set to 40 in advance) were randomly distributed. Black rectangles represent obstacles, the blue circle in the middle represents the bottom of the offshore island port, and a key area is designated as a blue dashed rectangle. Finally, as shown in Table 4, the UUV starts from its initial position and performs target search and detection tasks, with a specified runtime of T = 2500 (steps).

[0167] Table 4 UUV initial position

[0168]

[0169] Obstacle avoidance was not the focus of this study; it was only used to verify the rationality of the mission planning module and to prevent collisions during the UUV swarm's search and detection process, while also avoiding entering enemy harbor areas. Therefore, only one obstacle was set in the simulation environment. Furthermore, the MGA parameters were set as follows: the population size K was set to 240, and the maximum number of iterations Iter was set to 10,000. Based on expert experience, the total parameters used in the MGA, elite MGA, and elite MDC-MGA are summarized in Table 5.

[0170] Table 5 Parameters of MGA, Elite MGA and Elite MDC-MGA

[0171]

[0172] According to (Genetic algorithm with adapted crossover operators for multiple traveling salesmen problem with visiting constraints. IEEE International Conference on Systems, Man, and Cybernetics (SMC). IEEE, 2022, pp.3033-3039), genetic algorithm with three improved crossover operators is adopted to solve the multiple traveling salesman problem (MTSP). As mentioned in the experiment, the partial map crossover (PMX) operator performs best. The core of PMX is to exchange partial segments of two parent chromosomes, unlike the method mentioned in the first section, PMX can handle TSP with different starting points. It utilizes the mapping on the original parent chromosomes to correct the remaining part of the child chromosomes, ensuring that the generated child is feasible. However, PMX is not suitable for multi-UUV search tasks. Because in PMX, the starting point and the ending point of each individual are the same. PMX-GA algorithm is inspired by the PMX operator, which adds the mutation operator mentioned in step 3 on the basis of the PMX operator. Among them, the crossover probability P c is set to 0.5, the mutation probability P m is set to 0.1. Therefore, the PMX-GA algorithm can solve the cooperative search problem with different starting points and ending points.

[0173] Five algorithms are introduced in the task planning experiment to verify their performance, as shown in Table 6. Among them, the performance of the proposed MGA and elite MGA is verified by using the clustering algorithm (GACA), PMX-GA, and the improved two-stage wolf search algorithm 2 (MTWPS2).

[0174] Table 6 Comparison of algorithm results

[0175]

[0176] Table 6 shows the comparison results of GACA, PMX-GA, MTWPS2, MGA, and elite MGA in 30 rounds of simulation. The "average", "best" and "worst" in the table represent the average, minimum and maximum of the total search and detection path obtained by the related algorithm for all UUV individuals. The best result in this column is shown in bold text.

[0177] As can be seen, in the four-UUV collaborative search task, the performance of the elite MGA is almost identical to that of the MGA, but the number of iterations of the elite MGA is one-third of that of the MGA. For the six- and eight-UUV collaborative search tasks, GACA and PMX-GA appear to have the fewest iterations, respectively, but considering the average distance, it is clear that they are local optima. In the first and second experiments, the elite MGA has a larger search space and achieves better performance than the MGA. Although the MGA achieves better results in terms of average distance in the third experiment, this is due to the introduction of historical elite chromosomes by the elite MGA. Furthermore, due to the increased number of UUVs, the larger search space negatively affects the convergence of the elite MGA, resulting in an increase in the average number of iterations. Furthermore, the proposed MGA and elite MGA also have certain advantages in terms of the time required to achieve over 90% area coverage.

[0178] The overall simulation results of the MDC-MGA task planning method are as follows: Figure 8 As shown in Figure 2. The purple Δ represents a static target. The results show that there are two mission planning stages in the area search and detection process. Figure 9 -(a) and Figure 10 -(a) The mission points are generated by referring to the forward-looking sonar model. Figure 9 -(b) and Figure 10 In (b), the lines represent the path of each UUV. The small circle on each line represents the current position of each UUV, which is also the starting point of the search process. Figure 9 -(c) and Figure 10 -(c) Verifies the effectiveness of the proposed algorithm. Since the probability of UUV4 marked with the yellow path finding the target in the key area is higher than the average level of other UUVs, the two UUVs closest to each other (marked with the blue and red paths) are combined to form a search team, and the UUV marked with the green path continues to perform the current task. This not only shortens the total search time, but also saves energy consumption. When the search team arrives at the target, the target will be detected. Figure 8 -(a) The second phase of mission planning begins when the focus area is shown in the upper right corner. Figure 6 The elite MGA module mentioned in the report will issue a new single mission, and at the same time, the UUVs in the search group will set off from the key areas. Figure 10 -(a) As shown in the upper right corner, the density of task points in key areas is higher than the average level.

[0179] Compared with the first stage of mission planning (i.e. Figure 11The MDC has an advantage of finding the target in the key area immediately. In the first stage, the performance of the MTWPS2 is almost the same as that of the elite MGA and the elite MDC-MGA, especially before the 180th second. The slope of the purple curve increases after the marked point, which proves the advantage of the MDC task planning module. All the cooperative search algorithms are based on the path planning of the 3-DOF UUV model and the forward-looking sonar model, so the potential danger of entering the enemy port area can be relieved in time. Before the task is executed, the position of the static target is unknown, and the total search time may not fully reflect the performance of the algorithm. From the perspective of the time axis, the elite MGA and the elite MDC-MGA perform better at most time points, which to a large extent proves the effectiveness of the algorithm.

[0180] In summary, the present application proposes a UUV cluster dynamic task planning method based on a multi-objective genetic algorithm, aiming to cope with the problems of limited number of UUVs in a multi-UUV regional detection environment, communication range and energy consumption constraints of the UUV cluster, and the difficulty of adjusting task decisions according to the real-time position and working state of the UUV cluster for the existing algorithm. The proposed MGA cooperative regional search method is integrated into the MGA module, the coding and decoding scheme is set, the fitness of each individual in the population is calculated, and the optimal individual in the current population is recorded. Genetic operation is performed on each selected individual, various crossover, mutation and crossover-mutation operations are performed on the genes of the selected chromosome with a certain crossover and mutation probability, thereby generating new chromosomes, finding the individual with the highest fitness among all recorded optimal individuals and returning it as the global optimal solution. Finally, according to the actual underwater environment, the search area is divided into multiple task sub-areas, and the dynamic task planning of the UUV cluster is completed according to the target distribution and the task load of each UUV. Through the MDC-MGA method, the UUV nodes are dynamically scheduled to reduce the adverse effects of target distribution non-uniformity and UUV cluster target discovery efficiency difference. In addition, in order to further verify the effectiveness of the proposed method, a simulated underwater environment is constructed to simulate the complex layout of the port area. The research evaluates the performance of key indicators such as the number of platforms, time cost, total path length and search efficiency in the regional search and detection task, fully verifying the efficiency of the proposed MGA regional search algorithm and MDC-MGA dynamic task planning method.

[0181] In particular, in some preferred embodiments of the present application, a computer device is also provided, which includes a memory and a processor, and a computer program stored in the memory, wherein the processor implements the steps of the UUV cluster dynamic task planning method based on the multi-objective genetic algorithm described in any of the above embodiments when executing the computer program.

[0182] In some other preferred embodiments of the present application, a computer readable storage medium is also provided, which stores computer programs / instructions, and the computer programs are executed by a processor to implement the steps of the UUV cluster dynamic task planning method based on the multi-objective genetic algorithm in any of the above embodiments.

[0183] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by a computer program instructing related hardware, and the computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, it can include the processes of the above-mentioned embodiments of the UUV cluster dynamic task planning method based on the multi-objective genetic algorithm, which will not be repeated here.

[0184] In the description of the present application, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example", or "some examples" means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the present application, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, the skilled in the art can combine and combine the different embodiments or examples described in the present application and the features of the different embodiments or examples without contradiction.

[0185] In addition, the terms "first", "second" are only for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include at least one of the features. In the description of the present application, the meaning of "N" is at least two, such as two, three, etc., unless otherwise specifically limited.

[0186] Any process or method descriptions in flow charts or otherwise described herein can be understood as representing code modules, segments, or portions of code which include one or more executable instructions for implementing the specified logical functions or steps, and the preferred embodiments of the present application also include additional implementation involving other processes or methods. Those skilled in the art should understand that the preferred embodiments of the present application can be implemented in other orders than those shown or discussed, including functions according to the functions involved, in a substantially simultaneous manner, or in reverse order, which should be understood by those skilled in the art.

[0187] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or N wires (electronic devices), a portable computer disk cartridge (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and programmable read-only memory (EPROM or flash memory), fiber optic devices, and a portable compact disc read-only memory (CDROM). Furthermore, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium and then editing, interpreting or processing it in another suitable manner if necessary, and then storing it in a computer memory.

[0188] It should be understood that various parts of the present invention can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiment, the N steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having logic gate circuits for implementing logic functions on data signals, an application-specific integrated circuit having suitable combinational logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0189] Those skilled in the art will understand that all or part of the steps in the method of the above embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0190] In addition, the functional units in the various embodiments of the present invention may be integrated into a single processing module, or each unit may exist physically separately, or two or more units may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or in the form of software functional modules. If the integrated modules are implemented in the form of software functional modules and sold or used as independent products, they may also be stored in a computer-readable storage medium.

[0191] The storage medium mentioned above may be a read-only memory, a magnetic disk, or an optical disk, etc. Although the embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and are not to be construed as limiting the present invention. Persons skilled in the art may make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present invention.

Claims

1. A UUV swarm dynamic mission planning method based on a multi-objective genetic algorithm, characterized in that: include: Path planning is performed for regional task points based on a multi-objective genetic algorithm; the multi-objective genetic algorithm outputs the optimal individual with the highest fitness value after iteration to obtain the task planning path; A UUV cluster dynamic task planning method is used to search and detect the area, and the search area is divided into multiple task sub-areas. The dynamic task planning of the UUV cluster is completed according to the target distribution in the sub-area and the conditions of each UUV. The UUV cluster dynamic task planning method is subject to local locking, kinematics, communication distance, endurance, total path length, and time consumption constraints when allocating task planning paths.

2. A UUV cluster dynamic mission planning method based on a multi-objective genetic algorithm according to claim 1, characterized in that: The multi-objective genetic algorithm is used for path planning. The specific process is as follows: Step 1.1: Encode the path planning of the regional task points in sequence to generate an initial population containing K chromosomes; X k =(x1,x2,…,x z )1≤k≤K Where X is the total search path of each UUV; K is the number of chromosomes; z is the length of each chromosome, which also depends on the number of total mission points; Step 1.2: Calculate the fitness value of each chromosome individual in each group of the population, and select the chromosome with the highest fitness value from each group as the optimal individual; Step 1.3: Perform color crossover and chromosome mutation on each group of optimal individuals according to the crossover probability and mutation probability to generate new chromosomes, and then update the population to obtain the next generation population. Return to step 1.1; Step 1.4: Repeat steps 1.1 to 1.3 above until the maximum number of iterations is reached, obtain the optimal individual with the highest fitness value in each group, and output the task planning path result.

3. The method for dynamic mission planning of UUV clusters based on a multi-objective genetic algorithm according to claim 2, characterized in that: The fitness function of the chromosome individual is: in, is the total length of collision-free paths of all K UUVs under the scheme corresponding to chromosome X; K·d task_point is the task point spacing; T j J is the time consumed by the UUV to reach the predicted allocation area from the current position and perform target search, which is the larger value of the UUV movement time and the turning time; max Take the maximum penalty result in history; α, β, λ are weight parameters, and the multi-objective problem is transformed into a single-objective problem by adjusting the weight parameters; J(X) is the constraint condition of the chromosome; T max The maximum penalty time in the history of UUV mission execution; The constraints of the chromosome are designed based on the performance parameters of the UUV cluster and the mission requirements. The constraints are as follows: J=c x f x +c T T+c v,a (f v +f a )+c t f t +c s f s +c ξ ζ Among them, c x 、c T 、c v,a 、c t 、c s 、c ξ is a weight parameter; ζ is a "slack" variable used to measure the degree to which the current solution exceeds the threshold on various hard constraints; The relative position f between the UUV and the target sub-area x for: Among them, x (m) is the center position of the mth sub-region, q represents the current position of the UUV, and n is the number of sub-regions; The time consumption T of the UUV from the current position to the predicted assigned area and performing target search is: Among them, {P j ,ξ j } is the position of the next movement point of UUV and heading angle leagth(P(·)) is the length of the collision-free path; v max and are the maximum speed and yaw rate respectively; the time consumption T is calculated by calculating the navigation time and the UUV turning time, and the larger value of the two times represents the total time loss in the execution of the task; The control penalty is: The UUV kinematic constraints are: Among them, f v 、f a These are the velocity constraints and acceleration kinematic constraints added respectively; is the speed of the UUV at the next moving point; is the acceleration of the next moving point of the UUV; is the angular velocity of the next moving point of the UUV; is the angular acceleration of the next moving point of the UUV; a max is the maximum acceleration; Communication distance constraint f for multiple UUVs in the process of regional collaborative search t for: Among them, x c,j,μ and x c,i,μ are the current positions of the j-th UUV and the i-th UUV respectively; The endurance constraint of UUV during mission execution s for: Among them, S max is the maximum range of UUV, S c,i,μ The current distance traveled by the UUV.

4. The method for dynamic mission planning of UUV clusters based on a multi-objective genetic algorithm according to claim 2, characterized in that: In step 1.3, for each group of optimal individuals, the crossover probability of the optimal individual is calculated based on the difference between the individual fitness and the optimal / average fitness of the group. and mutation probability Among them, F k is the fitness value of the i-th individual in the current generation; F best is the maximum fitness of all individuals in the current generation, F avg is the arithmetic mean of all individual fitness in the current generation; P c,min With P m,min is the lower limit of the adaptive probability; k c and k m They represent two constant parameters, which are used to adjust the weights in the task.

5. The method for dynamic mission planning of UUV clusters based on a multi-objective genetic algorithm according to claim 1, characterized in that: The UUV cluster dynamic task planning method is used to search and detect the area. The specific process is as follows: Step 2.1: Divide the search area into multiple sub-areas; Step 2.2: The UUV executes the partition search strategy and assigns the obtained mission planning path to each UUV node; Step 2.3: Execute the sub-region obstacle avoidance process; If the UUV finds an obstacle, it will re-update the mission planning path based on the current UUV real-time position; If the UUV does not find an obstacle, it executes the sub-area target search process; Step 2.4: Execute the sub-area target search process; the UUV performs the search task along the current mission planning path. When the number of mission point targets identified by the UUV in a sub-area exceeds that of other sub-areas or the distance between adjacent mission points is smaller than the distance between adjacent mission points in other areas, the UUV cluster will synchronously complete mutual communication and mark the key area. The UUV closest to the key area will also enter the key area. Multiple UUVs in the key area will search for mission points according to the newly updated mission planning path. Step 2.5: After the UUV completes the search of the sub-area according to the mission planning path; if more than 15% of the mission points in the current sub-area have not been searched, the mission planning path is updated again and steps 2.3 to 2.4 are repeated until less than 15% of the mission points have not been searched, and the target search process ends.

6. The method for dynamic mission planning of UUV clusters based on a multi-objective genetic algorithm according to claim 5, characterized in that: When the UUV executes the partition search strategy, it first calculates the search cost of each task sub-region and determines whether the sub-region with the minimum search cost is preset as locked, that is: V i =min{f x ,LockArea(Sub n (um)=0} Among them, LockArea(Sub n um) is the sub-area Sub n um's locking flag. If the flag is equal to 0, it means that the sub-area has not been locked as a mission area by any UUV; otherwise, the sub-area has been locked by a UUV; V i Represents the cost of the sub-region Subnum, considering the sub-region closest to the current UUV node and not locked as the search task sub-region that the UUV needs to perform.

7. The method for dynamic mission planning of UUV clusters based on a multi-objective genetic algorithm according to claim 6, characterized in that: When the UUV executes the partition search strategy, there are key constraints. Through dynamic local optimization of the constraints, real-time task allocation for the current sub-area is achieved; The constraints are: J=c x f x +c T T+c v,a (f v +f a )+c t f t +c s f s +c ξ ζ 8. A computer device / apparatus / system comprising a memory, a processor, and a computer program stored in the memory, characterized in that: The processor executes the computer program to implement the steps of the method according to any one of claims 1 to 7.

9. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer program product comprising a computer program / instructions, characterized in that: When the computer program / instructions are executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images