Cloud multitask planning method for ocean satellite internet of things

Abstract

The application relates to a cloud multi-task planning method for a marine satellite Internet of Things, which comprises the following steps: acquiring a total cost function for minimizing the total task cost of multiple marine tasks, and acquiring a constraint condition set corresponding to the total cost function; the constraint condition set comprises a communication time delay constraint and a task allocation constraint; a task type with an undetermined decision is selected as a row branch, the total cost function is solved based on the constraint condition set, and a marine task decision is obtained; the marine task decision comprises a decision variable of allocating the marine task to a corresponding edge subnet; and corresponding marine tasks are sent to the corresponding edge subnet through the marine Internet of Things based on the marine task decision. The method can flexibly cope with emergencies and quickly execute task planning.

Description

Technical Field

[0001] This application relates to the field of marine mission planning technology, and in particular to a cloud-based multi-mission planning method for marine satellite Internet of Things. Background Technology

[0002] With the rapid development of intelligent and unmanned maritime operations, the maritime satellite Internet of Things (IoT) has become a crucial information infrastructure supporting marine environmental perception, maritime situation monitoring, emergency response, and long-range operations. In the vast spatiotemporal scale of the marine environment, the unified scheduling and collaborative operation of widely distributed heterogeneous terminals such as unmanned vessels, unmanned surface vessels, drones, communication buoys, and various detection buoys, relying on satellite communication networks, is a key technological direction for improving the overall efficiency, response speed, and system resilience of the maritime operation system. Among these, cloud-based multi-task planning and resource allocation capabilities for complex and ever-changing marine environments are a vital foundation for achieving intelligent operation and efficient service support of the maritime satellite IoT.

[0003] However, the platform types, mission attributes, and service quality requirements of maritime satellite IoT are highly diverse, including gigabit-level continuous buoy detection data, megabit-level unmanned surface vessel (USV) reconnaissance data, and UAV remote sensing data with strict time constraints. In existing application models, maritime multi-missions are usually conducted in subnets or formations, with shore-based systems uniformly pre-assigning tasks before execution. Various unmanned platforms then perform predetermined remote sensing, communication, reconnaissance, or detection tasks upon arriving at the designated operational area. Traditional mission planning methods typically assume that the mission type, objectives, and execution sequence are largely determined during the planning phase, employing pre-defined operational procedures. However, traditional mission planning methods suffer from an inability to flexibly respond to unforeseen events. Summary of the Invention

[0004] Therefore, it is necessary to provide a cloud-based multi-task planning method for marine satellite IoT that can flexibly respond to emergencies and address the aforementioned technical problems.

[0005] Firstly, this application provides a cloud-based multi-task planning method for marine satellite Internet of Things, applied in the cloud, the method comprising: Obtain the total cost function that minimizes the total cost of various maritime missions, and obtain the set of constraints corresponding to the total cost function; the maritime missions include early warning missions, strike missions, and communication missions; the set of constraints includes communication delay constraints and mission allocation constraints. Select the task type with an uncertain decision as the row branch, solve the total cost function, and generate nodes that meet the constraint set; the nodes include partial solutions and the lower bound values ​​corresponding to the partial solutions; For the node, return to the step of selecting the task type with the undetermined decision as the row branch, solve the total cost function, and generate a node that meets the set of constraints, until all task types have been decided, and obtain the complete solution and the total cost function value corresponding to the complete solution; Based on the complete solution and the total cost function value corresponding to the complete solution, the complete solution with the minimum total cost function value is determined to obtain the ocean mission decision; the ocean mission decision includes the decision variable of assigning the ocean mission to the corresponding edge subnet; Based on the aforementioned marine mission decision, the corresponding marine mission is sent to the corresponding edge subnet via the marine Internet of Things.

[0006] Secondly, this application also provides a cloud-based multi-task planning device for marine satellite Internet of Things, comprising: The acquisition module is used to acquire the total cost function that minimizes the total task cost of various marine missions, and to acquire the set of constraints corresponding to the total cost function; the marine missions include early warning missions, strike missions, and communication missions; the set of constraints includes communication delay constraints and task allocation constraints. The solution module is used to select the task type with an uncertain decision as the row branch, solve the total cost function, and generate nodes that meet the constraint set; the nodes include partial solutions and the lower bound values ​​corresponding to the partial solutions; For the node, return to the step of selecting the task type with the undetermined decision as the row branch, solve the total cost function, and generate a node that meets the set of constraints, until all task types have been decided, and obtain the complete solution and the total cost function value corresponding to the complete solution; Based on the complete solution and the total cost function value corresponding to the complete solution, the complete solution with the minimum total cost function value is determined to obtain the ocean mission decision; the ocean mission decision includes the decision variable of assigning the ocean mission to the corresponding edge subnet; The sending module is used to send the corresponding marine mission to the corresponding edge subnet through the marine Internet of Things based on the marine mission decision.

[0007] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps described in the cloud-based multi-task planning method for the Internet of Things for marine satellites.

[0008] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps described in the cloud-based multi-task planning method for ocean satellite Internet of Things.

[0009] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps described in the cloud-based multi-task planning method for the Internet of Things for marine satellites.

[0010] The aforementioned cloud-based multi-task planning method for marine satellite IoT obtains the total cost function that minimizes the total cost of various marine tasks and the corresponding set of constraints. It selects the task type with undetermined decisions as the row branch, solves the total cost function to generate partial solutions that satisfy the constraint set and their corresponding lower bounds, until all tasks are decided, obtaining a complete solution and the corresponding total cost function value. From this, it determines the complete solution with the smallest total cost function value, thus obtaining the marine task decision. By determining the decision variables row by row, the multi-task joint scheduling problem is decomposed into several sub-problems decided sequentially according to task type. Each search layer only needs to enumerate within the candidate edge subnet of the current row, without traversing all combinations, quickly converging to the complete solution with the smallest total cost. This yields the decision variable for assigning which marine task to which edge subnet, enabling the edge subnet to execute the corresponding marine task. This method is suitable for scenarios with large task changes and frequent responses to emergencies, allowing for flexible handling of unexpected events. Furthermore, it quickly solves the total cost function, enabling rapid task planning and execution of marine tasks. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 This is an application environment diagram of a cloud-based multi-task planning method for marine satellite IoT in one embodiment; Figure 2 This is an application environment diagram for a cloud-based multi-task planning method for marine satellite IoT in another embodiment; Figure 3 This is a schematic diagram of task allocation in one embodiment; Figure 4 This is a flowchart illustrating a cloud-based multi-task planning method for marine satellite IoT in one embodiment; Figure 5 This is a schematic diagram of edge subnet information backhaul in one embodiment; Figure 6 This is a schematic diagram illustrating the simulation results of the total cost of task planning in one embodiment; Figure 7This is a schematic diagram comparing the optimized time of the traditional task planning mode in one embodiment and the task planning mode in this embodiment; Figure 8 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0013] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0014] The cloud-based multi-task planning method for marine satellite IoT provided in this application can be applied to, for example... Figure 1 The application environment shown. Figure 1 This diagram illustrates the application environment of a cloud-based multi-task planning method for a marine satellite Internet of Things (IoT) in one embodiment. The cloud platform 102 communicates with multiple edge subnets 104 via the marine IoT. The cloud platform 102 can be a standalone physical server, a server cluster or distributed system comprised of multiple physical servers, or a cloud server providing cloud computing services. A data storage system stores the data that the cloud platform 102 needs to process. The data storage system can be integrated onto the cloud platform 102 or hosted on other network servers. Each operational team deployed in the target sea area can be considered an edge subnet 104.

[0015] Figure 2 illustrates the application environment of a cloud-based multi-task planning method for marine satellite IoT in another embodiment. This embodiment uses a maritime early warning and joint strike scenario as an example to illustrate the problem of cloud-based ad hoc task allocation during emergencies. Each operational group deployed in the target sea area—namely, the communication relay subnet, early warning subnet 1, and early warning subnet 2—can be considered an edge subnet. Each edge subnet has a periodic task chain, performing routine inspection work. When an emergency occurs (e.g., a conflict target enters the target sea area), the cloud will initiate an ad hoc task scheduling process, disrupting the original periodic task coordination. For example... Figure 3 The diagram illustrates task allocation in one embodiment. Edge subnets A, B, and C perform routine tasks from time t1 to t6. Meanwhile, the cloud will urgently insert sudden emergency tasks (early warning tasks or strike tasks) into the task queue, and simultaneously distribute the task list to the edge subnets via heterogeneous satellites.

[0016] Unlike traditional task planning methods, the maritime task planning method proposed in this embodiment uses the edge subnet as the smallest scheduling unit and distributes the scheduling results to the master station nodes of the edge subnet. Traditional task planning methods mainly perform task scheduling for all terminal nodes in the global scope. Although the scheduling granularity of traditional methods is relatively fine, if traditional methods are simply applied to the maritime scenario of this embodiment, task scheduling needs to be performed for hundreds of maritime terminal nodes, which will significantly increase the scheduling time.

[0017] When contingent events occur, how to efficiently plan maritime operations and effectively respond to emergencies with minimal scheduling costs becomes an urgent problem to be solved. This embodiment presents the cost calculation for mapping edge subnets to operational tasks, and based on the cost calculation results, designs a multi-task planning model to effectively respond to maritime emergencies with minimal cost.

[0018] like Figure 4 The diagram shown is a flowchart illustrating a cloud-based multi-task planning method for marine satellite IoT in one embodiment, including: Step 402: Obtain the total cost function that minimizes the total cost of various ocean missions, and obtain the set of constraints corresponding to the total cost function; ocean missions include early warning missions, strike missions and communication missions; the set of constraints includes communication delay constraints and mission allocation constraints.

[0019] The total cost function is the desired objective form expressed in terms of design variables; therefore, the total cost function is a function of the design variables.

[0020] Maritime missions primarily encompass three types: early warning missions, strike missions, and communication missions. Other types of missions may also be included, but are not limited here. Early warning missions involve using detection, surveillance, and identification methods within the maritime area to proactively detect, track, and warn of maritime targets and threats. Early warning missions provide advance information for subsequent decision-making, strikes, and evasion, serving as the perception front end of the maritime mission system.

[0021] Strike missions are operations conducted in the marine environment to carry out fire strikes, damage, suppression, and blockade of targets; they are the execution end of the marine mission system. Communication missions are operations conducted in the marine environment to realize information transmission, data exchange, command issuance, and network relay; they are the information hub of the marine mission system.

[0022] Specifically, the total cost function can include the sum of functions that minimize the cost of early warning missions, the cost of strike missions, and the cost of communication missions. The constraint set includes communication delay constraints and task allocation constraints. Communication delay constraints require each edge subnet to meet communication delay conditions, while task allocation constraints include requirements such as early warning and strike missions belonging to different task subnets. When a target is detected entering the target sea area, the cloud begins mission planning, obtaining the total cost function that minimizes the total cost of multiple maritime missions, as well as the corresponding constraint set.

[0023] Step 404: Select the task type with uncertain decision as the row branch, solve the total cost function, and generate nodes that meet the constraint set; the nodes include partial solutions and the lower bound values ​​corresponding to the partial solutions.

[0024] Step 406: For a node, return to the step of selecting the task type with an uncertain decision as the row branch, solving the total cost function, and generating a node that meets the constraint set, until all task types have been decided, and obtain the complete solution and the total cost function value corresponding to the complete solution.

[0025] Step 408: Based on the complete solution and the total cost function value corresponding to the complete solution, determine the complete solution with the smallest total cost function value to obtain the ocean mission decision.

[0026] The basic idea of ​​branch and bound is to search the space of all feasible solutions (a finite number) to a constrained optimization problem. Row decomposition-based branch and bound algorithms reduce the computational complexity of subproblems and improve the overall efficiency of the algorithm by decomposing the constraint system of the original problem into rows (such as constraints). These algorithms use task types as row branching units, with each row corresponding to the allocation decision of a task type on the candidate edge subnet set, determining the decision variables row by row.

[0027] Row branching is the core branching rule of the branch and bound algorithm based on row decomposition. It is used to decompose subproblems based on the row constraints of the original problem.

[0028] A partial solution is a solution obtained during the branch and bound process where some decision variables of the original problem have been assigned definite feasible values, while the remaining decision variables have not yet been assigned values. It is an intermediate solution generated during the branching process. For example, if there are three task types, and one type is assigned a value while the other two are not, or two types are assigned a value while the third is not, then the result is a partial solution.

[0029] The complete solution is a feasible solution to the original problem where all decision variables are assigned definite feasible values ​​and all constraints of the original problem are satisfied. It is the feasible solution of the total cost function and is used to represent maritime mission decisions. The total cost function value of the complete solution is the value obtained by substituting the complete solution into the total cost function.

[0030] The number of nodes in the constraint set generated in the cloud each time is unlimited. Each node includes candidate partial solutions and their corresponding lower bounds. The lower bound for a candidate partial solution is a conservative lower bound estimate of the total cost of all complete solutions in the branch represented by that node. The lower bound for each node is determined independently and changes with the stage of the determined solution. For example, the lower bound for the k1 partial solution = k1 cost + k2k3 relaxation estimate; the lower bound for the k1k2 partial solution = k1k2 cost + k3 relaxation estimate; the complete solution has no lower bound, and the exact cost, i.e., the total cost function value, is calculated.

[0031] Specifically, the cloud-based branch and bound algorithm, based on row decomposition, solves the total cost function based on a set of constraints to obtain ocean mission decisions that satisfy this set of constraints. Ocean mission decisions include decision variables for assigning ocean missions to corresponding edge subnets. For example, matrix X is the final decision matrix for cloud-based multi-task planning, obtained by the branch and bound algorithm. Its matrix elements are binary variables: if an element is 1, it means the corresponding edge subnet executes the corresponding task at the current time; if it is 0, it means the task is not executed. The entire matrix depicts the mapping relationship between task types and edge subnets, and is a unified expression of the cloud-based task scheduling results. Since there are multiple types of ocean missions, a branch and bound algorithm based on row decomposition is used to solve the problem, with each task type serving as a row branch for a complete solution. Taking task types k1, k2, and k3 as examples, in the initial processing, no decision is made for any task type. Therefore, the selected task type with an uncertain decision is k1. The solution will then yield multiple nodes that satisfy the constraint set. Among all nodes, k1 is determined, and the resulting partial solution and corresponding lower bound are those for k1. The remaining task types with uncertain decisions are k2 and k3. For the node corresponding to k1, in the next iteration, k2 can be selected for solution. At this point, both k1 and k2 are determined, and the resulting partial solution and corresponding lower bound are those for both k1 and k2. The remaining task type with an uncertain decision is k3. When k3 is also determined, all task decisions are completed, resulting in a complete solution where k1, k2, and k3 are all determined, along with the corresponding total cost function value. The complete solution with the minimum total cost function value can then be determined, thus obtaining the ocean mission decision.

[0032] Step 410: Based on the marine mission decision, send the corresponding marine mission to the corresponding edge subnet through the marine Internet of Things.

[0033] Specifically, marine IoT is the application of IoT technology underwater. After obtaining marine mission decisions, the cloud can send corresponding marine missions to the corresponding edge subnets through marine IoT based on the decision variables.

[0034] In this embodiment, the total cost function that minimizes the total cost of multiple marine tasks and the corresponding set of constraints are obtained. A branch and bound algorithm based on row decomposition is used to select the task type with an uncertain decision as the row branch. The total cost function is solved to generate partial solutions that meet the set of constraints and the corresponding lower bound values ​​until all tasks are decided, obtaining the complete solution and the corresponding total cost function value. The complete solution with the smallest total cost function value is then determined to obtain the marine task decision. By determining the decision variables row by row, the multi-task joint scheduling problem is decomposed into several sub-problems that are decided sequentially according to the task type. Each search layer only needs to enumerate within the candidate edge subnet of the current row, without traversing all combinations, and can quickly converge to the complete solution with the smallest total cost. This gives the decision variable of which marine task to assign to which edge subnet, so that the edge subnet executes the corresponding marine task. This is suitable for scenarios with large task changes and frequent need to deal with emergencies, and can flexibly respond to emergencies. It also allows for rapid solution of the total cost function, thereby quickly planning and executing marine tasks.

[0035] In one embodiment, the total cost function is determined based on the sum of the cost functions corresponding to the early warning mission, the strike mission, and the communication mission, respectively.

[0036] Specifically, the total cost function may include the sum of the cost functions of the early warning mission, the strike mission, and the communication mission. Optionally, the total cost function can be obtained by adding the products of the cost functions and weights corresponding to the early warning mission, the strike mission, and the communication mission, respectively.

[0037] In this embodiment, the cost functions of early warning mission, strike mission, and communication mission are important scenario functions in maritime missions. By decomposing them into the sum of the total cost functions of the three missions, the maritime missions can be accurately classified, enabling precise planning of maritime emergency missions.

[0038] In one embodiment, the cost function for the early warning task is determined based on the distance between the coordinates of the edge subnet and the coordinates of the detected target; the cost function for the strike task is used to determine whether the detected target has entered the target sea area attack range; and the cost function for the communication task is used to determine the sum of the transmission power of the edge subnet.

[0039] Specifically, when a conflict target enters the target sea area, the remote cloud needs to sequentially issue "early warning tasks," "communication tasks," and "strike tasks" to each edge subnet. Considering the set of edge subnets in a maritime satellite IoT scenario, it is as follows: That is: shared An edge subnet. Let the set It means all of the sea The two-dimensional coordinates of the master station node in each edge subnet, where express Time of the first Two-dimensional coordinate information of the master station node in each edge subnet. and Representing edge subnets respectively Location (corresponding longitude and latitude).

[0040] Command symbol Indicates remote cloud access A set of tasks that need to be planned at all times, among which , , These represent three sets of tasks: early warning, communication, and strike. The model is established. Among the multiple tasks, there is a temporal coupling relationship. For example, the early warning task responds first, the communication relay task follows the early warning task and precedes the strike task, and the strike task is triggered last. This temporal coupling relationship between the tasks will serve as a constraint on the model. (Remote cloud in...) At any time, on-the-spot planning is required. An urgent, edge-of-the-road mission, obviously Established.

[0041] Let matrix express At any given time, the mapping relationship between maritime missions and edge subnets can be expressed by the following matrix: The above matrix will be abbreviated as follows: Arbitrary edge subnets Only specific tasks can be performed, therefore the set of edge subnets is defined as follows: Mapping function between multiple task sets at any given time As shown below: Among them, the conclusion This is clearly true.

[0042] make express At time [time], the two-dimensional coordinate information of target Q. Because the target is in [location]... The coordinates of a given moment are dynamic and time-varying, and cannot be easily predicted. The remote cloud can only obtain the current coordinates through early warning tasks. The location information of the target at any given time and the location information of historical times are recorded as a set. At each new moment, the remote cloud needs to periodically plan the tasks for the next moment based on the target's real-time location information, and then complete the tasks. With edge subnet On-demand matching ensures that each edge subnet is in the latest state in the event of an emergency.

[0043] For early warning mission When the probe moves to different locations, the cloud will select the edge subnet closest to the target to perform the monitoring task. At any given time, the remote cloud needs to... Select from the edge subnets Each edge subnet performs real-time monitoring of the target (obviously there are...). (Established), this monitoring will continue throughout the entire time series. The monitoring will continue until the target leaves the target area or is sunk. It is worth noting that selecting different edge subnets for monitoring will result in varying costs.

[0044] The total cost function C corresponding to the early warning task 预警 It can be expressed as the following formula: Where x represents the decision variable, t represents time, k represents the task type, h represents the edge subnet, and H represents the entire set of edge subnets. Indicates the main site node The coordinate information, let express At any given moment, the target's two-dimensional coordinate information is detected. Here, H represents the entire set of candidate edge subnets. The physical meaning of the summation is: during the cloud planning phase, a unified cost evaluation is performed on all edge subnets with execution capabilities, and ultimately, constraints ensure that only a specified number of subnets are selected to execute the task. The above formula measures the total cost function of the early warning detection task. As can be seen from the formula, this total cost function is mainly related to the location of the detection subnet and the target location. Specifically, during the optimization selection process, the cloud will prioritize edge subnets closer to the target, thereby improving early warning and reconnaissance efficiency.

[0045] Targeting strike missions When the detection target enters the adjacent waters of the target area, and that area is within the sinkable range [x] min ,x max And [y] min ,y max [Time] (i.e., the coordinates of the detected target) After authorization via the cloud, the strike command can be triggered directly to complete the strike mission.

[0046] The total cost function C of the strike mission 打击 It can be expressed as the following formula: in, Let A be the characteristic function, that is: if event A is true, then Conversely, it is 0. Event This indicates whether the target has entered the attack range of the attacking platform, i.e., the target's coordinates. Does it belong to a set? Furthermore, c0 is a constant coefficient, representing the cost incurred after hitting the target.

[0047] The total cost function for the communication task is shown below: Among them, P h This represents the average power of the transmitted signal.

[0048] Based on the above analysis, what is the total cost function that minimizes the total cost of various ocean missions? It can be written as the following formula: in, See the formula above. , and All are vector parameters, representing the cost coefficients converted into vector factors for various types of tasks (and tasks of the same type but different priorities). The values ​​can be taken reasonably.

[0049] In this embodiment, the total cost function corresponding to the early warning task is determined based on the distance between the edge subnet coordinates and the detection target, which can prioritize the edge subnets closer to the target, thereby improving the early warning and reconnaissance efficiency; the total cost function of the strike task is used to determine whether the target has entered the attack range, which enables the edge subnets within the attack range to respond to the strike; the total cost function corresponding to the communication task is used to determine the sum of the power of the edge subnets, which can minimize the communication cost.

[0050] In one embodiment, the communication delay constraint includes a minimum communication delay constraint and a data return delay upper limit constraint; the constraint set also includes a time-coupling constraint and a binary variable constraint for the decision variable, and the time-coupling constraint includes executing the early warning task before executing the strike task.

[0051] Specifically, communication delay constraints are mainly used to constrain the total cost function of communication tasks. Temporal coupling constraints include tasks such as executing the early warning task first, then the communication task, and finally the strike task.

[0052] For communication tasks The remote cloud needs to collect data from all edge subnets performing communication tasks. Information processing results (of which, Communication tasks are a type of accompanying task (or support task). Their execution target is not a "task type," but rather the edge subnet that currently needs to transmit data back. Early warning and strike tasks are operational task types; at any given planning moment, each edge subnet can execute at most one of these operational tasks. Communication tasks, however, are support tasks (accompanying tasks), used to handle data transmission needs generated during operational tasks or routine inspections. Therefore, "an edge subnet executing a communication task" does not mean that it has selected a communication task at the operational level, but rather that the edge subnet currently has a data transmission need. This need may originate from: 1) Edge subnets currently performing emergency early warning missions; 2) Edge subnets that are currently executing strike missions and need to transmit mission status or results back; 3) Edge subnets that are not assigned emergency tasks, only perform routine inspection tasks, but still need to periodically transmit data back.

[0053] The three types of edge subnets mentioned above are mutually exclusive in terms of task types, but they can be superimposed at the level of communication requirements. Communication tasks only describe their information return behavior and the corresponding latency and resource constraints, without changing their task attributes.

[0054] At this point, the edge subnet mainly consists of edge subnets that perform emergency early warning tasks. And the edge subnets that perform routine inspection tasks (i.e., edge subnets that did not perform emergency early warning tasks), obviously have Established. Due to the time-sensitive nature of early warning information, the communication feedback interval for emergency early warning tasks shall not exceed [a certain value]. The feedback interval for routine inspection tasks shall not exceed Due to the urgency of the early warning mission, This is clearly true. Let Represents the edge subnet The last time information was transmitted back was when the symbol was displayed. Represents the edge subnet The following formula holds true for the next decision on when to send back information: In the above formula, the first expression represents the edge subnet (h∈H) that undertakes the emergency early warning task. d The time interval between the "next information feedback time" and the "previous information feedback time" cannot exceed the emergency threshold T1. The second formula means that the edge subnet (h∈H) undertaking routine inspection tasks... r The time interval between the "next information transmission time" and the "previous information transmission time" cannot exceed the conventional threshold T2.

[0055] The physical meaning of the above formula is as follows: Figure 5 As shown, Figure 5 This is a schematic diagram illustrating edge subnet information backhaul in one embodiment. These figures represent the timestamps of a single communication backhaul on the four edge subnets. In this diagram, edge subnets 1, 2, and 3 all meet the communication backhaul latency constraints, meaning the communication latency of edge subnets 1 and 2 (early warning tasks) is less than or equal to T1, the communication latency of edge subnet 3 is less than or equal to T2, and edge subnet 4 does not meet the latency constraint (greater than T2). The failure to meet the latency constraints mainly stems from: ① After a sudden event, some edge subnets need to simultaneously undertake both early warning and routine inspection tasks, leading to competition for communication resources. ② Satellite links are limited by visibility and bandwidth, making it impossible to guarantee that all subnets simultaneously meet the backhaul latency limit. Therefore, the model explicitly characterizes the "backhaulable" and "non-backhaulable" states through latency constraints to guide the cloud in prioritizing scheduling under resource-constrained conditions.

[0056] Therefore, the constraints on the total cost function of the communication task are as follows: in, The constraint is a minimum communication delay constraint. The constraint is the upper limit constraint on data return latency. Molecules in constraints This indicates the total amount of data to be transmitted.

[0057] denominator This represents the transmission rate calculated based on Shannon's formula.

[0058] The timing coupling constraints are as follows: Constraints and The constraint represents the temporal coupling constraint of multiple tasks, t1 f It is the time when the early warning mission is completed, t2 f It is the earliest start time of the communication task, t3 f It is the earliest end time of the communication task, t4 f This is the earliest start time for the strike mission. C4 requires the completion of the early warning mission before the communication mission can begin, and the strike mission can only begin after the communication mission is completed.

[0059] The constraint states that each task can be assigned to at most one edge subnet. Constraints represent decision variables It is a binary variable.

[0060] In this embodiment, by setting communication latency constraints to define the time boundary of task execution, the scheduling scheme is prevented from deviating from the hardware capabilities of the communication system, thus solving the speed problem of task planning; by setting temporal coupling constraints, the task order in maritime scenarios is applied, thus solving the order problem of task planning and avoiding resource conflicts and operational chaos; by setting decision variables as binary variables, multi-task coordination can be performed.

[0061] In one embodiment, after generating nodes that meet the set of constraints, the method further includes: Add a node to the set of active nodes, and select a node from the set of active nodes. Based on the complete solution and its corresponding total cost function value, the complete solution with the minimum total cost function value is determined to obtain the ocean mission decision, including: Update the global upper bound to the minimum total cost function value determined from the total cost function values ​​corresponding to the complete solution; Update the global optimal solution to the complete solution corresponding to the minimum total cost function value; When the lower bound of a node is less than the global upper bound, return to the step of selecting a node from the set of active nodes, until the set of active nodes is empty or the number of iterations is reached, to obtain the global optimal solution and thus obtain the ocean mission decision.

[0062] The active node set is used to store partial solutions and their corresponding lower bound values, i.e., nodes for which no complete solution has been obtained.

[0063] The global upper bound refers to the minimum value of the total cost function of all found complete solutions in the current solution stage of the branch and bound problem for minimizing the total cost function. It is the upper bound of the total cost function of the complete solution of the original problem. That is, the total cost function of the global optimal solution of the original problem must be less than or equal to this value.

[0064] The ultimate goal of the branch and bound algorithm is to find a globally optimal solution that satisfies all constraints of the original problem and makes the total cost function value globally optimal. The number of iterations is set as needed and stored in the cloud.

[0065] Specifically, the cloud selects a task type with an uncertain decision as a row branch, solves the total cost function, generates multiple nodes that meet the set of constraints, and adds these nodes to the activation node set Q. At this time, the activation node set Q stores nodes where k1 is determined, as well as nodes where both k1 and k2 are determined.

[0066] Since the nodes in the active node set Q have not yet been fully explored, a node is selected from Q for solution to obtain the complete solution and its corresponding total cost function value. The minimum total cost function value among these complete solution values ​​is then used as the global upper bound, i.e., the maximum total cost function value of the complete solution. The global optimal solution (i.e., the final ocean mission decision) is the complete solution corresponding to this minimum total cost function value. When the lower bound of a node is less than the global upper bound, it indicates that continuing calculation on that node might yield a lower-cost result. Therefore, nodes are selected iteratively from the active node set until the active node set is empty, indicating that all decisions have been made, or the required number of iterations has been reached. At this point, the global optimal solution is used to obtain the ocean mission decision.

[0067] In this embodiment, nodes that have not completed their decisions are temporarily added to the set of active nodes. Nodes are then selected from the set of active nodes for further solution. The global upper bound is updated to the minimum total cost function value determined from the total cost function corresponding to the complete solution. The global optimal solution is updated to the complete solution corresponding to the minimum total cost function value. When the lower bound of a node is less than the global upper bound, it indicates that continuing the solution may yield a result with a lower cost. Therefore, nodes are selected until the set of active nodes is empty or the number of iterations is reached, thus completing the solution and obtaining the global optimal solution. This minimizes the cost of ocean mission decision-making and enables rapid response to unexpected missions.

[0068] In one embodiment, after selecting nodes from the set of active nodes, the method further includes: pruning the nodes to be pruned when there are nodes whose lower bound is greater than or equal to the global upper bound.

[0069] Specifically, the lower bound of the node to be pruned is greater than or equal to the global upper bound. The cloud obtains nodes from the set of activated nodes. When there is a node to be pruned whose lower bound is greater than or equal to the global upper bound, the node to be pruned is pruned, and nodes whose lower bound is less than the global upper bound are retained for further node expansion.

[0070] In this embodiment, when there is a node to be pruned whose lower bound is greater than or equal to the global upper bound, it means that the lower bound of the node's cost is no better than the current global optimal solution. This means that under this scheduling path, even if the optimal edge subnet is selected for the remaining tasks, a better global scheduling result cannot be obtained. At this time, the dynamic pruning mechanism is triggered, and the branch is directly discarded, thereby avoiding ineffective searches for a large number of infeasible or suboptimal scheduling schemes. Compared with traditional backtracking search, this mechanism can significantly reduce the computational burden on the cloud in large-scale marine scenarios.

[0071] In one embodiment, selecting nodes from the activated node set includes: Arrange the nodes in the active node set in ascending order of their lower bound values, and select the node with the smallest lower bound value from the active node set.

[0072] Specifically, the node is added to the set of active nodes Q, and the nodes in the set of active nodes are arranged in ascending order of their lower bound values. The node with the smallest lower bound value is selected from the set of active nodes.

[0073] In this implementation, the nodes in the active node set are sorted in ascending order of their lower bound values. The node with the smallest lower bound value is selected from the active node set. Combined with the pruning mechanism, the minimum cost function value can be obtained quickly. Subsequent nodes are easy to prune, which greatly reduces the amount of computation and allows for rapid ocean mission decisions.

[0074] In one embodiment, based on the above analysis, the total cost function C corresponding to the early warning task is... 预警 It can be expressed as the following formula: The total cost function C of the strike mission 打击 It can be expressed as the following formula: The total cost function for the communication task is shown below: Then, the total cost function that minimizes the total cost of multiple ocean missions. It can be written as the following formula: in, See the formula above.

[0075] A branch-and-bound algorithm based on row decomposition is used to solve the total cost function based on the constraint set, obtaining the ocean mission decision, including: Select the task type with an uncertain decision as the row branch, solve the total cost function, and generate nodes that meet the constraint set. The nodes include partial solutions and the lower bound values ​​corresponding to the partial solutions. Add the node to the set of active nodes, arrange the nodes in the set of active nodes in ascending order of their lower bound values, and select the node with the smallest lower bound value from the set of active nodes. If there is a node to be pruned whose lower bound is greater than or equal to the global upper bound, then the node to be pruned is pruned.

[0076] When the lower bound of a node is less than the global upper bound, return to the step of selecting the task type with an uncertain decision as the row branch, solving the total cost function, and generating nodes that meet the constraint set, until all task types have been decided and the complete solution and the total cost function value corresponding to the complete solution are obtained. Update the global upper bound to the minimum total cost function value determined from the total cost function values ​​corresponding to the complete solution; Update the global optimal solution to the complete solution corresponding to the minimum total cost function value; Return to the step of selecting the node with the smallest lower bound from the set of active nodes, until the set of active nodes is empty or the number of iterations is reached, and obtain the global optimal solution to obtain the ocean mission decision.

[0077] Specifically, taking the task type with an undetermined decision as k1, we solve for the total cost function, generating multiple nodes for the k1 row branch. These nodes include partial solutions and their corresponding lower bounds. We then add these k1 nodes to the activation node set Q, arranging them in ascending order of their lower bounds. We select the node with the smallest lower bound from Q to quickly obtain the global optimum and facilitate subsequent fast pruning. Since there are three task types, all three task types must be decided before a complete solution and its corresponding total cost function value can be generated. Therefore, the activation node set Q contains two types of nodes: k1 and k1k2.

[0078] At the beginning of the loop, the global upper bound is set to infinity, and the global optimum is empty. Once the cloud obtains the complete solution and its corresponding total cost function value, the global upper bound is updated to the minimum total cost function value determined from the total cost function values ​​corresponding to the complete solution, and the global optimum is updated to the complete solution corresponding to the minimum total cost function value. At this point, nodes with lower bounds greater than or equal to the global upper bound (ub) can be pruned to avoid further searching for those nodes and save time. If the lower bound is less than the global upper bound, it indicates that the node may have a better solution, so the node search continues until the complete solution is obtained. Finally, the global optimum is obtained when the set of active nodes is empty or the number of iterations is reached, thus enabling the ocean mission decision.

[0079] In this embodiment, the simplified code for the branch and bound algorithm based on row decomposition is as follows: Specifically, as shown in the algorithm, this algorithm is designed for cloud-based multi-task planning scenarios under ad hoc events in the marine satellite IoT. The algorithm first constructs a task scheduling search tree with an empty edge subnet-task mapping matrix as the root node. Each layer corresponds to the cloud's allocation decision for a certain type of ad hoc task across the candidate edge subnet set. The cloud uses a priority queue Q to sort the nodes to be expanded according to the lower bound of the total cost of the current scheduling scheme, and initializes the global upper bound to infinity to characterize the currently known optimal task orchestration cost.

[0080] In the main loop of the algorithm, the cloud preferentially pops the node to be expanded with the smallest lower bound, which corresponds to a "partially determined edge subnetwork-task scheduling scheme". If the lower bound of the cost of this node is no better than the current global optimal solution, it indicates that even if an optimal edge subnetwork is subsequently selected for the remaining tasks under this scheduling path, a better global scheduling result cannot be obtained. At this time, a dynamic pruning mechanism is triggered, and the branch is directly discarded, thereby avoiding ineffective searching for a large number of infeasible or suboptimal scheduling schemes. Compared with traditional backtracking search, this mechanism can significantly reduce the computational burden of the cloud in large-scale sea area scenarios.

[0081] In the branching phase (line 11), the algorithm adopts a row-first decomposition strategy, that is, it takes "the allocation decision of a certain type of opportunistic tasks on the candidate edge subnetwork set" as a branch unit, preferentially selects the first undetermined allocation task row k, and generates the candidate edge subnetwork set A_k that simultaneously satisfies the task executability constraint and the timing coupling constraint. This design fully incorporates the differences in spatial position, communication capability and operational attributes of different edge subnetworks in marine satellite Internet of Things, so that the search process is only carried out within the range of subnetworks with actual execution capability, thereby greatly reducing the number of invalid candidate solutions.

[0082] Subsequently, during the expansion of child nodes (line 14), the cloud calculates the optimistic lower bound lb' of the current partial scheduling scheme by relaxing the undetermined task allocation relationships. If the partial solution corresponding to this child node satisfies all hard constraints (such as task uniqueness, timing constraints and resource constraints) while still having the potential to further reduce the overall scheduling cost (lb'<ub), the depth search of this scheduling path will be continued; otherwise, the pruning is performed directly. When the scheduling scheme corresponding to a child node has covered all opportunistic tasks, the system uses it as a complete set of cloud task planning results and updates the global optimal solution.

[0083] In this embodiment, through the above design, the algorithm decomposes the multi-task planning problem in marine satellite Internet of Things layer by layer into "task-edge subnetwork" mapping decisions, and achieves rapid response to opportunistic events on the premise of ensuring global optimality. Simulation results show that when the scale of edge subnetworks reaches K=200, the algorithm can still maintain sub-second planning response capability and meet the real-time scheduling requirement of the cloud for sudden maritime situations. Overall, through hierarchical problem decomposition, scenario-aware pruning strategy and tightly coupled constraint verification mechanism, the method reduces the computational complexity from exponential growth, which was originally exponential to polynomial order .

[0084] III. Simulation Results This section presents the simulation results of the event-driven task planning method. The main simulation parameters are set as follows: the total task planning time slot is T=360, and each time slot is 1 minute long. All edge subnets are distributed in a 1000km×1000km sea area, and the number of edge subnets is approximately [10,60]. The number of maritime missions at any given time is randomly selected from [50, 120].

[0085] Figure 6 This is a schematic diagram illustrating the simulation results of the total cost of task planning in one embodiment. Figure 6 The horizontal axis represents the number of edge subnets, and the vertical axis represents the total task planning cost. Simulation results show that, under the premise of satisfying task execution constraints, the total task planning cost decreases with the increase of the number of edge subnets, and this decreasing trend exhibits a clear step-like decreasing pattern. This is because the more edge subnets there are, the larger the feasible set for task deployment. Only when a better-performing deployment scheme exists and modifications are made based on this better scheme will the task planning cost be significantly reduced. However, in most cases, the newly added edge subnets may be poorly positioned, in which case the original deployment scheme will not change, thus causing the task planning to exhibit a step-like decreasing trend. Furthermore, when the number of edge subnets continues to increase (for example, when the number of edge task subnets exceeds 50), the task planning cost will further decrease.

[0086] Figure 7 This is a schematic diagram comparing the optimization time of the traditional task planning mode and the task planning mode in this embodiment in one example. Figure 7 Primarily used for latency comparison in cloud-based task planning. Specifically, such as... Figure 7 As shown by the dotted line, traditional cloud-based task planning directly plans all operational nodes across the entire sea area. In this case, task planning time increases non-linearly with network size, making it difficult to adapt to the rapid response requirements of large-scale ad-hoc maritime operations. In the cloud-edge collaborative architecture proposed in this embodiment, scattered operational nodes at sea form an edge subnet through weak connections with the main station node. Under this architecture, the remote cloud only needs to quickly orchestrate and schedule ad-hoc tasks within the edge subnet, significantly reducing the task planning time of the remote cloud. The simulation results are shown in Figure 7. Therefore, the cloud-edge collaborative architecture can reduce task planning time from the traditional hours to minutes through this cascaded task planning method, effectively improving the ad-hoc response speed of the maritime satellite IoT.

[0087] In one embodiment, a cloud-based multi-task planning method for marine satellite IoT, applied in the cloud, includes: Step (a1) involves obtaining the total cost function that minimizes the total cost of various ocean missions, and obtaining the set of constraints corresponding to the total cost function. The set of constraints includes communication delay constraints and mission allocation constraints. The total cost function is determined by the sum of the cost functions corresponding to the early warning mission, the strike mission, and the communication mission. The cost function corresponding to the early warning mission is determined by the distance between the coordinates of the edge subnet and the coordinates of the detection target. The cost function of the strike mission is used to determine whether the detection target has entered the attackable range. The cost function corresponding to the communication mission is used to determine the sum of the transmission power of the edge subnet. The communication delay constraints include minimum communication delay constraints and upper limit constraints for data return delay. The set of constraints also includes temporal coupling constraints and constraints that the decision variables are binary variables. The temporal coupling constraints include executing the early warning mission first and then executing the strike mission.

[0088] Step (a2): Select the task type with uncertain decision as the row branch, solve the total cost function, and generate nodes that meet the constraint set. The nodes include partial solutions and the lower bound values ​​corresponding to the partial solutions.

[0089] Step (a3): Add the node to the set of active nodes, arrange the nodes in the set of active nodes in ascending order of their lower bound values, and select the node with the smallest lower bound value from the set of active nodes.

[0090] Step (a4): When there is a node to be pruned whose lower bound is greater than or equal to the global upper bound, prune the node to be pruned.

[0091] Step (a5): When the lower bound of a node is less than the global upper bound, return to the step of selecting the task type with an uncertain decision as the row branch, solving the total cost function, and generating nodes that meet the constraint set, until all task types have been decided and the complete solution and the total cost function value corresponding to the complete solution are obtained.

[0092] Step (a6) updates the global upper bound to the minimum total cost function value determined from the total cost function values ​​corresponding to the complete solution.

[0093] Step (a7) updates the global optimal solution to the complete solution corresponding to the minimum total cost function value.

[0094] Step (a8) returns to the step of selecting the node with the smallest lower bound from the set of active nodes, until the set of active nodes is empty or the number of iterations is reached, to obtain the global optimal solution and thus obtain the ocean mission decision. The ocean mission decision includes the decision variables for assigning the ocean mission to the corresponding edge subnet.

[0095] Step (a9): Based on the marine mission decision, the corresponding marine mission is sent to the corresponding edge subnet through the marine Internet of Things.

[0096] In this embodiment, by obtaining the total cost function that minimizes the total cost of multiple marine tasks and the corresponding set of constraints, and using a branch and bound algorithm based on row decomposition to solve the total cost function, marine task decisions that meet the set of constraints are obtained. This yields the decision variable of which marine task to assign to which edge subnet, enabling the edge subnet to execute the corresponding marine task. This approach is suitable for scenarios with large task changes and frequent responses to emergencies, allowing for flexible responses to emergencies. Furthermore, the branch and bound algorithm based on row decomposition can quickly solve the total cost function, thereby enabling rapid task planning and execution of marine tasks.

[0097] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0098] Based on the same inventive concept, this application also provides a cloud-based multi-task planning device for marine satellite IoT, used to implement the cloud-based multi-task planning method for marine satellite IoT described above. The solution provided by this device is similar to the implementation described in the above method. Therefore, the specific limitations of one or more embodiments of the cloud-based multi-task planning device for marine satellite IoT provided below can be found in the limitations of the cloud-based multi-task planning method for marine satellite IoT described above, and will not be repeated here.

[0099] In one exemplary embodiment, a cloud-based multi-task planning device for marine satellite Internet of Things is provided, used to implement the steps of the methods in various embodiments, including: an acquisition module, a solution module, and a transmission module, wherein: The acquisition module is used to acquire the total cost function that minimizes the total task cost of various ocean missions, and to acquire the set of constraints corresponding to the total cost function; the set of constraints includes communication delay constraints and task allocation constraints. The solution module is used to solve the total cost function based on the constraint set using a row decomposition-based branch and bound algorithm to obtain the ocean mission decision; the ocean mission decision includes the decision variables for assigning ocean missions to the corresponding edge subnets; The sending module is used to send the corresponding marine mission to the corresponding edge subnet via the marine Internet of Things based on the marine mission decision.

[0100] In one embodiment, the total cost function is determined based on the sum of the cost functions corresponding to the early warning mission, the strike mission, and the communication mission, respectively.

[0101] In one embodiment, the cost function for the early warning task is determined based on the distance between the coordinates of the edge subnet and the coordinates of the detected target; the cost function for the strike task is used to determine whether the detected target has entered the attackable range; and the cost function for the communication task is used to determine the sum of the transmission power of the edge subnet.

[0102] In one embodiment, the communication delay constraint includes a minimum communication delay constraint and a data return delay upper limit constraint; the constraint set also includes a time-coupling constraint and a binary variable constraint for the decision variable, and the time-coupling constraint includes executing the early warning task before executing the strike task.

[0103] In one embodiment, the solver module is used to: Add a node to the set of active nodes, and select a node from the set of active nodes. Update the global upper bound to the minimum total cost function value determined from the total cost function values ​​corresponding to the complete solution; Update the global optimal solution to the complete solution corresponding to the minimum total cost function value; When the lower bound of a node is less than the global upper bound, return to the step of selecting a node from the set of active nodes, until the set of active nodes is empty or the number of iterations is reached, to obtain the global optimal solution and thus obtain the ocean mission decision.

[0104] In one embodiment, the solver module is further configured to: If there is a node to be pruned whose lower bound is greater than or equal to the global upper bound, then the node to be pruned is pruned.

[0105] In one embodiment, the solver module is further configured to: Arrange the nodes in the active node set in ascending order of their lower bound values, and select the node with the smallest lower bound value from the active node set.

[0106] The modules in the aforementioned cloud-based multi-task planning device for marine satellite IoT can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the corresponding operations of each module.

[0107] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 8 As shown, this computer device includes a processor, memory, input / output (I / O) interfaces, and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network. When the computer program is executed by the processor, it implements a cloud-based multi-task planning method for a marine satellite Internet of Things (IoT).

[0108] Those skilled in the art will understand that Figure 8 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0109] In one embodiment, a computer device is provided, including a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement the steps of the cloud-based multi-task planning method for ocean satellite Internet of Things.

[0110] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps described in the cloud-based multi-task planning method for ocean satellite Internet of Things.

[0111] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps described in the cloud-based multi-task planning method for the Internet of Things for marine satellites.

[0112] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0113] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0114] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0115] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A cloud-based multi-task planning method for marine satellite Internet of Things, characterized in that, Applied to the cloud, the method includes: The method involves obtaining a total cost function that minimizes the overall cost of various maritime missions, and a set of constraints corresponding to this total cost function. The maritime missions include early warning missions, strike missions, and communication missions. The total cost function is determined by the sum of the cost functions corresponding to the early warning mission, strike mission, and communication mission. The cost function for the early warning mission is determined based on the distance between the coordinates of the edge subnet and the coordinates of the detected target. The cost function for the strike mission is used to determine whether the detected target has entered the attackable range. The cost function for the communication mission is used to determine the sum of the transmission power of the edge subnet. The set of constraints includes communication delay constraints and mission allocation constraints. The communication delay constraints include minimum communication delay constraints and upper limit constraints for data return delay. The set of constraints also includes temporal coupling constraints and constraints where all decision variables are binary variables. The temporal coupling constraints include executing the early warning mission before executing the strike mission. Select the task type with an uncertain decision as the row branch, solve the total cost function, and generate nodes that meet the constraint set; add the nodes to the activation node set, arrange the nodes in the activation node set in ascending order of lower bound value, and select the node with the smallest lower bound value from the activation node set; the node includes a partial solution and the lower bound value corresponding to the partial solution. For the node, return to the step of selecting the task type with the undetermined decision as the row branch, solve the total cost function, and generate a node that meets the set of constraints, until all task types have been decided, and obtain the complete solution and the total cost function value corresponding to the complete solution; Based on the complete solution and the total cost function value corresponding to the complete solution, the complete solution with the minimum total cost function value is determined to obtain the marine mission decision; the marine mission decision includes the decision variable of assigning the marine mission to the corresponding edge subnet; Based on the aforementioned marine mission decision, the corresponding marine mission is sent to the corresponding edge subnet via the marine Internet of Things.

2. The method according to claim 1, characterized in that, The step of determining the complete solution with the minimum total cost function value based on the complete solution and the total cost function value corresponding to the complete solution, and obtaining the ocean mission decision, includes: The global upper bound is updated to the minimum total cost function value determined from the total cost function values ​​corresponding to the complete solution; Update the global optimal solution to the complete solution corresponding to the minimum total cost function value; When the lower bound of the node is less than the global upper bound, the step of selecting a node from the set of active nodes is returned until the set of active nodes is empty or the number of iterations is reached, and the global optimal solution is obtained to obtain the ocean mission decision.

3. The method according to claim 2, characterized in that, After selecting the node with the smallest lower bound from the set of activated nodes, the method further includes: When there is a node to be pruned whose lower bound is greater than or equal to the global upper bound, the node to be pruned is pruned.

4. A cloud-based multi-task planning device for marine satellite Internet of Things, characterized in that, The steps for implementing the method according to any one of claims 1 to 3 include: The acquisition module is used to acquire the total cost function that minimizes the total task cost of multiple maritime missions, and to acquire the constraint set corresponding to the total cost function. The maritime missions include early warning missions, strike missions, and communication missions. The total cost function is determined by the sum of the cost functions corresponding to the early warning mission, strike mission, and communication mission respectively. The cost function corresponding to the early warning mission is determined based on the distance between the coordinates of the edge subnet and the coordinates of the detected target. The cost function for the strike mission is used to determine whether the detected target has entered the attackable range. The cost function corresponding to the communication mission is used to determine the sum of the transmission power of the edge subnet. The constraint set includes communication delay constraints and mission allocation constraints. The communication delay constraints include minimum communication delay constraints and upper limit constraints for data return delay. The constraint set also includes temporal coupling constraints and constraints that all decision variables are binary variables. The temporal coupling constraints include executing the early warning mission first, followed by the strike mission. The solution module is used to select the task type with an uncertain decision as the row branch, solve the total cost function, and generate nodes that meet the constraint set; add the nodes to the activation node set, arrange the nodes in the activation node set in ascending order of the lower bound value, and select the node with the smallest lower bound value from the activation node set; the node includes a partial solution and the lower bound value corresponding to the partial solution. For the node, return to the step of selecting the task type with the undetermined decision as the row branch, solve the total cost function, and generate a node that meets the set of constraints, until all task types have been decided, and obtain the complete solution and the total cost function value corresponding to the complete solution; Based on the complete solution and the total cost function value corresponding to the complete solution, the complete solution with the minimum total cost function value is determined to obtain the marine mission decision; the marine mission decision includes the decision variable of assigning the marine mission to the corresponding edge subnet; The sending module is used to send the corresponding marine mission to the corresponding edge subnet through the marine Internet of Things based on the marine mission decision.

5. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 3.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 3.

Images