Marine large-scale resource rescue scheduling method, device and equipment

Abstract

The invention provides a maritime large-scale resource rescue scheduling method. A to-be-scheduled rescue strength index set and rescue environment information are acquired. And according to the to-be-dispatched rescue force index set, constructing a highest comprehensive dispatching scoring function and a minimum to-be-dispatched rescue force function. And according to the highest comprehensive scheduling scoring function and the minimum to-be-scheduled rescue force function, an initial scheduling model of the to-be-scheduled rescue force in an initial quick response stage is constructed through an approximate ideal solution sorting method, and an initial scheduling result is determined. And according to the initial scheduling result, the rescue site capacity and the rescue environment information, taking the shortest rescue time as a target, and constructing a continuous scheduling model in a continuous dynamic ordered stage. The continuous scheduling model is solved through a Q-learning method, a continuous scheduling result is obtained, rapid and reasonable scheduling in an initial rapid response stage and ordered allocation in a continuous dynamic ordered stage are achieved, and the actual requirements of rapid response and continuous ordering of maritime large-scale resource rescue scheduling are met.

Description

Technical Field

[0001] This invention relates to the field of maritime rescue resource scheduling technology, and in particular to a method, apparatus and equipment for large-scale maritime resource rescue scheduling. Background Technology

[0002] With the booming development of the cruise industry and the continuous expansion of maritime tourism, accidents such as collisions and fires on cruise ships could lead to numerous casualties if evacuation and rescue are not timely. Maritime rescue is characterized by harsh environments, suddenness, urgency, and complex weather conditions, posing unprecedented challenges to large-scale maritime life-saving operations. How to scientifically, rationally, orderly, and efficiently allocate existing rescue resources to improve rescue efficiency has become an important issue urgently requiring research.

[0003] Currently, traditional methods, such as the rescue force optimization and ranking method based on fuzzy decision theory and the search and rescue risk assessment method based on the analytic hierarchy process (AHP), can achieve the allocation of maritime rescue resources to a certain extent. However, when faced with large-scale maritime life-saving missions, these traditional methods are insufficient to meet actual rescue needs due to the lack of systematic research on the allocation of life-saving resources in large-scale maritime accidents. Therefore, to solve the decision-making problem of large-scale maritime life-saving force allocation, existing technologies construct a first-stage decision model with the shortest time as the optimization objective; a second-stage decision model with the shortest rescue time and the fewest rescue forces as optimization objectives; and combine simulated annealing algorithm and genetic algorithm to simultaneously optimize rescue time and the number of rescue forces, thereby achieving the allocation of life-saving resources in large-scale maritime accidents.

[0004] However, when using existing technologies, the dispatching behavior is not divided into stages based on the logic of the actual rescue process, and the dispatching mechanism of "rapid response and continuous order" that rescue forces should follow in complex maritime environments is ignored. Therefore, it is difficult to effectively meet the actual resource needs in real large-scale maritime life rescue. Summary of the Invention

[0005] This invention provides a method, apparatus, and equipment for large-scale maritime resource rescue dispatching. By dividing large-scale maritime resource rescue dispatching into two stages—an initial rapid response stage and a continuous dynamic orderly stage—the method enables rapid and rational dispatching of rescue forces during the initial rapid response stage, thus fulfilling emergency rescue needs and effectively avoiding congestion and redundancy issues that may occur when dispatching rescue forces during the continuous dynamic orderly stage. Subsequently, in the continuous dynamic orderly stage, by comprehensively considering the capacity of the rescue site and rescue environment information, the method ensures the orderly allocation of rescue forces, enabling the entire rescue process to meet the practical requirements of "rapid response and continuous orderly" in large-scale maritime resource rescue dispatching.

[0006] To achieve the above objectives, in a first aspect, embodiments of the present invention provide a method for large-scale maritime resource rescue and dispatching, comprising: Obtain the set of indicators for the rescue forces to be dispatched and information on the rescue environment; Based on the set of indicators of the rescue forces to be dispatched, construct the highest comprehensive dispatch score function and the minimum rescue forces to be dispatched function; Based on the highest comprehensive dispatch score function and the minimum number of rescue forces to be dispatched function, the initial dispatch model of the rescue forces to be dispatched in the initial rapid response phase is constructed by using the approximation ideal solution sorting method, and the initial dispatch result of the initial rapid response phase is determined. Based on the initial scheduling results, the capacity of the rescue site, and the rescue environment information, a continuous scheduling model for the rescue forces to be dispatched is constructed in a continuous, dynamic, and orderly phase, with the goal of minimizing the rescue time. The continuous scheduling model is solved using the Q-learning method to obtain the continuous scheduling results for the continuous dynamic ordered phase.

[0007] In one embodiment, the set of indicators for the rescue forces to be dispatched includes: wind resistance level indicators, dispatch time indicators, passenger capacity indicators, flight speed indicators, and resource type indicators. The step of constructing the highest comprehensive dispatch score function and the minimum number of rescue forces to be dispatched based on the set of indicators includes: The evaluation score is obtained by expert scoring between any two of the wind resistance level index, dispatch time index, passenger capacity index, speed index and resource type index. The judgment matrix corresponding to the set of rescue forces to be dispatched is determined based on multiple evaluation scores. Based on the judgment matrix, construct the highest comprehensive scheduling score function; Based on the wind resistance level index, scheduling time index, passenger capacity index, speed index, and resource type index, multiple initial rapid response phase constraints are determined, and the minimum required rescue force function is determined based on these multiple initial rapid response phase constraints.

[0008] In one embodiment, the step of constructing an initial dispatch model for the rescue forces to be dispatched in the initial rapid response phase, based on the highest comprehensive dispatch score function and the least dispatchable rescue force function, using an approximation-ideal-solution ranking method, and determining the initial dispatch result of the initial rapid response phase, includes: Based on the highest comprehensive scheduling score function and the minimum number of rescue forces to be scheduled, determine the single-objective optimization proximity function; Based on the single-objective optimization proximity function, solve for the single-objective optimization proximity corresponding to the multiple initial scheduling results in the initial rapid response phase; Compare the magnitudes of the single-objective optimization proximity scores and determine the initial scheduling result corresponding to the maximum single-objective optimization proximity score as the initial scheduling result.

[0009] In one embodiment, determining the single-objective optimal proximity function based on the highest comprehensive scheduling score function and the least number of rescue forces to be scheduled includes: The highest comprehensive scheduling score function is normalized to obtain the normalized highest comprehensive scheduling score function; The minimum number of rescue forces to be dispatched is normalized to obtain the normalized minimum number of rescue forces to be dispatched function. Based on the normalized highest comprehensive scheduling score function and the normalized minimum number of rescue forces to be scheduled function, construct the positive ideal solution and negative ideal solution corresponding to each of the initial scheduling results; Based on the positive and negative ideal solutions, the single-objective optimization proximity function is determined.

[0010] In one embodiment, the step of constructing a continuous scheduling model for the rescue forces to be dispatched during a continuous, dynamic, and orderly phase, based on the initial dispatch results, the capacity of the rescue site, and the rescue environment information, with the goal of minimizing the rescue time, includes: Based on the initial scheduling results, the capacity of the rescue site, and the rescue environment information, multiple continuous dynamic and orderly phase constraints are determined. Based on the multiple continuous dynamic orderly stage constraints, and with the goal of minimizing the rescue time, a continuous scheduling model for the rescue forces to be dispatched during the continuous dynamic orderly stage is constructed.

[0011] In one embodiment, before solving the continuous scheduling model using the Q-learning method to obtain the continuous scheduling result of the continuous dynamic ordered phase, the method further includes: Based on the initial scheduling results, the capacity of the rescue site, the rescue environment information, and multiple continuous dynamic orderly stage constraints, multiple state space sets and the action space corresponding to each state space set are determined. The state space set includes: state spaces corresponding to multiple times, and each action space includes: rescue force execution actions corresponding to multiple times. Each rescue force execution action corresponds one-to-one with each state space. The state space at time t+1 is determined based on the execution results of the rescue force execution actions at time t. For each set of state spaces and action spaces, construct a cumulative reward function.

[0012] In one embodiment, solving the persistent scheduling model using the Q-learning method to obtain the persistent scheduling result of the persistent dynamic ordered phase includes: For each set of state spaces and action spaces, the Q-learning method is used to solve the problem, and the cumulative reward value corresponding to each set of state spaces and action spaces is obtained according to the cumulative reward function. By comparing the magnitudes of multiple cumulative reward values, the action space corresponding to the maximum cumulative reward value is determined as the continuous scheduling result of the continuous dynamic ordered stage.

[0013] In one embodiment, solving for each set of state spaces and action spaces using the Q-learning method includes: The state space is initialized based on the initial scheduling results, the capacity of the rescue site, and the rescue environment information. Based on the greed strategy, determine the action to be performed by the rescue force at the current moment among the multiple rescue actions; The execution of the rescue force action to be executed is performed, and the action value of the rescue force action to be executed is obtained according to the action value function. It is then determined whether the rescue is completed. If not, the execution is returned to determine the rescue force action to be executed at the current moment among multiple rescue force actions according to the greed strategy, and the state space at the next moment is updated according to the execution result at the current moment until the rescue is completed.

[0014] Secondly, embodiments of the present invention provide a large-scale maritime resource rescue and dispatching device, comprising: The acquisition module is used to acquire the set of indicators of the rescue forces to be dispatched and rescue environment information; The function construction module is used to construct the highest comprehensive dispatch score function and the minimum number of dispatchable rescue forces function based on the set of indicators of the rescue forces to be dispatched; The initial rapid response phase solution module is used to construct the initial scheduling model of the rescue forces to be scheduled in the initial rapid response phase based on the highest comprehensive scheduling score function and the minimum number of rescue forces to be scheduled, and to determine the initial scheduling result of the initial rapid response phase by using the approximation ideal solution sorting method. The continuous scheduling model construction module is used to construct a continuous scheduling model for the rescue forces to be dispatched in a continuous, dynamic, and orderly phase, based on the initial scheduling results, the capacity of the rescue site, and the rescue environment information, with the goal of minimizing the rescue time. The continuous dynamic ordered phase solution module is used to solve the continuous scheduling model using the Q-learning method to obtain the continuous scheduling results of the continuous dynamic ordered phase.

[0015] Thirdly, embodiments of the present invention provide an electronic 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 of the large-scale maritime resource rescue and dispatch method described in the first aspect.

[0016] The above technical solution has the following technical effects: Thus, the large-scale maritime resource rescue scheduling method provided in this embodiment obtains a set of indicators for the rescue forces to be scheduled and rescue environment information. Based on the set of indicators for the rescue forces to be scheduled, a highest comprehensive scheduling score function and a minimum number of rescue forces to be scheduled function are constructed. Based on the highest comprehensive scheduling score function and the minimum number of rescue forces to be scheduled function, an initial scheduling model for the rescue forces to be scheduled in the initial rapid response phase is constructed using an approximation ideal solution ranking method, and the initial scheduling result for the initial rapid response phase is determined. Based on the initial scheduling result, the capacity of the rescue site, and the rescue environment information, a continuous scheduling model for the rescue forces to be scheduled in the continuous dynamic and orderly phase is constructed with the shortest rescue time as the objective. The continuous scheduling model is solved using the Q-learning method to obtain the continuous scheduling result for the continuous dynamic and orderly phase. In this way, by dividing the large-scale maritime resource rescue scheduling into two phases—the initial rapid response phase and the continuous dynamic and orderly phase—the rapid and reasonable scheduling of rescue forces to be scheduled can be achieved in the initial rapid response phase, thereby meeting the emergency rescue needs in the initial rapid response phase and effectively avoiding congestion and redundancy problems when the rescue forces to be scheduled are dispatched in the continuous dynamic and orderly phase. Subsequently, in the continuous and orderly phase, by comprehensively considering the capacity of the rescue site and information about the rescue environment, it is possible to ensure the orderly allocation of rescue forces to be dispatched, so that the entire rescue process meets the actual needs of "rapid response and continuous order" for large-scale maritime resource rescue and dispatch. Attached Figure Description

[0017] Figure 1 A flowchart illustrating a large-scale maritime resource rescue and dispatch method according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a large-scale maritime resource rescue and dispatch scenario according to an embodiment of the present invention; Figure 3 A schematic diagram of a large-scale maritime resource rescue and dispatching device provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0018] To further illustrate the various embodiments, the present invention provides accompanying drawings. These drawings are part of the disclosure of the present invention, primarily used to illustrate the embodiments and to explain the operating principles of the embodiments in conjunction with the relevant descriptions in the specification. With reference to these drawings, those skilled in the art should be able to understand other possible implementations and the advantages of the present invention. Components in the drawings are not drawn to scale, and similar component symbols are generally used to represent similar components.

[0019] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments.

[0020] Example 1: Figure 1 This is a flowchart illustrating a method for large-scale maritime resource rescue and dispatch according to an embodiment of the present invention. Figure 2 This is a schematic diagram illustrating a scenario of large-scale maritime resource rescue and dispatch according to an embodiment of the present invention. This embodiment specifically includes the following steps: S10: Obtain the set of indicators for the rescue forces to be dispatched and information on the rescue environment.

[0021] Among them, the rescue forces to be dispatched refer to maritime search and rescue equipment that can be dispatched, such as... Figure 2 As shown, the rescue forces to be dispatched include: specialized rescue vessels, patrol boats, and specialized rescue helicopters. The set of indicators for the rescue forces to be dispatched refers to the rescue indicators specific to these forces. This set includes: wind resistance level, dispatch time, passenger capacity, speed, and resource type. The set of indicators allows for the measurement of rescue speed, safety, and carrying capacity during the rescue process. However, this invention is not limited to these specific indicators; those skilled in the art can set them according to actual circumstances.

[0022] Rescue environment information refers to information about the current location of the distressed vessel, such as information on the remaining personnel awaiting rescue, but it is not limited to this. This invention does not specifically limit it, and those skilled in the art can set it according to the actual situation.

[0023] S11: Based on the set of indicators of the rescue forces to be dispatched, construct the highest comprehensive dispatch score function and the minimum number of rescue forces to be dispatched function.

[0024] Specifically, the system obtains the set of indicators of available rescue forces and the rescue environment information of the distressed vessel. After obtaining the set of indicators of available rescue forces and the rescue environment information of the distressed vessel, it constructs the highest comprehensive dispatch score function and the minimum available rescue force function based on the set of indicators of available rescue forces.

[0025] Optionally, based on the above embodiments, in some embodiments of the present invention, one implementation of S11 may be: S111: Obtain evaluation scores from any two of the following indicators: wind resistance level, dispatch time, number of passengers, speed, and resource type, through expert scoring. Determine the judgment matrix corresponding to the set of rescue forces to be dispatched based on multiple evaluation scores.

[0026] S112: Construct the highest comprehensive scheduling score function based on the judgment matrix.

[0027] Specifically, the system targets a set of indicators for the rescue forces to be dispatched, including wind resistance level, dispatch time, passenger capacity, flight speed, and resource type. Expert scoring is used to obtain evaluation scores between any two of these indicators. Further, a judgment matrix is ​​determined based on these multiple evaluation scores to identify the corresponding set of rescue force indicators. After obtaining the judgment matrix, a highest-level comprehensive dispatch score function is constructed.

[0028] Optionally, based on the above embodiments, in some embodiments of the present invention, the highest comprehensive scheduling score function may be defined by the following expression:

[0029] in, Indicates the initial rapid response phase. The rescue forces awaiting dispatch are in the first Comprehensive scheduling score under each indicator Indicates the initial rapid response phase. The first rescue center that meets the criteria for being able to be dispatched One rescue force awaiting dispatch. When the value is 1, the rescue force to be dispatched is selected; when the value is 0, dispatch is not selected.

[0030] Optionally, based on the above embodiments, in some embodiments of the present invention, the comprehensive scheduling score may be defined by the following expression:

[0031] in, Indicates the initial rapid response phase. The weighting coefficients of each indicator Indicates the initial rapid response phase. The rescue forces awaiting dispatch are in the first Normalized values ​​for each indicator.

[0032] S113: Based on wind resistance level indicators, dispatch time indicators, passenger capacity indicators, speed indicators, and resource type indicators, determine multiple initial rapid response phase constraints, and determine the minimum number of rescue forces to be dispatched based on these multiple initial rapid response phase constraints.

[0033] Optionally, based on the above embodiments, in some embodiments of the present invention, multiple initial rapid response phase constraints include: navigation time constraints, dispatch time constraints, upper limit constraints on the dispatch time of the rescue forces to be dispatched, and wind resistance constraints on the rescue forces to be dispatched. Based on this, the navigation time constraint can be limited by the following expression:

[0034] in, This indicates the first phase of the initial rapid response. The first rescue center that meets the criteria for being able to be dispatched The sailing time of each rescue force awaiting dispatch. This indicates the first phase of the initial rapid response. The first rescue center that meets the criteria for being able to be dispatched The distance between the rescue forces awaiting dispatch and the distressed vessel. This indicates the first phase of the initial rapid response. The rescue center that meets the criteria for being able to be dispatched The speed of the rescue forces awaiting dispatch.

[0035] The scheduling time constraint can be limited by the following expression:

[0036] in, This indicates the first phase of the initial rapid response. The first rescue center that meets the criteria for being able to be dispatched The response time of each pending rescue force. This indicates the first phase of the initial rapid response. The first rescue center that meets the criteria for being able to be dispatched The dispatch time for each available rescue force.

[0037] The upper limit constraint on scheduling time can be limited by the following expression:

[0038] in, This indicates the golden rescue time during the initial rapid response phase.

[0039] Wind resistance constraints can be defined by the following expression:

[0040] in, This indicates the first phase of the initial rapid response. The first rescue center that meets the criteria for being able to be dispatched The wind resistance rating of the rescue forces awaiting dispatch. This indicates the wind force on the distressed vessel.

[0041] Specifically, based on wind resistance level indicators, scheduling time indicators, passenger capacity indicators, speed indicators, and resource type indicators, multiple initial rapid response phase constraints are determined. After obtaining these multiple initial rapid response phase constraints, the minimum number of rescue forces to be dispatched is further determined based on these constraints.

[0042] Optionally, based on the above embodiments, in some embodiments of the present invention, the minimum number of rescue forces to be dispatched can be defined by the following expression:

[0043] S12: Based on the highest comprehensive dispatch score function and the minimum number of rescue forces to be dispatched function, construct the initial dispatch model of the rescue forces to be dispatched in the initial rapid response phase by using the approximation ideal solution sorting method, and determine the initial dispatch results in the initial rapid response phase.

[0044] Optionally, based on the above embodiments, in some embodiments of the present invention, S12 may be implemented as follows: S121: Determine the single-objective optimization proximity function based on the highest comprehensive scheduling score function and the minimum number of rescue forces to be scheduled.

[0045] Optionally, based on the above embodiments, in some embodiments of the present invention, one implementation of S121 may be: S1211: Normalize the highest comprehensive scheduling score function to obtain the normalized highest comprehensive scheduling score function.

[0046] Specifically, the highest comprehensive scheduling score function is normalized to obtain a normalized highest comprehensive scheduling score function.

[0047] Optionally, based on the above embodiments, in some embodiments of the present invention, the normalized highest comprehensive scheduling score function may be defined by the following expression:

[0048] in, This represents the average of the overall scheduling score. This represents the minimum value of the overall scheduling score. This represents the maximum value of the comprehensive scheduling score.

[0049] S1212: Normalize the minimum number of rescue forces to be dispatched function to obtain the normalized minimum number of rescue forces to be dispatched function.

[0050] Specifically, the function of the minimum number of rescue forces to be dispatched is normalized to obtain the normalized function of the minimum number of rescue forces to be dispatched.

[0051] Optionally, based on the above embodiments, in some embodiments of the present invention, the normalized minimum number of rescue forces to be dispatched function may be defined by the following expression:

[0052] in, This indicates the maximum number of rescue forces to be dispatched during the initial rapid response phase. This indicates the number of rescue forces awaiting dispatch during the initial rapid response phase. This indicates the minimum number of rescue forces to be dispatched during the initial rapid response phase.

[0053] S1213: Based on the normalized highest comprehensive scheduling score function and the normalized minimum number of rescue forces to be scheduled function, construct the positive ideal solution and negative ideal solution corresponding to each initial scheduling result.

[0054] Specifically, after obtaining the normalized highest comprehensive scheduling score function and the normalized minimum number of rescue forces to be scheduled function, positive ideal solutions and negative ideal solutions corresponding to each initial scheduling result are constructed based on the normalized highest comprehensive scheduling score function and the normalized minimum number of rescue forces to be scheduled function.

[0055] Optionally, based on the above embodiments, in some embodiments of the present invention, the ideal solution can be defined by the following expression:

[0056] in, These represent the weight coefficients of the normalized highest comprehensive scheduling score function. The weight coefficients represent the normalized minimum number of rescue forces to be dispatched function.

[0057] Optionally, based on the above embodiments, in some embodiments of the present invention, the negative ideal solution may be defined by the following expression:

[0058] S1214: Determine the single-objective optimization proximity function based on the positive and negative ideal solutions.

[0059] Specifically, after obtaining the positive and negative ideal solutions, the single-objective optimization proximity function is determined based on the positive and negative ideal solutions.

[0060] Optionally, based on the above embodiments, in some embodiments of the present invention, the single-objective optimization proximity function may be defined by the following expression:

[0061] S122: Based on the single-objective optimization proximity function, solve for the single-objective optimization proximity corresponding to the multiple initial scheduling results in the initial rapid response phase.

[0062] The initial dispatch result refers to the combination of currently dispatchable rescue forces determined according to the constraints of the initial rapid response phase.

[0063] Specifically, for multiple initial scheduling results, the single-objective optimization proximity function is used to solve for the single-objective optimization proximity corresponding to each initial scheduling result.

[0064] S123: Compare the magnitudes of multiple single-objective optimization proximity scores and determine the initial scheduling result corresponding to the maximum single-objective optimization proximity score as the initial scheduling result.

[0065] Specifically, the magnitudes of the single-objective optimization proximity corresponding to multiple initial scheduling results are compared, and the initial scheduling result corresponding to the maximum single-objective optimization proximity is determined as the initial scheduling result.

[0066] For example, referring to Table 1, the initial dispatch results in the initial rapid response phase were as follows: two specialized rescue helicopters, numbered Helicopter 1 and Helicopter 4, one specialized rescue vessel, numbered Rescue Vessel 4, and three patrol boats, numbered Patrol Boat 28, Patrol Boat 24, and Patrol Boat 22, were dispatched. The comprehensive dispatch score for the specialized rescue helicopters was the highest at 0.936, indicating that they possess characteristics such as high deployment speed and strong maneuverability, but their carrying capacity is relatively low compared to other rescue forces awaiting dispatch. The specialized rescue vessels and patrol boats effectively supplemented the insufficient carrying capacity of the helicopters, thereby enabling the rapid and rational dispatch of rescue forces awaiting dispatch in the initial rapid response phase, thus fulfilling the emergency rescue needs in this phase.

[0067] Table 1. Initial scheduling results during the initial rapid response phase

[0068] S13: Based on the initial dispatch results, the capacity of the rescue site, and the rescue environment information, construct a continuous dispatch model for the rescue forces to be dispatched in a continuous, dynamic, and orderly phase, with the goal of minimizing the rescue time.

[0069] Optionally, based on the above embodiments, in some embodiments of the present invention, one implementation of S13 may be: S131: Based on the initial dispatch results, the capacity of the rescue site, and the rescue environment information, determine multiple continuous dynamic orderly phase constraints.

[0070] The capacity of the rescue site refers to the number of rescue forces that can still be accommodated at the rescue site after the initial rapid response phase has been completed and all other rescue efforts have been made.

[0071] Specifically, after obtaining and executing the initial scheduling results, multiple continuous dynamic orderly stage constraints are determined based on the initial scheduling results, the capacity of the rescue site, and the rescue environment information.

[0072] Optionally, based on the above embodiments, in some embodiments of the present invention, multiple continuous dynamic ordered stage constraints include: constraints on the number of people rescued by the dispatched rescue forces, constraints on the dispatch quantity of the dispatched rescue forces, constraints on the capacity of the dispatched rescue forces, constraints on the rescue time of the dispatched rescue forces at the distressed vessel, and constraints on the selection of the dispatched rescue forces according to risk level. The constraint on the number of people rescued by the dispatched rescue forces can be limited by the following expression:

[0073] in, Indicates the continuous dynamic ordered stage. The first rescue center that meets the criteria for being able to be dispatched A value of 1 indicates that a rescue force awaiting dispatch is selected for dispatch, while a value of 0 indicates that no dispatch is selected. Indicates the continuous dynamic ordered stage. The first rescue center that meets the criteria for being able to be dispatched The number of people in each rescue force awaiting dispatch. This indicates the total number of people in distress at the distressed vessel.

[0074] The scheduling quantity constraint can be limited by the following expression:

[0075] in, This indicates the total number of available and dispatchable rescue forces.

[0076] Capacity constraints can be limited by the following expression:

[0077] in, This represents the minimum remaining passenger capacity at the location of the distressed vessel, accommodating all available rescue forces awaiting dispatch. This represents the maximum number of people that can be accommodated at the distressed vessel, including the sum of all available rescue forces awaiting dispatch. Indicates the continuous dynamic ordered stage. The first rescue center that meets the criteria for being able to be dispatched The remaining passenger capacity of the rescue forces awaiting dispatch.

[0078] The rescue time constraint can be limited by the following expression:

[0079] in, Indicates the continuous dynamic ordered stage. The first rescue center that meets the criteria for being able to be dispatched The efficiency of rescue efforts by the available rescue forces.

[0080] Selection constraints can be qualified by the following expressions:

[0081] in, Indicates the risk level, when At that time, the risk level was high. At that time, the risk level was medium risk. At this time, the risk level is low. At high risk, in order to ensure the safety of the rescue forces to be dispatched, priority is given to dispatching rescue forces with a large number of people; at medium and low risk, in order to improve rescue efficiency, priority is given to dispatching rescue forces with short time and fast speed.

[0082] S132: Based on multiple constraints of continuous dynamic orderly stages, and with the goal of minimizing rescue time, construct a continuous dispatch model for the rescue forces to be dispatched during the continuous dynamic orderly stages.

[0083] Specifically, after obtaining multiple continuous dynamic orderly stage constraints, and based on these constraints, a continuous scheduling model for the rescue forces to be dispatched during the continuous dynamic orderly stage is constructed with the goal of minimizing the rescue time.

[0084] Optionally, based on the above embodiments, in some embodiments of the present invention, the continuous scheduling model of the continuous dynamic ordered phase can be defined by the following expression:

[0085] in, Indicates the time of rescue.

[0086] S14: Solve the continuous scheduling model using the Q-learning method to obtain the continuous scheduling results for the continuous dynamic ordered phase.

[0087] Optionally, based on the above embodiments, in some embodiments of the present invention, the method further includes the following before performing S14: S20: Based on the initial scheduling results, the capacity of the rescue site, the rescue environment information, and multiple continuous dynamic orderly stage constraints, determine multiple state space sets and the action space corresponding to each state space set.

[0088] The state space set includes: state spaces corresponding to multiple time points. These state spaces are determined based on the resource allocation results during the rapid response phase, the capacity of the rescue site, and information about the rescue environment. ,in, This represents the sum of the remaining manpower of the rescue forces awaiting dispatch at the current distressed vessel. Risk level; The utilization rate of the rescue forces to be dispatched (the ratio of the number of people rescued by the rescue forces to the number of people they can carry) is used to measure the efficiency of the use of the rescue forces to be dispatched. The minimum dispatch time for the rescue forces to be dispatched is the earliest expected arrival time of the rescue forces to be dispatched, which can indirectly reflect the duration of the rescue. This represents the number of people currently in distress at the vessel that have not yet been successfully rescued.

[0089] Each action space includes: multiple rescue actions performed by rescue forces at different times, and each rescue action corresponds one-to-one with each state space. For example, the set of state spaces is as follows: The action space is then: .

[0090] The state space at time t+1 is determined based on the execution result of the rescue force's action at time t. It can be understood that after executing the rescue force's action at time t, the state space at time t+1 will be updated based on the execution result of that action.

[0091] S21: For each set of state spaces and action spaces, construct a cumulative reward function.

[0092] Specifically, for each set of state spaces and action spaces, a cumulative reward function is constructed.

[0093] Optionally, based on the above embodiments, in some embodiments of the present invention, the cumulative return function may be defined by the following expression:

[0094] in, This represents the weighting coefficient of the cumulative assistance received in the k-th instance. Indicates the first Rewards for timely rescue.

[0095] Optional, according to the formula Determine the first Rewards for timely rescue.

[0096] Optionally, based on the above embodiments, in some embodiments of the present invention, one implementation of S14 may be: S141: For each set of state spaces and action spaces, solve using the Q-learning method, and obtain the cumulative reward value corresponding to each set of state spaces and action spaces based on the cumulative reward function.

[0097] Optionally, based on the above embodiments, in some embodiments of the present invention, one implementation of solving for each set of state spaces and action spaces using the Q-learning method may be: S30: Initialize the state space based on the initial scheduling results, the capacity of the rescue site, and the rescue environment information.

[0098] Specifically, the state space is initialized based on the initial scheduling results of the initial rapid response phase, the capacity of the rescue site, and the rescue environment information.

[0099] It should be noted that Q-learning is a tabular reinforcement learning method. For the initial Q-table, it contains the state space and action space at different time points, i.e. Q(s, a) , indicating in the state space s Rescue forces in the lower execution space perform actions a The cumulative rewards obtained.

[0100] S31: Based on the greed strategy, determine the action to be performed by the rescue force at the current moment among multiple rescue actions.

[0101] S32: Execute the action to be performed by the rescue force, and obtain the action value of the action to be performed by the rescue force according to the action value function. Determine whether the rescue is completed. If not, return to the previous step. According to the greedy strategy, determine the action to be performed by the rescue force at the current moment among multiple rescue force actions, and update the state space at the next moment according to the execution result at the current moment, until the rescue is completed.

[0102] Specifically, based on the greedy strategy, the action to be executed by a rescue force at the current moment is determined among multiple rescue force actions. This action is then executed, and its action value is obtained according to the action value function. Further determination is made as to whether the rescue has been completed. If not, the process returns to the previous step, updating the state space for the next moment based on the execution results of the previously executed actions, until the rescue is completed.

[0103] S142: Compare the magnitudes of multiple cumulative reward values ​​and determine the action space corresponding to the maximum cumulative reward value as the continuous scheduling result of the continuous dynamic ordered phase.

[0104] Specifically, for each set of state spaces and action spaces, the Q-learning method is used to solve the problem, and the cumulative reward value corresponding to each set of state spaces and action spaces is obtained according to the cumulative reward function. After obtaining multiple cumulative reward values, the values ​​are compared, and the action space corresponding to the largest cumulative reward value is determined as the result of the dynamic ordered stage rescue force scheduling.

[0105] For example, as shown in Table 2, the continuous scheduling results during the continuous dynamic and orderly phase are as follows: a total of 18 rescue forces to be dispatched were dispatched, including multiple patrol boats and professional rescue vessels. The arrival times of the rescue forces to be dispatched at the rescue site were distributed from 8:00 to 21:00. At the last moment, 1 to 2 rescue forces to be dispatched were dispatched for the final rescue. This can comprehensively consider the capacity of the rescue site and the rescue environment information to ensure the orderly allocation of the rescue forces to be dispatched and the rationality of the subsequent rescue forces to be dispatched.

[0106] Table 2. Continuous scheduling results during the continuous dynamic ordering phase

[0107] Thus, the large-scale maritime resource rescue scheduling method provided in this embodiment obtains a set of indicators for the rescue forces to be scheduled and rescue environment information. Based on the set of indicators for the rescue forces to be scheduled, a highest comprehensive scheduling score function and a minimum number of rescue forces to be scheduled function are constructed. Based on the highest comprehensive scheduling score function and the minimum number of rescue forces to be scheduled function, an initial scheduling model for the rescue forces to be scheduled in the initial rapid response phase is constructed using an approximation ideal solution ranking method, and the initial scheduling result for the initial rapid response phase is determined. Based on the initial scheduling result, the capacity of the rescue site, and the rescue environment information, a continuous scheduling model for the rescue forces to be scheduled in the continuous dynamic and orderly phase is constructed with the shortest rescue time as the objective. The continuous scheduling model is solved using the Q-learning method to obtain the continuous scheduling result for the continuous dynamic and orderly phase. In this way, by dividing the large-scale maritime resource rescue scheduling into two phases—the initial rapid response phase and the continuous dynamic and orderly phase—the rapid and reasonable scheduling of rescue forces to be scheduled can be achieved in the initial rapid response phase, thereby meeting the emergency rescue needs in the initial rapid response phase and effectively avoiding congestion and redundancy problems when the rescue forces to be scheduled are dispatched in the continuous dynamic and orderly phase. Subsequently, in the continuous and orderly phase, by comprehensively considering the capacity of the rescue site and information about the rescue environment, it is possible to ensure the orderly allocation of rescue forces to be dispatched, so that the entire rescue process meets the actual needs of "rapid response and continuous order" for large-scale maritime resource rescue and dispatch.

[0108] Example 2: Figure 3 This is a schematic diagram of a large-scale maritime resource rescue and dispatching device provided in an embodiment of the present invention, as shown below. Figure 3 As shown, it includes: acquisition module 10, function construction module 11, initial fast response phase solution module 12, continuous scheduling model construction module 13, and continuous dynamic ordered phase solution module 14.

[0109] Among them, the acquisition module 10 is used to acquire the set of indicators of the rescue forces to be dispatched and the rescue environment information.

[0110] Function construction module 11 is used to construct the highest comprehensive dispatch score function and the minimum number of dispatchable rescue forces based on the set of indicators of the rescue forces to be dispatched.

[0111] The initial rapid response phase solution module 12 is used to construct the initial scheduling model of the rescue forces to be scheduled in the initial rapid response phase based on the highest comprehensive scheduling score function and the minimum number of rescue forces to be scheduled, and to determine the initial scheduling results of the initial rapid response phase by using the approximation ideal solution sorting method.

[0112] The continuous scheduling model construction module 13 is used to construct a continuous scheduling model for the rescue forces to be dispatched in a continuous, dynamic, and orderly phase, based on the initial scheduling results, the capacity of the rescue site, and the rescue environment information, with the goal of minimizing the rescue time.

[0113] The continuous dynamic ordered phase solution module 14 is used to solve the continuous scheduling model using the Q-learning method to obtain the continuous scheduling results of the continuous dynamic ordered phase.

[0114] In this embodiment, the acquisition module obtains the set of indicators for the rescue forces to be dispatched and the rescue environment information. The function construction module constructs the highest comprehensive dispatch score function and the minimum number of rescue forces to be dispatched based on the set of indicators. The initial rapid response phase solution module, based on the highest comprehensive dispatch score function and the minimum number of rescue forces to be dispatched, constructs the initial dispatch model for the rescue forces to be dispatched in the initial rapid response phase using the approximation ideal solution ranking method, and determines the initial dispatch result for the initial rapid response phase. The continuous dispatch model construction module, based on the initial dispatch result, the capacity of the rescue site, and the rescue environment information, constructs a continuous dispatch model for the rescue forces to be dispatched in the continuous dynamic and orderly phase, with the shortest rescue time as the objective. The continuous dynamic and orderly phase solution module is used to solve the continuous dispatch model using the Q-learning method to obtain the continuous dispatch result for the continuous dynamic and orderly phase. Thus, by dividing large-scale maritime resource rescue dispatch into two phases—the initial rapid response phase and the continuous dynamic and orderly phase—the rapid and reasonable dispatch of rescue forces to be dispatched in the initial rapid response phase can be achieved, thereby fulfilling the emergency rescue needs in the initial rapid response phase and effectively avoiding congestion and redundancy problems when rescue forces to be dispatched are dispatched in the continuous dynamic and orderly phase. Subsequently, in the continuous and orderly phase, by comprehensively considering the capacity of the rescue site and information about the rescue environment, it is possible to ensure the orderly allocation of rescue forces to be dispatched, so that the entire rescue process meets the actual needs of "rapid response and continuous order" for large-scale maritime resource rescue and dispatch.

[0115] Example 3: The present invention also provides an electronic device, such as... Figure 4 As shown, the electronic device includes a processor 1101, a memory 1102, a bus 1103, and a computer program stored in the memory 1102 and executable on the processor 1101. The processor 1101 includes one or more processing cores. The memory 1102 is connected to the processor 1101 via the bus 1103. The memory 1102 is used to store program instructions. When the processor executes the computer program, it implements the steps in the above-described method embodiment of Embodiment 1 of the present invention.

[0116] Furthermore, as an executable solution, the electronic device can be a computer unit, which can be a desktop computer, laptop, handheld computer, cloud server, or other computing device. The computer unit may include, but is not limited to, a processor and memory. Those skilled in the art will understand that the above-described structure of the computer unit is merely an example and does not constitute a limitation on the computer unit. It may include more or fewer components, or combine certain components, or use different components. For example, the computer unit may also include input / output devices, network access devices, buses, etc., and this embodiment of the invention does not limit this.

[0117] Furthermore, as an executable solution, the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc. The processor is the control center of the computer unit, connecting various parts of the entire computer unit via various interfaces and lines.

[0118] The memory can be used to store the computer programs and / or modules. The processor implements various functions of the computer unit by running or executing the computer programs and / or modules stored in the memory and by calling data stored in the memory. The memory may mainly include a program storage area and a data storage area. The program storage area may store the operating system and at least one application program required for a function; the data storage area may store data created based on the use of the mobile phone, etc. In addition, the memory may include high-speed random access memory and may also include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital card (SD card), flash card, at least one disk storage device, flash memory device, or other volatile solid-state storage device.

[0119] Although the invention has been specifically shown and described in conjunction with preferred embodiments, those skilled in the art should understand that various changes in form and detail may be made to the invention without departing from the spirit and scope of the invention as defined in the appended claims, all of which shall be within the scope of protection of the invention.

Claims

1. A method for large-scale maritime resource rescue and dispatch, characterized in that, The method includes: Obtain the set of indicators for the rescue forces to be dispatched and information on the rescue environment; Based on the set of indicators of the rescue forces to be dispatched, construct the highest comprehensive dispatch score function and the minimum rescue forces to be dispatched function; Based on the highest comprehensive dispatch score function and the minimum number of rescue forces to be dispatched function, the initial dispatch model of the rescue forces to be dispatched in the initial rapid response phase is constructed by using the approximation ideal solution sorting method, and the initial dispatch result of the initial rapid response phase is determined. Based on the initial scheduling results, the capacity of the rescue site, and the rescue environment information, a continuous scheduling model for the rescue forces to be dispatched is constructed in a continuous, dynamic, and orderly phase, with the goal of minimizing the rescue time. The continuous scheduling model is solved using the Q-learning method to obtain the continuous scheduling results for the continuous dynamic ordered phase.

2. The method according to claim 1, characterized in that, The set of indicators for the rescue forces to be dispatched includes: wind resistance level, dispatch time, number of passengers, speed, and resource type. The process of constructing the highest comprehensive dispatch score function and the minimum number of rescue forces to be dispatched based on this set of indicators includes: The evaluation score is obtained by expert scoring between any two of the wind resistance level index, dispatch time index, passenger capacity index, speed index and resource type index. The judgment matrix corresponding to the set of rescue forces to be dispatched is determined based on multiple evaluation scores. Based on the judgment matrix, construct the highest comprehensive scheduling score function; Based on the wind resistance level index, scheduling time index, passenger capacity index, speed index, and resource type index, multiple initial rapid response phase constraints are determined, and the minimum required rescue force function is determined based on these multiple initial rapid response phase constraints.

3. The method according to claim 2, characterized in that, The step involves constructing an initial dispatch model for the rescue forces to be dispatched in the initial rapid response phase based on the highest comprehensive dispatch score function and the minimum number of rescue forces to be dispatched function, using an approximation of the ideal solution ranking method, and determining the initial dispatch results for the initial rapid response phase, including: Based on the highest comprehensive scheduling score function and the minimum number of rescue forces to be scheduled, determine the single-objective optimization proximity function; Based on the single-objective optimization proximity function, solve for the single-objective optimization proximity corresponding to the multiple initial scheduling results in the initial rapid response phase; Compare the magnitudes of the single-objective optimization proximity scores and determine the initial scheduling result corresponding to the maximum single-objective optimization proximity score as the initial scheduling result.

4. The method according to claim 3, characterized in that, The step of determining the single-objective optimal proximity function based on the highest comprehensive scheduling score function and the minimum number of rescue forces to be scheduled includes: The highest comprehensive scheduling score function is normalized to obtain the normalized highest comprehensive scheduling score function; The minimum number of rescue forces to be dispatched is normalized to obtain the normalized minimum number of rescue forces to be dispatched function. Based on the normalized highest comprehensive scheduling score function and the normalized minimum number of rescue forces to be scheduled function, construct the positive ideal solution and the negative ideal solution corresponding to each of the initial scheduling results; Based on the positive and negative ideal solutions, the single-objective optimization proximity function is determined.

5. The method according to claim 1, characterized in that, Based on the initial dispatch results, the capacity of the rescue site, and the rescue environment information, and with the goal of minimizing the rescue time, a continuous dispatch model is constructed for the rescue forces to be dispatched during the continuous, dynamic, and orderly phase. This includes: Based on the initial scheduling results, the capacity of the rescue site, and the rescue environment information, multiple continuous dynamic and orderly phase constraints are determined. Based on the multiple continuous dynamic orderly stage constraints, and with the goal of minimizing the rescue time, a continuous scheduling model for the rescue forces to be dispatched during the continuous dynamic orderly stage is constructed.

6. The method according to claim 5, characterized in that, Before solving the persistent scheduling model using the Q-learning method to obtain the persistent scheduling results for the persistent dynamic ordered phase, the method further includes: Based on the initial scheduling results, the capacity of the rescue site, the rescue environment information, and multiple continuous dynamic orderly stage constraints, multiple state space sets and the action space corresponding to each state space set are determined. The state space set includes: state spaces corresponding to multiple times, and each action space includes: rescue force execution actions corresponding to multiple times. Each rescue force execution action corresponds one-to-one with each state space. The state space at time t+1 is determined based on the execution results of the rescue force execution actions at time t. For each set of state spaces and action spaces, construct a cumulative reward function.

7. The method according to claim 6, characterized in that, The step of solving the persistent scheduling model using the Q-learning method to obtain the persistent scheduling results for the persistent dynamic ordered phase includes: For each set of state spaces and action spaces, the Q-learning method is used to solve the problem, and the cumulative reward value corresponding to each set of state spaces and action spaces is obtained according to the cumulative reward function. By comparing the magnitudes of multiple cumulative reward values, the action space corresponding to the maximum cumulative reward value is determined as the continuous scheduling result of the continuous dynamic ordered stage.

8. The method according to claim 7, characterized in that, The process of solving for each set of state spaces and action spaces using the Q-learning method includes: The state space is initialized based on the initial scheduling results, the capacity of the rescue site, and the rescue environment information. Based on the greed strategy, determine the action to be performed by the rescue force at the current moment among the multiple rescue actions; The execution of the rescue force action to be executed is performed, and the action value of the rescue force action to be executed is obtained according to the action value function. It is then determined whether the rescue is completed. If not, the execution is returned to determine the rescue force action to be executed at the current moment among multiple rescue force actions according to the greed strategy, and the state space at the next moment is updated according to the execution result at the current moment until the rescue is completed.

9. A large-scale maritime resource rescue and dispatch device, characterized in that, include: The acquisition module is used to acquire the set of indicators of the rescue forces to be dispatched and rescue environment information; The function construction module is used to construct the highest comprehensive dispatch score function and the minimum number of dispatchable rescue forces function based on the set of indicators of the rescue forces to be dispatched; The initial rapid response phase solution module is used to construct the initial scheduling model of the rescue forces to be scheduled in the initial rapid response phase based on the highest comprehensive scheduling score function and the minimum number of rescue forces to be scheduled, and to determine the initial scheduling result of the initial rapid response phase by using the approximation ideal solution sorting method. The continuous scheduling model construction module is used to construct a continuous scheduling model for the rescue forces to be scheduled in a continuous, dynamic, and orderly phase, based on the initial scheduling results, the capacity of the rescue site, and the rescue environment information, with the goal of minimizing the rescue time. The continuous dynamic ordered phase solution module is used to solve the continuous scheduling model using the Q-learning method to obtain the continuous scheduling results of the continuous dynamic ordered phase.

10. An electronic 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 large-scale maritime resource rescue and dispatch method according to any one of claims 1 to 8.

Images