CN-122022412-A - Multi-disaster emergency collaboration-oriented dynamic priority scheduling method

CN122022412ACN 122022412 ACN122022412 ACN 122022412ACN-122022412-A

Abstract

The invention relates to the technical field of emergency management informatization and intelligent cooperative scheduling control, in particular to a multi-disaster-oriented dynamic priority scheduling method for emergency cooperation, which comprises the steps of obtaining multi-modal perception data uploaded by a preset fixed perception node to generate modal incomplete degree representing a data missing state, inputting the multi-modal perception data into a preset physical evolution prediction model, calculating a data uncertainty penalty term of a task node, calculating a current global situation perception entropy based on the evolution emergency degree and the data uncertainty penalty term, judging whether the global situation perception entropy is larger than a preset perception entropy safety threshold to generate an absolute optimal scheduling instruction, and sending the robust suboptimal scheduling instruction or the absolute optimal scheduling instruction to a preset maneuvering execution node as a target scheduling instruction to update the global situation perception entropy in a global task distribution network.

Inventors

Liu nanyu
CHEN YIN
ZHANG JUN
JIANG HAO

Assignees

贵州华泰智远大数据服务有限公司

Dates

Publication Date: 20260512
Application Date: 20260414

Claims (10)

1. The dynamic priority scheduling method for multi-disaster emergency collaboration is characterized by comprising the following steps of: s1, acquiring multi-mode sensing data uploaded by a preset fixed sensing node, and analyzing the multi-mode sensing data to generate a mode defect degree representing a data missing state; s2, inputting the multi-mode perception data into a preset physical evolution prediction model to output the evolution emergency degree of a corresponding task node, and calculating a data uncertainty penalty term of the task node based on the mode incomplete degree; s3, calculating an initial priority score of the task node based on the evolution urgency and the data uncertainty penalty term, and calculating a current global situation awareness entropy based on a preset global task distribution network; S4, judging whether the global situation awareness entropy is larger than a preset awareness entropy safety threshold, if so, multiplying the initial priority score by a preset entropy suppression weight to generate a robust suboptimal scheduling instruction, and if so, directly based on the initial priority score to generate an absolute optimal scheduling instruction; S5, the robust suboptimal scheduling instruction or the absolute optimal scheduling instruction is used as a target scheduling instruction to be issued to a preset maneuvering executing node, and the global situation awareness entropy in the global task distribution network is updated based on the response state of the maneuvering executing node.
2. The dynamic priority scheduling method for multi-disaster emergency collaboration according to claim 1, wherein S1 comprises the following sub-steps: s11, receiving original multi-mode data acquired by the fixed sensing node in a preset time window; s12, comparing the original multi-mode data with a preset complete mode template to identify a missing data mode identifier; s13, acquiring preset penalty weights corresponding to the missing data mode identifiers, and adding the preset penalty weights corresponding to all the missing data mode identifiers to generate the mode incomplete degree.
3. The dynamic priority scheduling method for multi-disaster emergency collaboration according to claim 1, wherein the step S2 comprises the following sub-steps: S21, extracting an uploading time stamp and a current time stamp of the multi-mode sensing data; S22, calculating the time difference between the current time stamp and the uploading time stamp, and inputting the time difference into a preset exponential decay function to generate a communication delay decay factor; S23, multiplying the communication delay attenuation factor with the modal incomplete degree to generate the data uncertainty penalty term.
4. The dynamic priority scheduling method for multi-disaster emergency collaboration according to claim 1, wherein the step of calculating the initial priority score of the task node based on the evolutionary emergency degree and the data uncertainty penalty term in S3 includes the sub-steps of: S31, acquiring a preset emergency gain coefficient and a preset penalty attenuation coefficient; S32, multiplying the evolution emergency degree by the emergency degree gain coefficient to generate a first weighted value; s33, multiplying the data uncertainty penalty term by the penalty term attenuation coefficient to generate a second weighted value; s34, subtracting the second weighted value from the first weighted value to generate the initial priority score.
5. The dynamic priority scheduling method for multi-disaster emergency collaboration according to claim 1, wherein the step of calculating the current global situation awareness entropy based on the preset global task distribution network in S3 includes the following sub-steps: S35, acquiring current calculation coverage rate and historical blind area duration time of all task nodes in the global task distribution network; S36, calculating the local information confusion of each task node based on the current calculation coverage rate and the historical blind area duration; s37, calculating indexes of products of the local information chaos of all the task nodes and preset node weights, and summing all the indexes to generate the global situation awareness entropy.
6. The dynamic priority scheduling method for multi-disaster emergency coordination according to claim 1, wherein the step of generating a robust sub-optimal scheduling instruction in S4 comprises the following sub-steps: S41, acquiring a preset candidate scheduling scheme set; S42, when the global situation awareness entropy is larger than the awareness entropy safety threshold, removing schemes which cause the maneuver execution node to perform cross-region preemption from the candidate scheduling scheme set so as to generate a limited scheme subset; s43, selecting a target scheme with highest global network connectivity from the limited scheme subset, and taking the target scheme as the robust suboptimal scheduling instruction.
7. The dynamic priority scheduling method for multi-disaster emergency coordination according to claim 1, wherein the step of generating an absolute optimal scheduling instruction in S4 includes the following sub-steps: s44, selecting a task node with the highest initial priority score from the candidate scheduling scheme set when the global situation awareness entropy is smaller than or equal to the awareness entropy safety threshold; S45, taking the task node with the highest initial priority score as a target preemption node to generate the absolute optimal scheduling instruction.
8. The dynamic priority scheduling method for multi-disaster-oriented emergency collaboration according to claim 1, wherein the step of updating the global situation awareness entropy in the global task distribution network in S5 comprises the following sub-steps: s51, receiving an arrival confirmation signal returned by the maneuvering executing node and on-site real-time sensing data; S52, resetting the duration of a historical blind zone of a corresponding task node to zero based on the on-site real-time perception data; s53, recalculating and updating the global situation awareness entropy based on the duration of the history blind zone after resetting to zero.
9. The multi-disaster-oriented emergency collaboration dynamic priority scheduling method of claim 1, wherein the fixed sensing nodes comprise tower emergency sentry, the mobile execution nodes comprise emergency unmanned aerial vehicles, and the multi-mode sensing data comprise temperature data, image data and water level data.
10. The dynamic priority scheduling method for multi-disaster-oriented emergency coordination according to claim 1, wherein the physical evolution prediction model is used for predicting an evolution critical point of a secondary disaster chain, and the robust suboptimal scheduling instruction is used for avoiding that the task node is in a system calculation dead zone within a continuous preset time threshold.

Description

Multi-disaster emergency collaboration-oriented dynamic priority scheduling method Technical Field The invention relates to the technical field of emergency management informatization and intelligent cooperative scheduling control, in particular to a dynamic priority scheduling method for multi-disaster emergency cooperation. Background In a multi-disaster emergency cooperative scene, dynamic priority scheduling is a key for ensuring accurate resource release, a traditional emergency scheduling method generally relies on complete sensing data uploaded by a sensor, and an absolute optimal scheduling method is formulated by calculating the urgency degree of each task node so as to pursue response efficiency maximization; However, in a high-risk environment of cascade disasters caused by extreme weather, the existing scheduling mode faces the technical limitations that a communication link is easily damaged to generate space data faults, the damage degree of the asymmetrical interference to the data integrity is difficult to quantify by the traditional method, prediction deviation caused by data missing cannot be processed, disaster evolution has nonlinear characteristics, data reference weight is exponentially reduced along with time, the prior art lacks deep fusion of time delay and space defect to cause that a system is easily trapped by obsolete or false data to cause resource mismatch, when a huge gap exists in mobile resources, the internal information disorder of the system is greatly increased if preemptive scheduling aiming at a single-point extreme value is only executed, and the traditional scheduling logic cannot sense the paralysis risk of a global situation and is easily trapped into calculation deadlock or perception blind area; therefore, how to construct a dispatching framework with redundant fault tolerance and interference self-healing capacity under the double constraint of heterogeneous data loss and limited communication, and realize dynamic balance of local disaster response speed and global system perception stability, thereby becoming the problem to be solved in multi-disaster-oriented emergency cooperative dispatching. Disclosure of Invention The invention aims to provide a dynamic priority scheduling method for multi-disaster emergency collaboration, which solves the following technical problems: Scheduling decision deviation, false early warning decoy and system perception paralysis risk caused by bottom layer perception data deletion or communication lag under extreme disaster environment are avoided, and the vulnerability resistance and network connectivity of a scheduling system are maintained by dynamically balancing global resource allocation while the local disaster extreme response speed is ensured more easily. The aim of the invention can be achieved by the following technical scheme: A dynamic priority scheduling method for multi-disaster emergency collaboration comprises the following steps: s1, acquiring multi-mode sensing data uploaded by a preset fixed sensing node, and analyzing the multi-mode sensing data to generate a mode defect degree representing a data missing state; s2, inputting the multi-mode perception data into a preset physical evolution prediction model to output the evolution emergency degree of a corresponding task node, and calculating a data uncertainty penalty term of the task node based on the mode incomplete degree; s3, calculating an initial priority score of the task node based on the evolution urgency and the data uncertainty penalty term, and calculating a current global situation awareness entropy based on a preset global task distribution network; S4, judging whether the global situation awareness entropy is larger than a preset awareness entropy safety threshold, if so, multiplying the initial priority score by a preset entropy suppression weight to generate a robust suboptimal scheduling instruction, and if so, directly based on the initial priority score to generate an absolute optimal scheduling instruction; S5, the robust suboptimal scheduling instruction or the absolute optimal scheduling instruction is used as a target scheduling instruction to be issued to a preset maneuvering executing node, and the global situation awareness entropy in the global task distribution network is updated based on the response state of the maneuvering executing node. Further, S1 comprises the following sub-steps: s11, receiving original multi-mode data acquired by the fixed sensing node in a preset time window; s12, comparing the original multi-mode data with a preset complete mode template to identify a missing data mode identifier; s13, acquiring preset penalty weights corresponding to the missing data mode identifiers, and adding the preset penalty weights corresponding to all the missing data mode identifiers to generate the mode incomplete degree. Further, S2 comprises the following sub-steps: S21, extracting an uploading time stamp and a current time