Transportation hub reliability scheduling performance optimization method based on variable topology super network decomposition

Abstract

The application discloses a kind of based on the transport hub maintenance scheduling efficiency optimization method of deconstruction of variable topology super network, comprising the following steps: step one, the maintenance element involved in the current real-time maintenance scheduling scene of transport hub and its interrelated relationship mapping modeling is modeled into supergraph network model;Step two, the basic uncertainty characteristics of supergraph network CNN topological structure are characterized;Step three, according to the supergraph network model, the evaluation value of measurement index is calculated;Step four, the three indexes calculated are compared with the reasonable value interval range of historical scheduling result, and the rationality of current scheduling state is judged;Step five, based on the measurement index, the maintenance scheduling efficiency of transport hub is improved.The beneficial effects of the present application are that the execution efficiency, resource collaboration degree and dynamic robustness of maintenance scheduling can be effectively improved through super network modeling and quantitative analysis, and bottleneck positioning and optimization are realized.

Description

TECHNICAL FIELD

[0001] The present application relates to a maintenance scheduling optimization method, more particularly to a transfer hub maintenance scheduling efficiency optimization method based on variable topology super network decomposition. BACKGROUND

[0002] Airport, port and other transfer hubs as the core nodes of regional logistics network, its efficient operation is highly dependent on timely and reliable maintenance support. The daily maintenance of the transfer hub not only contains the field infrastructure, conventional equipment, special equipment, but also involves the temporary parking of transport carriers (aircraft, ships, etc.). Maintenance tasks usually need to be arranged periodically according to the actual needs of maintenance, such as daily, weekly, monthly, seasonal and annual time windows, or immediate dynamic response scheduling (equipment inspection, facility maintenance, equipment repair, fault finding, emergency repair, etc.). Maintenance work not only involves routine equipment repair, but also a series of highly specialized and technical complex tasks, covering mechanical, electrical, structural, energy, safety and other fields, so it needs to involve different maintenance resource scheduling, such as different professional technicians (different job responsibilities, professional fields, technical proficiency, professional qualifications, etc.), different professional maintenance tools (such as oscilloscope, logic analyzer, etc.).

[0003] In the daily operation of the transfer hub, maintenance scheduling has the characteristics of dynamic, multi-concurrent and high linkage. At present, although intelligent operation and maintenance monitoring and control, large model driven auxiliary decision making, special AI wisdom assistant and other means have been introduced to help improve the maintenance response and scheduling service capability, there are still problems such as low efficiency of multi-element cooperation, poor global allocation of resources, and insufficient response to dynamic disturbance, which have not been effectively solved. In view of the dynamic and uncertain maintenance requirements and task environment in the daily operation of the transfer hub, it is of great significance to reasonably model the multi-collaboration elements such as maintenance tasks and maintenance resources, identify and analyze the coupling relationship between elements in multiple dimensions, and quantitatively measure the expected execution efficiency and robustness of immediate maintenance scheduling arrangement from the perspective of global optimization, in order to optimize the maintenance process and scheduling arrangement, maximize the release of limited maintenance resources, and improve the maintenance service capability under the constraint of maintenance cost, and to ensure and improve the efficiency of daily operation of the transfer hub. SUMMARY

[0004] In view of the deficiencies of the prior art, the present application provides a transfer hub maintenance scheduling efficiency optimization method based on variable topology super network decomposition, which solves the problems of insufficient processing capacity, low scheduling efficiency, poor resource cooperation and weak robustness of the existing maintenance scheduling method in dynamic and uncertain task environment. The present application realizes real-time optimization of transfer hub maintenance scheduling through variable topology super network decomposition and quantitative analysis, and improves the overall efficiency.

[0005] To achieve the above object, the present application provides the following technical solutions:

[0006] A transport hub maintenance scheduling performance optimization method based on variable topology super network decomposition, comprising the following steps: step one, mapping and modeling the maintenance elements involved in the current real-time maintenance scheduling scene of the transport hub and the correlation therebetween into a supergraph network model; step two, characterizing two basic uncertain features of the supergraph network CNN topology structure under a dynamic and uncertain maintenance task environment; step three, based on the variable topology super network model mapped and constructed for the current real-time maintenance scheduling scene in steps one and two, fusing network feature quantitative analysis, and combining the multi-coupling maintenance task concurrent collaborative response scheduling situation characteristics of the transport hub, calculating the expected execution performance, resource collaboration degree and dynamic robustness of the current real-time maintenance scheduling in three aspects, and outputting the quantitative measurement analysis results; step four, judging the rationality of the current scheduling state by comparing the evaluation values of the expected execution performance, resource collaboration degree and dynamic robustness calculated for the current real-time maintenance scheduling with the reasonable value interval range of the historical scheduling results; step five, based on the expected execution performance, resource collaboration degree and dynamic robustness obtained in step four, first positioning the maintenance scheduling bottleneck, then optimizing through dynamic scheduling, an integrated information platform and a hierarchical emergency plan, and finally iterating and verifying to realize the maintenance scheduling performance improvement of the transport hub.

[0007] As a further improvement of the present application, the supergraph network model in step one comprises: a maintenance resource sub-network for integrating the current schedulable maintenance resources, explicitly defining the historical collaboration correlation among the resources, and providing a resource pool and collaboration basis for resource optimization configuration; a maintenance task sub-network for sorting out the internal relationship of the current concurrent maintenance tasks, explicitly defining the task importance, dependency order and execution mode, and providing decision rules for task scheduling and resource inclination; and an inter-layer correlation network for connecting the maintenance resource sub-network and the maintenance task sub-network to explicitly define the boundary and reliability of the task undertaken by the resources, and provide a basis for accurate matching of resources and tasks.

[0008] As a further improvement of the present application, the specific steps of characterizing the two basic uncertain features of the supergraph network CNN topology structure under a dynamic and uncertain maintenance task environment in step two are as follows: step two one, modeling and characterizing the uncertainty feature of the node-to-node edge weight disturbance; and step two two, modeling and characterizing the uncertainty feature of the task node probability failure.

[0009] As a further improvement of the present application, the specific way of modeling and characterizing the uncertainty feature of the node-to-node edge weight disturbance in step two one is as follows: considering the dynamic and uncertain maintenance task environment and the fuzziness of the quantitative evaluation value, for an associated edge s in the edge weight matrix Qa m×m , Da m×m , and Ea n×m ,ij The attribute value represents the range of edge weight changes defined as [ , We use methods such as estimates and empirical values ​​to determine the range of this edge weight perturbation.

[0010] As a further improvement of the present invention, the specific method for modeling and representing the uncertainty characteristics of the probabilistic failure of task nodes in step two is as follows: Considering the uncertainty of the probabilistic failure of task nodes under dynamic and uncertain maintenance task environment, the representation is defined as follows: when the maintenance resource node rn l Undertake and participate in maintenance tasks (task tn) i At that time, task node tn i The failure probability is tfp i .

[0011] As a further improvement of the present invention, the specific steps for calculating the expected execution efficiency, resource cooperation degree, and dynamic robustness of the current real-time maintenance scheduling in step three, and outputting the quantitative measurement analysis results are as follows: Step three-one, quantitative measurement of expected execution efficiency, which is measured by combining the probability rework impact coefficient and the execution efficiency of the critical chain in the current maintenance task sub-network; Step three-two, quantitative measurement of resource cooperation, which is measured by combining the corrected weighting degree of the maintenance resource sub-network and the expected reliability of resources; Step three-three, quantitative measurement of dynamic robustness, which is measured by combining the topology change response performance under CNN node failure and the attribute change response performance under edge weight perturbation.

[0012] The beneficial effects of this invention are:

[0013] This invention addresses the shortcomings of existing technologies in handling the dynamic uncertainties, high concurrency, and high interconnectivity of daily maintenance and scheduling in transit hubs, as well as the lack of quantitative evaluation methods for global scheduling effectiveness. It features support for real-world mapping of maintenance and scheduling elements through variable topology hypernetwork modeling, support for the expected execution efficiency of real-time maintenance and scheduling arrangements, and robust global optimization perspective quantitative measurement. It can effectively help optimize maintenance processes and scheduling arrangements, maximize the release of limited maintenance resources and maintenance service capabilities under cost constraints, and ensure and improve the daily operational efficiency of transit hubs. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the CCN network structure in the hypernetwork model.

[0015] Figure 2 This is a schematic diagram illustrating the topological uncertainty of the CCN (Cyclic Network Network) model. Detailed Implementation

[0016] According to the specific implementation method described in the claim, the content format of the specific implementation method is as follows:

[0017] The proposed method for optimizing the maintenance and scheduling efficiency of transport hubs based on variable topology hypernetwork deconstruction comprises four main steps:

[0018] (1) The maintenance elements and their interrelationships involved in the current real-time maintenance scheduling scenario of the transfer hub are mapped and modeled as a hypergraph network model:

[0019] The maintenance resources, maintenance tasks, and other maintenance elements involved in the current real-time maintenance scheduling scenario of the transfer hub, as well as the relationships between them, are represented and modeled as a multi-modal, multi-layer hypernetwork model: CCN=(RN,TN,NRR), where:

[0020] RN=(Rn,Sc) represents the maintenance resource subnetwork, where the subnetwork node set Rn=(rn1,rn2,…,rn n Let be the n maintenance resource nodes currently being scheduled; Sc be the edge weight matrix. n×n SC ij The value is the maintenance resource node rn in the sub-network. i and rn j The historical collaboration associated edge attribute value between them is taken as sc when associated edges exist. ij =1, take sc when there are no associated edges. ij =0.

[0021] TN=(Tn,Qa,Da,Es) represents the maintenance task sub-network, where the sub-network node set Tn=(tn1,tn2,…,tn m Let m be the m maintenance task nodes currently being concurrently scheduled; and let Qa be the edge weight matrix. m×m Da m×m Zhongqa ij and da ij The values ​​are respectively for task node tn i and tn j Quality-efficiency related edge attribute values ​​and execution-dependency related edge attribute values, qa ij and da ij The value can be obtained by assessing the degree of joint impact of the two tasks on the relevant aspects of the normal operation of the transfer hub, and the degree of interactive dependency during execution; the edge weight matrix Es m×m es ij The value is the attribute value of the execution sequence and parallel relationship between tasks, es ij A value of 1 indicates task tn i With tn j There is a serial-parallel execution relationship between them and tn i For tn j The immediate preceding task, if there is no serial-parallel association, take sc. ij =0.

[0022] NRR= (Ea, Ir), representing the association relationship between the nodes of the sub-network RN and TN layers, wherein the attribute value of the edge weight matrix Ea n×m The value of ea ij is the execution reliability associated edge between the resource node rn i and the task node tn j , which can be obtained by comprehensively evaluating its competence in executing similar tasks in historical records; the value of ir n×m in the edge weight matrix Ir ij is the attribute value of the execution association edge between the resource node rn i and the task node tn j , which is sc ij = 1 when the associated edge exists, and sc ij = 0 when the associated edge does not exist.

[0023] (2) Two basic uncertain characteristics of the supergraph network CNN topology under dynamic and uncertain maintenance task environment are represented:

[0024] The uncertainty characteristics of the edge weight disturbance between nodes are modeled and represented: considering the dynamic and uncertain maintenance task environment and the fuzziness of the quantitative evaluation value, for the attribute value of an associated edge s m×m in the edge weight matrix Qa m×m , Da n×m , Ea ij , the edge weight change interval is defined as [ , ], and the range of such edge weight disturbance can be determined by estimation value, experience value, etc.

[0025] The uncertainty characteristics of the task node probability failure are modeled and represented: considering the uncertainty of the task node probability failure under the dynamic and uncertain maintenance task environment, the failure probability tfp l of the task node tn i is defined when the maintenance resource node rn i participates in the maintenance task tn i , which can be calculated by the following formula:

[0026] [1]

[0027] [2]

[0028] Where Sm i,l is the skill requirement matching satisfaction degree of the resource node rn l and the task node tn i ; Qsp i,l is the resource node rn lHistorical mission records and mission tn i The overall probability of completing similar task experiences; e is the number of similar task experiences included in the evaluation statistics; For task tn i The similarity to the k-th task among e similar task items; Ev represents the completion satisfaction of the k-th task among e similar task items; T For task tn i The satisfaction threshold corresponding to achieving the requirements.

[0029] (3) Based on the variable topology hypernetwork model constructed in steps (1) and (2) for the current real-time maintenance scheduling scenario, integrate network feature quantitative analysis, and combine the characteristics of concurrent collaborative response scheduling of multi-coupled maintenance tasks in the transfer hub to calculate the expected execution efficiency of the current real-time maintenance scheduling. Resource collaboration Dynamic robustness Three metrics are used to output quantitative metric analysis results.

[0030] (3-1) Expected Execution Effectiveness Quantitative measurement

[0031] Expected Execution Performance By comprehensively considering the probability rework impact coefficient in the current maintenance task subnetwork and the execution efficiency of the critical chain To measure, calculate using the following formula:

[0032] [3]

[0033] [4]

[0034] [5]

[0035] in: For upstream task tn i The impact of probabilistic iterative rework on related downstream tasks tn j Average expected rework time; Tet j The expected task duration assuming no rework occurs; Irr j For the affected downstream task tn j The impact of rework is relative to, taking In j To communicate with downstream task tn j Total number of upstream tasks with execution dependencies; TCT TIN Tct is the number of closed loops in the current task sequence. i Let I be the number of tasks in the closed loop of the i-th task;i The execution of the ith task closed loop depends on the correlation strength evaluation value; c is the number of tasks contained in the longest critical task chain with the sum of expected average completion time; tw i The task node tn i The comprehensive importance degree in the current task network is measured by Calculation; The node importance degree of the node in the perspective of Da correlation edge is measured by PageRank analysis; The node importance degree of the node in the perspective of Qa correlation edge is measured by calculating the network node degree centrality; wtt i The independent pre-determination value of the task node importance degree.

[0036] (3-2) Resource cooperation degree Quantitative measurement

[0037] Resource cooperation degree It is measured by the comprehensive maintenance resource sub-network correction weighting degree And the expected reliability of resources Drr RN , which is calculated by the following formula:

[0038] [6]

[0039] [7]

[0040] [8]

[0041] Where, rw i is the initial weight of the resource node, which takes the weight tw i of the task node participated by the resource node. The correction weighting degree of the resource node rn i , which takes .

[0042] (3-3) Dynamic robustness Quantitative measurement

[0043] Dynamic robustness It is measured by the comprehensive super-network CNN node failure topology change response performance And the edge weight disturbance attribute change response performance , which is calculated by the following formula:

[0044] [9]

[0045]

[10]

[0046]

[11]

[0047] wherein, is the expected average value of the expected execution performance of the current task network under different edge weight value combinations within the perturbation variation interval of each edge weight attribute value , is the variance thereof, both of which can be calculated by means of random simulation.

[0048] (4) According to the expected execution performance , resource cooperation degree , and dynamic robustness of the current real-time maintenance scheduling conversion calculation, the evaluation values of the three aspects of measurement indicators are compared with the reasonable value interval range of the historical scheduling results to determine the rationality of the current scheduling state.

[0049] (5) Based on the expected execution performance , resource cooperation degree , and dynamic robustness obtained above, the maintenance scheduling bottleneck is first located, and then optimized through dynamic scheduling, an integrated information platform, and hierarchical emergency plans, and finally iteratively verified to realize the optimization and improvement of the maintenance scheduling efficiency of the transfer hub.

[0050] The following examples are provided in this embodiment:

[0051] For a transfer hub A, the real-time maintenance scheduling scenario at time t is analyzed, the maintenance resources, maintenance tasks and other maintenance elements involved and their correlation are sorted out, and are mapped, converted and modeled into a corresponding variable topology super network model: CCN=(RN,TN,NRR), wherein the maintenance task subnetwork includes 6 maintenance task nodes Tn=(tn1,tn2,tn3,tn4,tn5,tn6), the maintenance resource subnetwork includes 9 maintenance resource nodes RN=(rn1,rn2,rn3,rn4,rn5,rn6,rn7,rn8,rn9), and the edge weight matrix Sc 9×9 , Qa 6×6 , Da 6×6 , Es 6×6 , Ea 9×6 , Ir 9×6 between the network nodes is quantitatively converted and evaluated, wherein the uncertain topology edge weight perturbation is considered, and the attribute values in the matrices Qa 6×6 , Da 6×6 , Ea 9×6 are taken as interval range values.

[0052]

[0053] ​

[0054]

[0055]

[0056] Based on the degree to which the skill requirements of each resource node with an execution-related edge are matched with those of the maintenance task node, a quantitative matrix Sm is obtained through evaluation. 9×6 Based on the historical operation and maintenance data of the transfer hub, the top 10 (e=10) similar task experiences with the highest similarity to the maintenance tasks performed in the historical records of each resource node were selected as the evaluation criteria. The comprehensive completion probability matrix Qsp of similar tasks was calculated according to Formula 2. 9×6 Then, by substituting into Formula 1, the failure probability values ​​of each task node under the current scheduling configuration are calculated as tfp1=0.13, tfp2=0.21, tfp3=0.22, tfp4=0.16, tfp5=0.18, and tfp6=0.25.

[0057] The pre-determined importance values ​​for each task node were obtained from the evaluation: wtt1, wtt2, wtt3, wtt4, wtt5, and wtt6. PageRank analysis was then used to measure the importance of each task node in Da... 6×6 Node importance values ​​from the perspective of associated edges , , , , , The degree centrality of network nodes is used to determine the degree centrality of each task node in Qa. 6×6 Node importance values ​​from the perspective of associated edges , , , , , Then, substitute into the formula. The calculated overall importance of each task node in the current task network is tw1=0.19, tw2=0.25, tw3=0.18, tw4=0.15, tw5=0.11, and tw6=0.12.

[0058] The expected execution cycle times for each task, Tet1, Tet2, Tet3, Tet4, Tet5, and Tet6, were obtained under the condition that no rework events occur. Based on matrix Qa... 6×6 Da 6×6 Es 6×6 Based on the information on the quality and efficiency correlation / degree, execution dependency / degree, and serial / parallel relationships between tasks, an average expected value matrix of rework time for each task is obtained. Then bring into the formula The rework influence ratio Irr1, Irr2, Irr3, Irr4, Irr5, Irr6 of each task is calculated. Combined with Es 6×6 and information, the number of tasks c contained in the critical chain with the longest expected average completion time is determined, and then the execution efficiency value of the critical chain is calculated according to formula 4 . Combined with Da 6×6 and Es 6×6 information, the number of task loops TCT TIN in the current task sequence and the number of tasks and execution dependency correlation strength evaluation value in each task loop are determined, and then the probability rework influence coefficient in the current maintenance task sub-network is calculated according to formula 5 . Then the expected execution performance quantitative measurement value under the current maintenance scheduling configuration is calculated according to formula 3 = 36.84.

[0059] Let the initial importance value of each resource node be equal to the weight of the task node participated by the resource node (take the maximum value if multiple tasks are participated), and the initial weight value rw1, rw2, rw3, rw4, rw5, rw6 of each resource node is obtained. According to formula , the modified weight degree of each resource node is calculated, and the modified weight degree of the maintenance resource sub-network is calculated according to formula 7 , and then the resource expected reliability Drr 9×6 is calculated according to Ea RN and formula 8, and the resource cooperation quantitative measurement value is calculated according to formula 6 = 0.34.

[0060] According to formula 10, the topology change response performance measurement value under the super network CNN node failure is calculated . According to the attribute interval range value in the edge weight matrix Qa 6×6 , Da 6×6 , Ea 9×6 , the value of and is calculated by means of random simulation method, and the attribute change response performance measurement value under edge weight disturbance is calculated according to formula 11 . Then the values of and are brought into formula 9, and the dynamic robustness quantitative measurement value is calculated = 15.73.

[0061] The expected execution performance calculated by the above-mentioned conversion of the current real-time maintenance scheduling , resource cooperation degree , dynamic robustness The three aspects of the evaluation value of the quantitative index can be judged according to the reasonable value interval range of the historical scheduling result, the rationality of the current scheduling state, or the execution sequence of the maintenance task or the maintenance resource configuration scheme is adjusted, and then the three quantitative indexes are calculated and evaluated again, compared and judged, and then the optimization of the maintenance task is realized according to the judgment result.

[0062] As described above, the transport hub maintenance scheduling efficiency optimization method based on the variable topology super network decomposition of the embodiment can effectively help to optimize the maintenance process and scheduling arrangement through the evaluation method.

[0063] The above only describes the preferred embodiments of the present application, and the protection scope of the present application is not limited to the above-mentioned embodiments. Any technical solution falling within the concept of the present application shall fall within the protection scope of the present application. It should be noted that, for ordinary skilled persons in the art, some improvements and refinements without departing from the principles of the present application shall also be considered as the protection scope of the present application.

Claims

1. A method for optimizing the performance of a transport hub based on variable topology hypernetwork deconstruction, characterized in that: Comprising the following steps: Step one, map the involved maintenance elements and their inter-relationships in the current real-time maintenance scheduling scenario of the transfer hub to a hypergraph network model; Step two, characterize two basic uncertain features of the hypergraph network CNN topology structure under dynamic and uncertain maintenance task environment; Step three, based on the variable topology hypernetwork model mapped and constructed in step one and step two for the current real-time maintenance scheduling scenario, fuse network feature quantitative analysis, and combined with the characteristics of concurrent collaborative response scheduling of multi-coupling maintenance tasks in the transfer hub, calculate the expected execution efficiency, resource collaboration degree, and dynamic robustness of the current real-time maintenance scheduling in three aspects of the measurement index, and output the quantitative measurement analysis result; Step four, according to the expected execution efficiency, resource collaboration degree, and dynamic robustness of the three aspects of the measurement index evaluation value calculated for the current real-time maintenance scheduling, compare with the reasonable value interval range of the historical scheduling result, judge the rationality of the current scheduling state; Step five, based on the expected execution efficiency, resource collaboration degree, and dynamic robustness obtained in step four, first locate the maintenance scheduling bottleneck, then optimize through dynamic scheduling, integrated information platform, and hierarchical emergency plan, and finally iterate and verify to realize the efficiency improvement of the maintenance scheduling of the transfer hub.

2. The method of claim 1, wherein the method is based on the deconstruction of a variable topology supernetwork. The hypergraph network model in step one comprises: A maintenance resource subnetwork for integrating current schedulable maintenance resources, clarifying historical collaboration relationships between resources, and providing a "resource pool" and collaboration basis for resource optimization configuration; A maintenance task subnetwork for clarifying the internal relationship of current concurrent maintenance tasks, clarifying task importance, dependency order, and execution mode, and providing decision rules for task scheduling and resource inclination; An interlayer association relationship network for connecting the maintenance resource subnetwork and the maintenance task subnetwork to clarify the boundary and reliability of resource task acceptance, and provide basis for resource-task precise matching.

3. The method of claim 1 or 2, wherein the method is based on the deconstruction of a variable topology supernetwork. The specific steps of characterizing two basic uncertain features of the hypergraph network CNN topology structure under dynamic and uncertain maintenance task environment in step two are as follows: Step two one, model and characterize the uncertainty feature of the node inter-edge weight disturbance; Step two two, model and characterize the uncertainty feature of the task node probability failure.

4. The method of claim 3, wherein the method is characterized by: The specific way of modeling the uncertainty feature of the task node probability failure in step two is as follows: considering the uncertainty of the task node probability failure under the dynamic and uncertain maintenance task environment, the definition is that when the maintenance resource node rn l participates in the maintenance task tn i , the failure probability of the task node tn i is tfp i .

5. The method of claim 1 or 2, wherein the method is based on the deconstruction of a variable topology supernetwork. The specific steps of calculating the expected execution efficiency, resource collaboration degree, and dynamic robustness of the current real-time maintenance scheduling in three aspects of the measurement index, and outputting the quantitative measurement analysis result in step three are as follows: Step three one, expected execution efficiency quantitative measurement, which measures by comprehensively considering the probability rework influence coefficient in the current maintenance task subnetwork and the execution efficiency of the critical chain; Step three two, resource collaboration quantitative measurement, which measures by comprehensively considering the revised weighted degree of the maintenance resource subnetwork and the expected reliability of the resource; Step three three, dynamic robustness quantitative measurement, which measures by comprehensively considering the topology change response performance under hypernetwork CNN node failure and the attribute change response performance under edge weight disturbance.

Images