A reliable reverse unloading method based on multimodal transport participants scenario

Abstract

The present invention relates to the field of intelligent traffic control technology, and discloses a reliable reverse unloading method based on a multimodal transport participant scenario, comprising the following steps: constructing a multimodal transport reverse unloading system model to obtain reverse unloading of traffic participants and edge computing servers; based on the multimodal transport reverse unloading system model, obtaining the honesty, coordination, task achievement rate, and computing resource and time availability of traffic participants, and calculating the reliability of traffic participants; calculating the efficiency of traffic participants based on the reliability of traffic participants; based on the greedy method, evaluating the reliability and computing power of traffic participants based on the reliability and efficiency of traffic participants, and selecting the optimal traffic participant for reverse unloading; the method promotes resource coordination between multimodal transport vehicles, significantly improves the diversity and integrity of the system, can share loads and reduce delays, so as to quickly and comprehensively evaluate the performance of each traffic participant node.

Description

Technical Field

[0001] The present invention relates to the field of intelligent traffic control technology, and in particular to a reliable reverse unloading method based on a multimodal transport participant scenario. Background Art

[0002] Intelligent Transportation Systems (ITS) are highly autonomous, next-generation transportation systems characterized by self-perception, self-adaptation, self-learning, and self-organization. The rapid development of cooperative vehicle-infrastructure systems (CVIS) and vehicle edge computing (VEC) has laid the foundation for the edge computing paradigm of VEC. However, due to computational limitations and the increasing number of traffic participants, VEC cannot complete all computations before deadlines. Offloading computation tasks to the cloud incurs significant latency. To reduce latency and balance computational load within the system, existing technologies have proposed methods for offloading computation tasks to traffic participants and various approaches for selecting task nodes. While these approaches effectively address the issue of offloading, most focus on a single network architecture, such as the Internet of Vehicles (IoV), while ignoring the complexity of a wider range of traffic participants. Furthermore, existing load distribution strategies lack dynamic adaptability, resulting in uneven load distribution. Some nodes may be overloaded while others remain idle, failing to fully utilize all system resources. Summary of the Invention

[0003] In response to the above-mentioned deficiencies in the prior art, the present invention provides a reliable reverse unloading method based on the scenario of multimodal transport participants. By introducing a reverse unloading mechanism, that is, when the edge server is overloaded, the computing tasks are dynamically unloaded to local traffic participants to share the load, thereby solving the problems of uneven load distribution and low system resource utilization.

[0004] In order to achieve the above-mentioned object of the invention, the technical solution adopted by the present invention is:

[0005] A reliable reverse unloading method based on a multimodal transport participant scenario includes the following steps:

[0006] S1. Construct a multimodal reverse unloading system model that includes a set of traffic participants, edge devices, and cloud servers, and obtain reverse unloading between traffic participants and edge computing servers;

[0007] S2. Based on the constructed multimodal reverse unloading system model, the honesty, coordination, task completion rate, computing resources and time availability of traffic participants are obtained, and the reliability of traffic participants is calculated;

[0008] S3. Calculate the efficiency of the traffic participants based on the calculated reliability of the traffic participants;

[0009] S4. Based on the greedy method, the reliability and computing power of traffic participants are evaluated according to the calculated reliability and efficiency of traffic participants, so as to select the optimal traffic participant for reverse unloading.

[0010] Furthermore, step S1 specifically includes:

[0011] S11. Construct a multimodal reverse unloading system model that includes a set of traffic participants, edge devices, and cloud servers. The traffic participants include cars, buses, and trains. Cars are The bus is Train for Indicates the i v cars, L represents the number of cars, V b,ib Indicates the i b buses, B represents the number of buses, Indicates the i t′ trains, N represents the number of trains;

[0012] S12. All traffic participants in the traffic participant set communicate with the edge computing node close to the edge device and have the computing capacity of local non-offload tasks. At the same time, all traffic participants communicate within the Internet of Vehicles or the public network system from the train to the ground, upload data and trigger computing tasks at the edge. The edge device is connected to the cloud server and communicates with each traffic participant. It is responsible for scheduling all edge devices and completing the offloading of cross-modal computing tasks.

[0013] Furthermore, step S2 specifically includes:

[0014] S21. Calculate the honesty of each traffic participant in the set of traffic participants, that is:

[0015]

[0016] Among them, h represents honesty, q represents the set of traffic participants, v represents cars, b represents buses, t′ represents trains, M represents the number of communication rounds, P q,i,h represents the honesty of the i-th traffic participant, represents the honest information sent by the i-th traffic participant in the m-th round of communication, represents the number of messages sent by the i-th traffic participant in the m-th round of communication;

[0017] S22. Calculate the coordination of each traffic participant in the traffic participant set, that is:

[0018]

[0019] in, represents the empty set, c represents the synergy, V represents the total number of vehicles, which is the sum of the number of cars and buses, P q,i,c represents the cooperation of the i-th traffic participant, represents the total number of vehicles participating in the collaboration in the mth round of communication;

[0020] S23. Calculate the task achievement rate of each traffic participant in the traffic participant set, that is:

[0021]

[0022] Among them, r represents the task achievement rate, P q,i,r represents the task arrival rate of the i-th traffic participant, represents the number of successful arrivals of the i-th traffic participant, represents the number of arrival failures of the i-th traffic participant;

[0023] S24. Calculate the computing resources of the traffic participants in the traffic participant set, namely:

[0024]

[0025] Among them, Γ q,i represents the computing resources of the i-th traffic participant, a i represents the discount factor of the computing resources of the i-th traffic participant, represents the CPU frequency of reverse unloading of the i-th traffic participant;

[0026] S25. Calculate the time availability of cars and buses in the set of traffic participants, that is:

[0027]

[0028] in, represents the time availability of the i-th traffic participant, which is a car or a bus, t ave represents the average usage time of the car-bus link, represents the time that the i-th traffic participant is connected between the car and the bus;

[0029] S26. Calculate the time availability of the train as a transport participant in the transport participant set, namely:

[0030]

[0031] in, Indicates the i t′ The time availability of trains, t pre Indicates the estimated time the train will stay at the site. Indicates the i t′ The time that each train link has been used;

[0032] S27. Calculate the reliability of traffic participants based on their honesty, collaboration, and task completion rate, namely:

[0033]

[0034] in, represents the reliability of the i-th traffic participant, α q,e represents the trust value weight, if represents the judgment condition, P th represents the honesty threshold, and e represents an optional variable.

[0035] Furthermore, step S3 specifically includes:

[0036] S31. Based on the calculated reliability of traffic participants, define the reliability matrix D, utility matrix C′, and availability matrix A′ as follows:

[0037]

[0038] Among them, d 11 represents the probability of a traffic participant changing from state 1 to state 1, d 12 represents the probability of a traffic participant changing from state 1 to state 2, d 21 represents the probability of a traffic participant changing from state 2 to state 1, d 22 represents the probability that a traffic participant changes from state 2 to state 2, c1′ and c2′ represent capability vectors, a1′ represents the probability that a traffic participant has no computational task at any time, a2′ represents the probability that a traffic participant is engaged in a computational task, and T represents transposition;

[0039] S32. Define the local task rate of the intermodal reverse unloading system model as λ, define the task completion rate of the intermodal reverse unloading system model as μ, and calculate the reliability matrix D, that is:

[0040]

[0041] Where exp(·) represents the natural exponential function, l represents the size of the calculated data, represents the computational workload of reverse unloading of the i-th traffic participant;

[0042] S33. Calculate the utility matrix C′, that is:

[0043]

[0044] S34. Calculate the efficiency of traffic participants based on the reliability matrix D, the utility matrix C′, and the availability matrix A′, namely:

[0045] E=A′ T DC′

[0046] Among them, E represents efficiency.

[0047] Furthermore, step S4 specifically includes:

[0048] S41, the number of communication rounds M, the number of trains N, the number of buses B, the number of cars L, the number of malicious vehicles, the number of unloaded tasks, and the CPU frequency Calculate data size l and discount factor a i , The train is expected to stay at the site for t pre , the average usage time t of the car and bus link ave , calculation workload Local task rate λ, task completion rate μ, honesty threshold P th And the trust value weight α q,e As input to the greedy method;

[0049] S42. Traverse each vehicle in the set of traffic participants, calculate the reliability and efficiency of each vehicle, and store them in the corresponding reliability matrix and utility matrix. The specific process is as follows:

[0050] Calculate the reliability of each vehicle and store it in the reliability matrix;

[0051] Calculate the efficiency of each vehicle and store it in the efficiency matrix;

[0052] S43: Set the selected vehicle to be empty and the best comprehensive score to negative infinity, and traverse the vehicles in the set of traffic participants. The specific process is as follows:

[0053] Determine whether the vehicle's reliability is less than the optimal comprehensive score. If so, exit the traversal. Otherwise, if the vehicle's reliability is greater than the optimal comprehensive score, assign the vehicle to the selected vehicle and assign the vehicle's reliability to the optimal comprehensive score. If the vehicle's reliability is equal to the optimal comprehensive score and the vehicle's efficiency is greater than the efficiency of the selected vehicle, assign the vehicle to the selected vehicle.

[0054] S44, obtaining a selected vehicle;

[0055] S45: The selected vehicle is regarded as the optimal traffic participant, so as to select the optimal traffic participant for reverse unloading.

[0056] Furthermore, the specific process of selecting the best traffic participant for reverse unloading in step S45 is as follows:

[0057] First, traffic participants upload data and generate computationally intensive tasks in the multimodal reverse unloading system model. When the computational load is high, edge computing nodes close to edge devices are prompted to generate unloading requests.

[0058] Secondly, the edge device uses the greedy method to determine whether reverse offloading is feasible based on the connected traffic participants. If so, it calculates the offloading process. Otherwise, if the connected traffic participants are participating in other computing tasks or no trustworthy nodes are found, it sends a cross-node offloading request to the cloud server;

[0059] Then, the cloud server selects the optimal traffic participants to participate in the computational task offloading based on the greedy method;

[0060] Finally, the optimal traffic participant processes the computation task and returns the offloaded results to the edge.

[0061] The present invention has the following beneficial effects:

[0062] 1. This paper proposes a reliable reverse unloading method based on a multimodal transport participant scenario, and constructs a multimodal transport reverse unloading system model. This model is applicable not only to a single type of transportation network, but also to a variety of transportation participants (such as cars, trains, and buses), thereby expanding the system's scope of application.

[0063] 2. The constructed multimodal reverse unloading system model significantly improves the diversity and overall performance of the system by promoting resource coordination among multimodal transportation vehicles;

[0064] 3. The intermodal reverse offloading system model can dynamically offload computing tasks to local traffic participants when edge servers are overloaded, thereby sharing the load and reducing latency.

[0065] 4. When selecting an unloading node, the present invention can comprehensively consider the historical communication behavior and current link status of traffic participants, and select the optimal traffic participant for reverse unloading based on the reliability and efficiency of the traffic participants, thereby quickly and comprehensively evaluating the performance of each traffic participant node. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 This is a flow chart of a reliable reverse unloading method based on a multimodal transport participant scenario proposed by the present invention;

[0067] Figure 2 This is a schematic diagram of the principle of the multimodal transport reverse unloading system model;

[0068] Figure 3 Schematic diagram showing the effect of the number of communication rounds on the number of successes in the method proposed in the present invention;

[0069] Figure 4 Schematic diagram showing the effect of the number of communication rounds on the number of successes in the traditional random reverse offloading method;

[0070] Figure 5 This is a schematic diagram showing the effect of the number of communication rounds on the number of successes when the method proposed in the present invention removes efficiency;

[0071] Figure 6 Schematic diagram showing the effect of the number of communication rounds on the number of successes when the reliability is removed in the method proposed in the present invention;

[0072] Figure 7 This is a schematic diagram showing the effect of the number of communication rounds on the number of successes under a larger transportation scale according to the method proposed in the present invention;

[0073] Figure 8 Schematic diagram showing the effect of the number of communication rounds on the number of successes for the traditional random reverse unloading method under a larger transport scale;

[0074] Figure 9 Schematic diagram of the effect of the number of communication rounds for multimodal transport participants on the number of successes in the method proposed by the present invention. DETAILED DESCRIPTION

[0075] The specific embodiments of the present invention are described below to facilitate understanding of the present invention by those skilled in the art. However, it should be clear that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, as long as various changes are within the spirit and scope of the present invention as defined and determined by the appended claims, these changes are obvious, and all inventions and creations utilizing the concepts of the present invention are protected.

[0076] like Figure 1 As shown, a reliable reverse unloading method based on a multimodal transport participant scenario includes the following steps:

[0077] S1. Construct a multimodal reverse unloading system model that includes a set of traffic participants, edge devices, and cloud servers to obtain the reverse unloading of traffic participants and edge computing servers.

[0078] In this embodiment, based on the actual scene of the railway station, such as Figure 2As shown, the intermodal reverse unloading system model integrates cars, buses and trains in a vehicle-edge-cloud three-layer architecture. In the intermodal reverse unloading system model, the edge computing server (VEC) receives information from traffic participants and generates computing tasks. However, due to its limited computing power and the dense data received, the edge computing server may need help to complete tasks in a timely manner, which affects the accuracy of business decisions. Therefore, the edge server alleviates this pressure by offloading some tasks to the fog computing nodes on the traffic participant side. Utilizing the powerful computing resources of the fog computing nodes on the traffic participant side can reduce the load on the edge server and shorten the task processing delay, thereby improving the overall performance and efficiency of the intelligent transportation system (ITS). In this embodiment, only the case where the edge computing server has a heavy computing burden is considered, and the edge computing server is used to reversely offload the load to the traffic participants.

[0079] First, the set of traffic participants includes cars, buses, and trains. All traffic participants can communicate with edge computing nodes and have the computing power of local non-offloaded tasks. Here, edge nodes refer to computing nodes that are close to edge devices and far away from server centers. These traffic participants communicate within the Internet of Vehicles or the train-to-ground public network system (T2G), upload data, and trigger computing tasks at the edge. Considering the real-world situation, there may be dishonest or selfish nodes among the traffic participants. These nodes may maliciously upload incorrect calculation results or receive tasks but avoid computing to protect their resources. In addition, vehicles (cars and buses) and trains are moving or temporarily stationary, resulting in a limited survival time within the node.

[0080] Secondly, at the edge, the roadside units (RSUs) in the Internet of Vehicles communicate with vehicles (cars and buses) through vehicle-to-infrastructure (V2I) communication. The edge computing server connected to the roadside unit generates edge tasks based on the collected information. However, due to the limited resources of the edge computing server, the computing tasks generated at the edge can be offloaded to resource-rich nodes (cars, buses, and trains) for calculation, which is called reverse offloading. Similarly, in the railway transportation system, the edge computing nodes communicate with the trains through the public network. The edge nodes in railway transportation generate computing tasks such as passenger density estimation and railway traffic target detection, which brings about the need for reverse offloading. In addition, since the edge communicates directly with the transportation participants, it selects the reverse offloading node.

[0081] Finally, the cloud, as an authoritative dispatching center, is connected to the edge devices, communicates with each traffic participant, and is responsible for dispatching all edge devices to complete the offloading of cross-modal computing tasks. The cloud server collects reliability and effectiveness information of traffic participants in the system. The number of traffic participants in the multimodal reverse unloading system model is huge, and the tasks and unloading demands generated are many, which may suddenly surge. In order to solve the problem of unbalanced computing loads between different traffic participants, the cloud needs to be comprehensively considered and reasonably dispatched. In this process, the selection of nodes is crucial. Therefore, based on the constructed multimodal reverse unloading system model, this embodiment calculates the reliability and efficiency of traffic participants, and proposes a node selection method based on the reliability, efficiency and computing power of traffic participants based on the greedy method, so as to improve the reliability of reverse unloading of each participant in the multimodal reverse unloading system model.

[0082] Specifically, step S1 includes S11-S12:

[0083] S11. Construct a multimodal reverse unloading system model that includes a set of traffic participants, edge devices, and cloud servers. The traffic participants include cars, buses, and trains. Cars are The bus is Train for Indicates the i v cars, L represents the number of cars, Indicates the i b buses, B represents the number of buses, Indicates the i t′ trains, N represents the number of trains.

[0084] S12. All traffic participants in the traffic participant set communicate with the edge computing node close to the edge device and have the computing capacity of local non-offload tasks. At the same time, all traffic participants communicate within the Internet of Vehicles or the public network system from the train to the ground, upload data and trigger computing tasks at the edge. The edge device is connected to the cloud server and communicates with each traffic participant. It is responsible for scheduling all edge devices and completing the offloading of cross-modal computing tasks.

[0085] S2. Based on the constructed multimodal reverse unloading system model, the honesty, coordination, task completion rate, computing resources and time availability of traffic participants are obtained, and the reliability of traffic participants is calculated.

[0086] Specifically, step S2 includes S21-S27:

[0087] In this embodiment, since the honesty of the transport participants is crucial to their reliability and time availability in the intermodal reverse unloading system model, it is necessary to evaluate the honesty of the transport participants. Through security protocols such as YD / T3957-2021, the edge can determine whether the messages sent by the transport participants are honest, which is crucial to ensuring the safe and reliable operation of the intermodal reverse unloading system model. Honest transport participants usually transmit true and complete data, while dishonest transport participants may send false or tampered data. If the data sent by the transport participants is unreliable, it may lead to errors in decision-making and task allocation control in the intermodal reverse unloading system model. In addition, dishonest transport participants may disrupt the normal operation of the system by sending a large number of false messages, causing congestion or chaos, thereby affecting the stability and reliability of the system. Therefore, the honesty is calculated as follows:

[0088] S21. Calculate the honesty of each traffic participant in the set of traffic participants, that is:

[0089]

[0090] Among them, h represents honesty, q represents the set of traffic participants, v represents cars, b represents buses, t′ represents trains, M represents the number of communication rounds, and p q,i,h represents the honesty of the i-th traffic participant, represents the honest information sent by the i-th traffic participant in the m-th round of communication, It represents the number of messages sent by the i-th traffic participant in the m-th round of communication.

[0091] In this embodiment, represents the honest information sent by traffic participants in history, Indicates the total number of messages sent in M ​​rounds of communication.

[0092] Because the selfishness of vehicles (cars and buses) has a significant impact on their reliability in the intermodal reverse unloading system model. Vehicles with high synergy (cars and buses) are more likely to share information and resources, and better coordinate task allocation and processing through mutual communication and collaboration. This helps to optimize the utilization of system resources and improve the efficiency of task completion. In addition, vehicles with high synergy (cars and buses) are more willing to participate in building a more stable and reliable network topology. Through communication and interaction with adjacent vehicles (cars and buses), traffic participants can dynamically adjust the network topology, optimize routing and data transmission paths, thereby reducing data transmission delays and packet loss, and enhancing the reliability of the intermodal reverse unloading system model. In addition, the synergy of trains is not considered in this embodiment, that is, it is assumed that trains are selfless when performing reverse unloading tasks. Therefore, the synergy is calculated as follows:

[0093] S22. Calculate the coordination of each traffic participant in the traffic participant set, that is:

[0094]

[0095] in, represents the empty set, c represents the synergy, V represents the total number of vehicles, which is the sum of the number of cars and buses, P q,i,c represents the cooperation of the i-th traffic participant, Represents the total number of vehicles participating in the collaboration in the mth round of communication.

[0096] In this embodiment, the task arrival rate represents the ratio of the number of successful interactions between the node that generates the unloading task and the traffic participant to the total number of interactions. This indicator directly reflects the level of transportation participation in task processing and unloading in the intermodal reverse unloading system model. A higher task arrival rate can promote information exchange and sharing among transportation participants, which is beneficial to the authenticity and integrity of the data. When transportation participants (traffic participants involved in transportation) can successfully interact and complete tasks, the data they generate is more trustworthy, which helps to improve the reliability of the system. The task arrival rate can speed up task processing and response speed. With a high task arrival rate, communication and interaction between transportation participants are more frequent, and the intermodal reverse unloading system model can respond and process tasks faster, thereby improving the real-time response capabilities of transportation participants and enhancing the robustness of the intermodal reverse unloading system model. Therefore, the task achievement rate is calculated as follows:

[0097] S23. Calculate the task achievement rate of each traffic participant in the traffic participant set, that is:

[0098]

[0099] Among them, r represents the task achievement rate, P q,i,r represents the task arrival rate of the i-th traffic participant, represents the number of successful arrivals of the i-th traffic participant, represents the number of failed arrivals of the i-th traffic participant.

[0100] In this embodiment, The purpose of increasing 1 / i is to prevent the value of this item from becoming 0 after the first uninstallation failure, thereby affecting subsequent selections.

[0101] The stability and reliability of a traffic participant's computing resources are key to ensuring the normal processing of tasks. If a traffic participant's computing resources are occupied by tasks, or if they become faulty or unstable, this may cause task interruptions or errors, impacting the availability of the traffic participant. Therefore, when offloading computing tasks to the vehicle network, the status of the traffic participant's computing resources must be fully considered and evaluated to ensure the normal operation of the traffic participant and the efficiency of task processing. Therefore, the computing resources of the traffic participant are calculated as follows:

[0102] S24. Calculate the computing resources of the traffic participants in the traffic participant set, namely:

[0103]

[0104] Among them, Γ q,i represents the computing resources of the i-th traffic participant, a i represents the discount factor of the computing resources of the i-th traffic participant, represents the CPU frequency of the reverse offloading of the i-th traffic participant.

[0105] In this embodiment, considering that when computing resources are occupied, continuing to assign tasks may cause queuing or a decrease in computing power, let a i ∈[0,1] is the discount factor of computing resources, The CPU frequency of the reverse offload node, in CPU cycles / second.

[0106] Since selecting vehicles (cars and buses) with longer availability times can improve task reliability and network connection stability, optimize task scheduling efficiency and system load balance, and thus improve the overall performance and reliability of the multimodal reverse unloading system model. Therefore, when selecting vehicles and assigning tasks, the availability time of vehicles must be considered to ensure that the multimodal reverse unloading system model operates in an optimal state. Therefore, the time availability of cars and buses is calculated as follows:

[0107] S25. Calculate the time availability of cars and buses in the set of traffic participants, that is:

[0108]

[0109] in, represents the time availability of the i-th traffic participant, which is a car or a bus, t ave represents the average usage time of the car-bus link, It represents the time that the i-th traffic participant is connected between the car and the bus.

[0110] In this embodiment, since the train available time is the time spent at the station, the train time availability is calculated as follows:

[0111] S26. Calculate the time availability of the train as a transport participant in the transport participant set, namely:

[0112]

[0113] in, Indicates the i t′ The time availability of trains, t pre Indicates the estimated time the train will stay at the site. Indicates the i t′ The time that a train link has been used.

[0114] S27. Calculate the reliability of traffic participants based on their honesty, collaboration, and task completion rate, namely:

[0115]

[0116] in, represents the reliability of the i-th traffic participant, α q,e represents the trust value weight, if represents the judgment condition, P th represents the honesty threshold, and e represents an optional variable.

[0117] In this embodiment, the trust value weight α q,e The weights of different trust values ​​of different transport agents are determined.

[0118] S3. Calculate the efficiency of the traffic participants based on the calculated reliability of the traffic participants.

[0119] In this embodiment, the system performance indicator calculation model (ADC model) is used to evaluate the efficiency of traffic participants. Traffic participants generate computational tasks locally and may reversely offload tasks received from previous communications, thereby depleting their computational resources. Furthermore, based on the reliability calculated in the previous step, it is known that the lifetime of a node affects its efficiency. Therefore, the efficiency of a traffic participant is calculated by comprehensively considering its reliability and the time availability of the traffic participant.

[0120] Specifically, step S3 includes S31-S34:

[0121] S31. Based on the calculated reliability of traffic participants, define the reliability matrix D, utility matrix C′, and availability matrix A′ as follows:

[0122]

[0123] Among them, d 11 represents the probability of a traffic participant changing from state 1 to state 1, d 12represents the probability of a traffic participant changing from state 1 to state 2, d 21 represents the probability of a traffic participant changing from state 2 to state 1, d 22 represents the probability that the traffic participant changes from state 2 to state 2, c1′ and c2′ represent the capability vectors, a1′ represents the probability that the traffic participant has no computational task at any time, a2′ represents the probability that the traffic participant is engaged in a computational task, and T represents the transpose.

[0124] S32. Define the local task rate of the intermodal reverse unloading system model as λ, define the task completion rate of the intermodal reverse unloading system model as μ, and calculate the reliability matrix D, that is:

[0125]

[0126] Where exp(·) represents the natural exponential function, l represents the size of the calculated data, represents the computational workload of reverse unloading of the i-th traffic participant.

[0127] In this embodiment, the calculation workload The number of CPU cycles consumed to calculate 1 bit of data.

[0128] S33. Calculate the utility matrix C′, that is:

[0129]

[0130] In this embodiment, the utility matrix is ​​the probability of completing the calculation under the current state.

[0131] S34. Calculate the efficiency of traffic participants based on the reliability matrix D, the utility matrix C′, and the availability matrix A′, namely:

[0132] E=A′ T DC′

[0133] Among them, E represents efficiency.

[0134] S4. Based on the greedy method, the reliability and computing power of traffic participants are evaluated according to the calculated reliability and efficiency of traffic participants, so as to select the optimal traffic participant for reverse unloading.

[0135] In this embodiment, aggregation parameters are used to consider the combined impact of various factors on the reliability and efficiency of traffic participants. A greedy approach based on reliability and efficiency parameters is employed for comprehensive decision-making. This algorithm selects the current optimal solution without considering future scenarios. This algorithm is computationally efficient and suitable for large-scale problems. In traffic participant selection scenarios, due to the short lifecycle of traffic participants, the greedy approach can sequentially select the best overall performing traffic participant based on the aggregation parameters until the desired number of traffic participants is reached or specific conditions are met.

[0136] Specifically, step S4 includes:

[0137] S41, the number of communication rounds M, the number of trains N, the number of buses B, the number of cars L, the number of malicious vehicles, the number of unloaded tasks, and the CPU frequency Calculate data size l and discount factor a i , The train is expected to stay at the site for t pre , the average usage time t of the car and bus link ave , calculation workload Local task rate λ, task completion rate μ, honesty threshold P th And the trust value weight α q,e As the input of the greedy method.

[0138] S42. Traverse each vehicle in the set of traffic participants, calculate the reliability and efficiency of each vehicle, and store them in the corresponding reliability matrix and utility matrix. The specific process is as follows:

[0139] Calculate the reliability of each vehicle and store it in the reliability matrix.

[0140] Calculate the efficiency of each vehicle and store it in the efficiency matrix.

[0141] S43: Set the selected vehicle to be empty and the best comprehensive score to negative infinity, and traverse the vehicles in the set of traffic participants. The specific process is as follows:

[0142] Determine whether the vehicle's reliability is less than the optimal comprehensive score. If so, exit the traversal. Otherwise, if the vehicle's reliability is greater than the optimal comprehensive score, assign the vehicle to the selected vehicle, and assign the vehicle's reliability to the optimal comprehensive score. If the vehicle's reliability is equal to the optimal comprehensive score and the vehicle's efficiency is greater than the efficiency of the selected vehicle, assign the vehicle to the selected vehicle.

[0143] S44. Get the selected vehicle.

[0144] S45: The selected vehicle is regarded as the optimal traffic participant, so as to select the optimal traffic participant for reverse unloading.

[0145] Specifically, the specific process of selecting the best traffic participant for reverse unloading in step S45 is as follows:

[0146] First, traffic participants upload data and generate computationally intensive tasks in the intermodal reverse unloading system model. When the computational load is high, the edge computing nodes close to the edge devices are prompted to generate unloading requests.

[0147] Secondly, the edge device determines whether reverse offloading is feasible based on the connected traffic participants based on the greedy method. If so, it calculates the offloading process. Otherwise, if the connected traffic participants are participating in other computing tasks or no trustworthy nodes are found, it sends a cross-node offloading request to the cloud server.

[0148] Then, the cloud server selects the optimal traffic participants based on the greedy method to participate in the computing task offloading.

[0149] Finally, the optimal traffic participant processes the computation task and returns the offloaded results to the edge.

[0150] In summary, step S4 comprehensively evaluates the reliability and computing power of the nodes and intelligently selects the most suitable node (selected vehicle) for task offloading, thereby improving the reliability and execution efficiency of the system.

[0151] In order to verify the effectiveness of the reliable reverse unloading method proposed in the present invention based on the multimodal transport participant scenario, this embodiment is verified through simulation experiments, as follows:

[0152] Set the simulation parameters as shown in Table 1:

[0153] Table 1 Simulation parameters

[0154]

[0155] In order to evaluate the rationality of the method proposed in this invention, the results with and without the algorithm are compared in simulation. In addition, an ablation experiment is conducted to demonstrate the role of reliability and efficiency discussed in this invention. In the simulation, unreliable traffic participants are set to send unreliable messages according to different probabilities, where dishonest traffic participants cause the task to fail. In addition, selfish traffic participants decide whether to execute the task according to their corresponding selfishness. The efficiency of each traffic participant is determined by the discount factor a. i and local task rate λ regulation.

[0156] The results are as follows Figure 3-6 As shown, from Figure 3-Figure 6As can be seen from the results, using the task completion rate as an indicator, the proposed method significantly improves the task completion rate. Furthermore, when reliability and efficiency are removed from the considerations, the completion rate of the intermodal reverse unloading system model decreases, demonstrating the rationality of the factors considered in this invention. Furthermore, it can be seen that reliability has a greater impact on the intermodal reverse unloading system model.

[0157] In addition, the effectiveness under larger transportation scale and its impact on the success rate of reverse unloading tasks are also discussed. As the number of traffic participants increases, a large number of tasks are issued, dishonest and selfish vehicles increase, and trustworthy traffic participants such as trains and buses take on more unloading tasks, resulting in a decrease in efficiency. However, even in such a scenario, the proposed method still achieves satisfactory results, which confirms the robustness of the proposed method. Among them, the results of larger-scale traffic simulation are as follows: Figure 7-Figure 8 shown.

[0158] Finally, the probability of selecting different traffic participants is also shown in the simulation. Considering the real-world scenario, it is assumed that trains have longer availability and higher reliability than other participants. Therefore, in theory, the probability of trains being selected is higher. In addition, buses show a higher level of cooperation and a lower degree of selfishness, which makes them more popular than cars to a certain extent. Simulations were also carried out on a larger traffic scale, and the results showed that the method proposed in this invention tends to select more reliable traffic participants. The simulation results of multimodal traffic participants are shown in Figure 2. Figure 9 shown.

[0159] In summary, the reliable reverse unloading method based on the multimodal transport participant scenario proposed in the present invention and the constructed multimodal transport reverse unloading system model are not only applicable to a single type of transportation network, but also applicable to multiple transportation participants (such as cars, trains, and buses), thereby expanding the reference scope of the system; at the same time, the constructed multimodal transport reverse unloading system model significantly improves the diversity and overall performance of the system by promoting resource collaboration between multimodal transportation tools; secondly, when the edge server is overloaded, the multimodal transport reverse unloading system model can dynamically offload computing tasks to local transportation participants to share the load and reduce latency; finally, when selecting the unloading node, the present invention can comprehensively consider the historical communication behavior and current link status of the transportation participants, and select the optimal transportation participant for reverse unloading based on the reliability and efficiency of the transportation participants, thereby quickly and comprehensively evaluating the performance of each transportation participant node.

[0160] Specific embodiments are used in the present invention to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core ideas. At the same time, for those skilled in the art, according to the ideas of the present invention, there may be changes in the specific implementation methods and application scopes. In summary, the contents of this specification should not be understood as limiting the present invention.

[0161] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the principles of the present invention, and it should be understood that the scope of protection of the present invention is not limited to such specific descriptions and embodiments. Those skilled in the art can make various other specific variations and combinations based on the technical teachings disclosed in the present invention without departing from the essence of the present invention, and such variations and combinations are still within the scope of protection of the present invention.

Claims

1. A reliable reverse unloading method based on a multimodal transport participant scenario, characterized in that: The following steps are involved: S1. Construct a multimodal reverse unloading system model that includes a set of traffic participants, edge devices, and cloud servers, and obtain reverse unloading between traffic participants and edge computing servers; the set of traffic participants includes cars, buses, and trains; S2. Based on the constructed multimodal reverse unloading system model, the honesty, coordination, task completion rate, computing resources and time availability of traffic participants are obtained, and the reliability of traffic participants is calculated; S3. Calculate the efficiency of the traffic participants based on the calculated reliability of the traffic participants; S4. Based on the calculated reliability and efficiency of traffic participants, the reliability and computing capacity of traffic participants are evaluated based on the greedy method, and the optimal traffic participant is selected for reverse unloading. Specifically, the method includes: S41, the number of communication rounds , number of trains , number of buses , number of cars , Number of malicious vehicles, Number of unloaded tasks, CPU frequency , calculate data size , discount factor , Estimated time the train will stay at the site , average usage time of car and bus connections , calculation workload , local task rate , task completion rate , honesty threshold and trust value weight As input to the greedy method; S42. Traverse each vehicle in the set of traffic participants, calculate the reliability and efficiency of each vehicle, and store them in the corresponding reliability matrix and utility matrix. The specific process is as follows: Calculate the reliability of each vehicle and store it in the reliability matrix; Calculate the efficiency of each vehicle and store it in the efficiency matrix; S43: Set the selected vehicle to be empty and the best comprehensive score to negative infinity, and traverse the vehicles in the set of traffic participants. The specific process is as follows: Determine whether the vehicle's reliability is less than the optimal comprehensive score. If so, exit the traversal. Otherwise, if the vehicle's reliability is greater than the optimal comprehensive score, assign the vehicle to the selected vehicle and assign the vehicle's reliability to the optimal comprehensive score. If the vehicle's reliability is equal to the optimal comprehensive score and the vehicle's efficiency is greater than the efficiency of the selected vehicle, assign the vehicle to the selected vehicle. S44, obtaining a selected vehicle; S45: The selected vehicle is regarded as the optimal traffic participant, so as to select the optimal traffic participant for reverse unloading.

2. The reliable reverse unloading method based on the multimodal transport participant scenario according to claim 1 is characterized in that: Step S1 specifically includes: S11. Construct a multimodal reverse unloading system model that includes a set of traffic participants, edge devices, and cloud servers. The traffic participants include cars, buses, and trains. Cars are , the bus is , the train is , Indicates the cars, Indicates the number of cars, Indicates the buses, Indicates the number of buses, Indicates the train, Indicates the number of trains; S12. All traffic participants in the traffic participant set communicate with the edge computing node close to the edge device and have the computing capacity of local non-offload tasks. At the same time, all traffic participants communicate within the Internet of Vehicles or the public network system from the train to the ground, upload data and trigger computing tasks at the edge. The edge device is connected to the cloud server and communicates with each traffic participant. It is responsible for scheduling all edge devices and completing the offloading of cross-modal computing tasks.

3. The reliable reverse unloading method based on the multimodal transport participant scenario according to claim 2 is characterized in that: Step S2 specifically includes: S21. Calculate the honesty of each traffic participant in the set of traffic participants, that is: in, Indicates honesty, represents the set of traffic participants, Indicates car, Indicates bus, It means train, Indicates the number of communication rounds, Indicates the The honesty of each traffic participant, Indicates the Traffic participants in the Honest information sent by round communication, Indicates the Traffic participants in the Number of messages sent in round communication; S22. Calculate the coordination of each traffic participant in the traffic participant set, that is: in, represents the empty set, Indicates collaboration, Represents the total number of vehicles, which is the sum of the number of cars and the number of buses. Indicates the The coordination of traffic participants, Indicates the The total number of vehicles participating in the coordinated wheel communication; S23. Calculate the task achievement rate of each traffic participant in the traffic participant set, that is: in, represents the task achievement rate, Indicates the The task arrival rate of each traffic participant, Indicates the The number of successful arrivals of traffic participants, Indicates the Number of failed arrivals of each traffic participant; S24. Calculate the computing resources of the traffic participants in the traffic participant set, namely: in, Indicates the The computing resources of traffic participants, Indicates the The discount factor of the computing resources of each traffic participant, Indicates the The CPU frequency of the reverse unloading of each traffic participant; S25. Calculate the time availability of cars and buses in the set of traffic participants, that is: in, Indicates the The time availability of each traffic participant is car and bus, represents the average usage time of the car-bus link, Indicates the The time that a traffic participant is connected by a car and a bus; S26. Calculate the time availability of the train as a transport participant in the transport participant set, namely: in, Indicates the The availability of train times, Indicates the estimated time the train will stay at the site. Indicates the The time that each train link has been used; S27. Calculate the reliability of traffic participants based on their honesty, collaboration, and task completion rate, namely: in, Indicates the The reliability of each traffic participant, represents the trust value weight, Indicates the judgment condition. represents the honesty threshold, Indicates an optional variable.

4. The reliable reverse unloading method based on the multimodal transport participant scenario according to claim 3 is characterized in that: Step S3 specifically includes: S31. Define the reliability matrix based on the calculated reliability of traffic participants , utility matrix , Availability Matrix They are: in, represents the probability of a traffic participant changing from state 1 to state 1, represents the probability of a traffic participant changing from state 1 to state 2, represents the probability of a traffic participant changing from state 2 to state 1, represents the probability of a traffic participant changing from state 2 to state 2, 、 They represent the capability vectors, represents the probability that a traffic participant has no computational tasks at any time, represents the probability that a traffic participant is engaged in a computing task, represents transpose; S32. Define the local task rate of the intermodal reverse unloading system model as , the task completion rate of the multimodal reverse unloading system model is defined as , calculate the reliability matrix ,Right now: in, represents the natural exponential function, Indicates the calculation data size, Indicates the The computational workload of reverse unloading of each traffic participant; S33. Calculate the utility matrix ,Right now: ； S34, according to the reliability matrix , utility matrix , Availability Matrix , calculate the efficiency of traffic participants, that is: in, Indicates efficiency.

5. The reliable reverse unloading method based on the multimodal transport participant scenario according to claim 1 is characterized in that: The specific process of selecting the best traffic participant for reverse unloading in step S45 is as follows: First, traffic participants upload data and generate computationally intensive tasks in the multimodal reverse unloading system model. When the computational load is high, edge computing nodes close to edge devices are prompted to generate unloading requests. Secondly, the edge device uses the greedy method to determine whether reverse offloading is feasible based on the connected traffic participants. If so, it calculates the offloading process. Otherwise, if the connected traffic participants are participating in other computing tasks or no trustworthy nodes are found, it sends a cross-node offloading request to the cloud server; Then, the cloud server selects the optimal traffic participants to participate in the computational task offloading based on the greedy method; Finally, the optimal traffic participant processes the computation task and returns the offloaded results to the edge.

Images