Cold chain vehicle dispatching method, system, apparatus, and non-transitory storage medium

Abstract

This invention discloses a method, system, device, and non-volatile storage medium for cold chain vehicle scheduling. The method includes: collecting status data corresponding to multiple orders and the current status of multiple cold chain vehicles in a target area; determining the order distribution density in the target area based on the order locations of the orders; determining clustering parameters based on the order distribution density; performing density clustering on the multiple orders based on the clustering parameters to obtain multiple order clusters; dividing the multiple order clusters into multiple task clusters based on the timeliness requirements and temperature zone types of the orders; and determining a target scheduling scheme based on the multiple task clusters and the operating status of the multiple cold chain vehicles. This invention solves the technical problems of current cold chain vehicle cluster energy consumption optimization, which is mostly limited to local optimization at the single vehicle or single station level, lacking global scheduling capabilities, and traditional optimization algorithms having slow convergence and being prone to getting trapped in local optima.

Description

Technical Field

[0001] This invention relates to the field of logistics scheduling technology, and more specifically, to a method, system, device, and non-volatile storage medium for scheduling cold chain vehicles. Background Technology

[0002] Currently, the operation of hydrogen fuel cell refrigerated vehicles in urban and intercity logistics scenarios still mainly relies on independent energy management and decentralized scheduling strategies for individual vehicles. Existing technologies mostly focus on using onboard controllers to construct linear regression or neural network models based on static parameters such as mileage, load, road condition index, and temperature to achieve energy consumption prediction and optimal power allocation for a single vehicle. Although some studies have introduced intelligent algorithms such as genetic algorithms and ant colony optimization for route and hydrogen refueling decisions, their computational complexity is high, convergence is slow, they are prone to getting trapped in local optima, and they lack the ability to accurately integrate dynamic environmental variables (such as real-time congestion, slope changes, traffic light timing, and ambient temperature and humidity). Meanwhile, some solutions attempt to optimize the coordinated supply of hydrogen and electricity at the energy station level through electricity price incentives, V2G interaction, or microgrid dispatch. However, their optimization goals focus on grid stability and energy station economics, rather than prioritizing the operational efficiency and minimization of total costs for cold chain logistics operators. This results in low vehicle-task matching efficiency, crude energy replenishment strategies, and weak cold chain quality assurance capabilities, failing to meet the demand for efficient, coordinated, reliable, and low-cost clustered dispatch for the large-scale operation of 4.5-ton hydrogen fuel cell cold chain vehicles.

[0003] There is currently no effective solution to the above problems. Summary of the Invention

[0004] This invention provides a method, system, device, and non-volatile storage medium for scheduling cold chain vehicles, which at least solves the technical problems that current energy consumption optimization of cold chain vehicle clusters is mostly limited to local optimization at the level of a single vehicle or a single station, lacks global scheduling capabilities, and that traditional optimization algorithms are slow to converge and prone to getting trapped in local optima.

[0005] According to one aspect of the present invention, a cold chain vehicle scheduling method is provided, comprising: collecting status data corresponding to multiple orders in a target area and the current status corresponding to multiple cold chain vehicles, wherein the status data includes at least order location, timeliness requirement, and temperature zone type, and the current status includes load status, current location, and current hydrogen storage capacity; determining the order distribution density in the target area based on the order locations corresponding to the multiple orders; determining clustering parameters based on the order distribution density, wherein the clustering parameters include neighborhood radius and target point number parameters; performing density clustering on the multiple orders based on the clustering parameters to obtain multiple order clusters; dividing the multiple order clusters based on the timeliness requirement and temperature zone type corresponding to the multiple orders to obtain multiple task clusters; and determining a target scheduling scheme based on the multiple task clusters and the operating status corresponding to the multiple cold chain vehicles.

[0006] Optionally, based on the respective operating states of multiple task clusters and multiple cold chain vehicles, a target scheduling scheme is determined, including: determining the geometric center of each of the multiple task clusters; determining an initial scheduling scheme based on the respective geometric centers of the multiple task clusters and the respective operating states of the multiple cold chain vehicles; iteratively adjusting the initial scheduling scheme based on preset sine and cosine operators and calculating the target value corresponding to the new scheduling scheme based on a preset objective function, wherein the preset objective function is determined based on energy cost, cargo loss cost, and time-delay penalty cost; and determining the target scheduling scheme based on the target value corresponding to the new scheduling scheme.

[0007] Optionally, if a new order is detected in the target area, the status data corresponding to the new order is obtained; based on the status data corresponding to the new order, the influencing task cluster is determined from multiple task clusters; the influencing cold chain vehicles corresponding to the influencing task cluster in the target scheduling scheme are determined; and the status data corresponding to the new order is sent to the influencing cold chain vehicles.

[0008] Optionally, when scheduling multiple cold chain vehicles based on a target scheduling scheme, the method includes: collecting real-time data of multiple cold chain vehicles during order execution, wherein the real-time data includes real-time location and real-time remaining hydrogen; obtaining road conditions and energy station operation information in the target area, wherein the energy station operation information includes current hydrogen price, electricity price, and number of vehicles in queue; and constructing a scheduling decision model based on road conditions, energy station operation information, and real-time data of each of the multiple cold chain vehicles, wherein the scheduling decision model is used to plan order execution routes, charging routes, and hydrogen refueling routes.

[0009] Optionally, if the real-time remaining hydrogen quantity corresponding to the first cold chain vehicle is lower than the first hydrogen storage threshold, based on the real-time location of the first cold chain vehicle, it is determined whether there is a hydrogen refueling station within a preset distance from the first cold chain vehicle; if there is a hydrogen refueling station within a preset distance from the first cold chain vehicle, a route to the hydrogen refueling station is generated based on the scheduling decision model; and the route is pushed to the first cold chain vehicle.

[0010] According to another aspect of the present invention, a cold chain vehicle dispatching system is also provided, comprising: multiple terminal devices, respectively deployed in multiple cold chain vehicles and order outlets in a target area, for acquiring status data corresponding to each of the multiple orders in the target area and the current status corresponding to each of the multiple cold chain vehicles, and sending them to multiple edge computing nodes and a cloud platform; a cloud platform, communicatively connected to the multiple edge computing nodes and the multiple terminal devices, for executing the cold chain vehicle dispatching method of any one of claims 1 to 5 and distributing the target dispatching scheme to the multiple edge computing nodes; and multiple edge computing nodes, communicatively connected to the multiple terminal devices, for distributing the target dispatching scheme to the cold chain vehicles within their respective preset communication distances.

[0011] Optionally, in the event of a communication interruption between the cloud platform and multiple edge computing nodes, the multiple edge computing nodes shall schedule the cold chain vehicles within their respective preset communication distances based on the order status data within their respective preset communication distances and the current status of the cold chain vehicles.

[0012] According to another aspect of the present invention, a cold chain vehicle scheduling device is also provided, comprising: a data acquisition module, configured to acquire status data corresponding to multiple orders in a target area and the current status corresponding to multiple cold chain vehicles, wherein the status data includes at least order location, timeliness requirement, and temperature zone type, and the current status includes load status, current location, and current hydrogen storage capacity; a first determination module, configured to determine the order distribution density in the target area based on the order locations corresponding to the multiple orders; a second determination module, configured to determine clustering parameters based on the order distribution density, wherein the clustering parameters include a neighborhood radius and a minimum number of points parameter; a clustering module, configured to perform density clustering on the multiple orders based on the clustering parameters to obtain multiple order clusters; a partitioning module, configured to partition the multiple order clusters based on the timeliness requirement and temperature zone type corresponding to the multiple orders to obtain multiple task clusters; and a scheduling module, configured to determine a target scheduling scheme based on the multiple task clusters and the operating status corresponding to the multiple cold chain vehicles.

[0013] According to another aspect of the present invention, a non-volatile storage medium is also provided, the non-volatile storage medium including a stored program, wherein, when the program is running, the device where the non-volatile storage medium is located is controlled to execute any of the above-described cold chain vehicle scheduling methods.

[0014] According to another aspect of the present invention, a computer device is also provided, the computer device including a processor for running a program, wherein the program executes any of the above-described cold chain vehicle scheduling methods during runtime.

[0015] According to another aspect of the present invention, a computer program product is also provided, including a computer program that, when executed by a processor, implements any of the above-described cold chain vehicle scheduling methods.

[0016] In this embodiment of the invention, a cold chain vehicle scheduling method is adopted. This method collects status data corresponding to multiple orders in a target area and the current status of multiple cold chain vehicles. The status data includes at least order location, timeliness requirement, and temperature zone type. The current status includes load status, current location, and current hydrogen storage. Based on the order locations of the multiple orders, the order distribution density in the target area is determined. Based on the order distribution density, clustering parameters are determined, including neighborhood radius and target point number parameters. Based on the clustering parameters, multiple orders are density-clustered to obtain multiple order clusters. Based on the timeliness requirements and temperature zone types of the multiple orders, the multiple order clusters are divided to obtain multiple task clusters. Based on the multiple task clusters and the operating status of the multiple cold chain vehicles, a target scheduling scheme is determined, achieving the goal of intelligently matching the optimal vehicle scheduling scheme. This improves scheduling efficiency and reduces transportation costs. Furthermore, this method solves the technical problems of current cold chain vehicle cluster energy consumption optimization, which is mostly limited to local optimization at the single vehicle or single station level, lacking global scheduling capabilities, and traditional optimization algorithms having slow convergence and being prone to local optima. Attached Figure Description

[0017] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0018] Figure 1 A hardware structure block diagram of a computer terminal for implementing a cold chain vehicle scheduling method is shown.

[0019] Figure 2 This is a flowchart illustrating the cold chain vehicle scheduling method provided according to an embodiment of the present invention;

[0020] Figure 3 This is an architecture diagram of a cold chain vehicle dispatching system provided according to an embodiment of the present invention;

[0021] Figure 4This is a schematic diagram of the "warehouse-station-vehicle-map-cloud" information structure provided by an optional embodiment of the present invention;

[0022] Figure 5 This is a structural block diagram of a cold chain vehicle dispatching device provided according to an embodiment of the present invention. Detailed Implementation

[0023] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0024] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0025] According to an embodiment of the present invention, a method embodiment of a cold chain vehicle scheduling method is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0026] The method embodiment provided in Embodiment 1 of this application can be executed on a mobile terminal, computer terminal, or similar computing device. Figure 1 A hardware block diagram of a computer terminal for implementing a cold chain vehicle scheduling method is shown. Figure 1As shown, the computer terminal 10 may include one or more processors (shown as 102a, 102b, ..., 102n in the figure) (the processor may include, but is not limited to, a microprocessor MCU or a programmable logic device FPGA, etc.) and a memory 104 for storing data. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of a BUS bus), a network interface, a power supply, and / or a camera. Those skilled in the art will understand that... Figure 1 The structure shown is for illustrative purposes only and does not limit the structure of the aforementioned electronic device. For example, computer terminal 10 may also include... Figure 1 The more or fewer components shown, or having the same Figure 1 The different configurations shown.

[0027] It should be noted that the aforementioned one or more processors and / or other data processing circuits are generally referred to herein as "data processing circuits". These data processing circuits may be embodied, in whole or in part, in software, hardware, firmware, or any other combination thereof. Furthermore, the data processing circuits may be a single, independent processing module, or may be integrated, in whole or in part, into any other element within the computer terminal 10. As involved in the embodiments of this application, the data processing circuits serve as a processor control mechanism (e.g., selection of a variable resistor termination path connected to an interface).

[0028] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the cold chain vehicle scheduling method in this embodiment of the invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory 104, thereby realizing the aforementioned application-based cold chain vehicle scheduling method. The memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include memory remotely located relative to the processor, and these remote memories can be connected to the computer terminal 10 via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0029] The display can be, for example, a touchscreen liquid crystal display (LCD) that allows the user to interact with the user interface of the computer terminal 10.

[0030] This invention primarily targets 4.5-ton-class hydrogen fuel cell cold chain transportation scenarios within urban and intercity areas, serving the high-frequency, high-efficiency delivery needs of temperature-sensitive goods such as food and pharmaceuticals. In this scenario, a large number of intelligent connected cold chain vehicles travel daily between warehousing hubs, community stores, and delivery points, operating in complex and diverse environments, including high-density urban traffic sections as well as areas with weak communication, such as tunnels and underground parking garages. They must meet stringent temperature control requirements while also coping with energy constraints such as uneven distribution of hydrogen refueling stations, fluctuating queue times, and dynamic changes in electricity and hydrogen prices. Traditional scheduling methods rely on manual experience or centralized cloud-based decision-making, making it difficult to respond in real-time to changes in road conditions, energy price fluctuations, and vehicle status anomalies, easily leading to empty runs, hydrogen refueling congestion, temperature control failures, or task backlogs. To address these technical problems, a cold chain vehicle scheduling method is proposed. Figure 2 This is a flowchart illustrating the cold chain vehicle scheduling method provided in an embodiment of the present invention, as shown below. Figure 2 As shown, the method includes the following steps:

[0031] Step S202: Collect the status data corresponding to each of the multiple orders in the target area and the current status of each of the multiple cold chain vehicles. The status data includes at least the order location, time requirement and temperature zone type, and the current status includes the load status, current location and current hydrogen storage.

[0032] In this step, before executing the scheduling task, the system first collects the status data corresponding to all orders to be delivered within the target area, as well as the current operating status of all hydrogen fuel cell cold chain vehicles within the area. The order status data includes at least: the geographical coordinates of the order, the temperature control range required for cargo transportation (i.e., temperature zone type, such as cryogenic ≤-18℃, refrigerated 2–8℃, or constant temperature 15–20℃), and the time window from the order creation time to the latest delivery time specified by the customer (i.e., timeliness requirements). The vehicle's current status includes at least: the total mass of cargo currently carried by the vehicle (i.e., load status), the precise location determined in real time by the BeiDou or GPS positioning module, and the current remaining hydrogen storage capacity of the fuel cell system (in kilograms). This data can be collected in real time by end devices deployed at warehouse smart gateways, vehicle terminals, and integrated hydrogen-electricity energy stations, and transmitted to regional edge computing nodes via 5G communication, V2X, or edge gateways, serving as input for subsequent task clustering and scheduling decisions.

[0033] Specifically, a three-dimensional parameter system encompassing vehicles, tasks, and the environment can be established in the cloud. Vehicle parameters include precise settings for vehicle type, cluster size, daily mileage range, load capacity range, average hydrogen storage, refrigeration unit power, and fuel cell system degradation, among other vehicle profile characteristics. Task parameters include integrated order information, such as cargo temperature zone, delivery time requirements, cargo volume / weight, and warehouse loading / unloading time windows. Environmental parameters include access to dynamic road conditions (congestion, gradient), traffic light timings, and ambient temperature and humidity information provided by high-precision maps. Based on these parameters, a multi-objective optimization function is established, with minimizing total operating costs as the core, while also maximizing delivery timeliness and ensuring the highest operational safety and reliability. Total costs include hydrogen / electricity consumption costs, time window default penalty costs, and cargo damage risk costs.

[0034] Through the above data collection, the system obtains the original status information of orders and vehicles in the spatial, temporal, and energy dimensions, ensuring that the dispatching system can perform collaborative optimization based on multi-dimensional, high-precision, and real-time updated information under real operating conditions, and avoid the risk of delivery delays, energy waste, or cargo damage caused by missing or delayed information.

[0035] Step S204: Determine the order distribution density in the target area based on the order locations corresponding to each of the multiple orders.

[0036] In this step, spatial clustering analysis is used to calculate the spatial distribution density of orders within a target area based on the location of each order. Specifically, the system maps the latitude and longitude coordinates of all orders to a two-dimensional geographic grid, calculates the spatial distance distribution between adjacent orders, and identifies hotspot and sparse areas where orders cluster. In order-dense areas (such as urban commercial cores and large community clusters), the number of orders per unit area is significantly higher than the average, exhibiting a high-density clustered distribution; while in suburban or roadside areas, orders show a discrete, low-density distribution. This density information does not rely on manually set thresholds but is dynamically identified through an adaptive algorithm, providing a real and objective spatial basis for subsequent order clustering, edge node task partitioning, and scheduling resource allocation. This ensures that the system can intelligently adjust the computational load and decision-making strategies according to the actual needs of different areas, avoiding excessive consumption of computational resources in low-density areas while ensuring scheduling response efficiency and coverage in high-density areas.

[0037] By analyzing the geographic coordinates of each order, its spatial distribution within the target area is identified, directly reflecting the degree of order clustering in different regions. Order locations serve as the raw input data, and their spatial arrangement is used to quantify the density or sparseness of orders within the area, forming a direct description of the spatial clustering characteristics of orders. This process does not involve the intervention of any other attributes or weight calculations; the spatial attribute of distribution density is derived solely from the location information itself.

[0038] Step S206: Based on the order distribution density, determine the clustering parameters, which include the neighborhood radius and the number of target points.

[0039] In this step, based on the spatial distribution density of orders within the target area, two key clustering parameters used by the density clustering algorithm—the neighborhood radius and the minimum target number of points—can be dynamically determined. The neighborhood radius refers to the farthest distance threshold used in spatial clustering to determine whether two orders belong to the same neighborhood. When the geographical distance between two orders is less than or equal to this radius, the system considers them spatially "adjacent" or "closely related"; if the distance exceeds this threshold, they are considered separate points. This parameter directly determines the "granularity" of clustering—a smaller radius results in finer clustering, suitable for high-density areas; a larger radius results in looser clustering, helping to connect orders in sparse areas. The minimum target number of points refers to the minimum number of orders a region must contain to be identified as a valid cluster. Only when an order, centered on a given order, contains at least this number of other orders within its neighborhood radius will the region be considered a "real cluster," rather than an isolated point or noise. This parameter is used to filter sparse or anomalously distributed orders, avoiding invalid clustering triggered by individual scattered orders, thereby improving the stability and schedulability of the clustering results.

[0040] When a high order density is detected in a certain area, the neighborhood radius is automatically reduced to improve the precision of clustering. This allows orders within dense areas to be divided into smaller, more precise sub-clusters, preventing mixing of different business points. Simultaneously, the minimum target number of points is increased to ensure effective clustering only occurs in areas with sufficient order aggregation intensity, filtering out noise points. Conversely, in sparsely distributed suburbs or at the ends of road networks, the system automatically expands the neighborhood radius and reduces the minimum target number of points. This allows previously discrete orders to be rationally aggregated into scheduleable transportation units, preventing isolated orders from failing to form groups due to overly strict parameters. This adaptive parameter adjustment mechanism is based on real-time calculation of the local point density distribution function, requiring no manual intervention. It ensures that the clustering results possess spatial rationality and scheduling feasibility under different urban structures (such as grid-like road networks in city centers and radial road networks in suburbs), providing a stable and adaptive input basis for subsequent vehicle-task matching.

[0041] Step S208: Based on the clustering parameters, perform density clustering on multiple orders to obtain multiple order clusters.

[0042] In this step, based on a dynamically determined neighborhood radius and minimum target number of points, a density clustering algorithm based on DBSCAN is executed on all orders within the target area. Orders with similar geographical locations and business characteristics are automatically aggregated into multiple independent order clusters. The algorithm treats each order as a data point, traversing other orders within its neighborhood radius. If an order's neighborhood contains other orders with at least the minimum target number of points, it is marked as a core point. The algorithm then expands outward from this core, merging all connected core points and their neighboring boundary points into a single cluster. Isolated orders not covered by any core point are considered noise points and are not included in the current batch scheduling, but will be processed later during rolling optimization. Ultimately, the system outputs several spatially contiguous, task-load-balanced order clusters, with orders within each cluster being geographically highly concentrated.

[0043] Specifically, starting with the first order, each order is checked one by one to see if it has been processed. If not, it is marked as "pending processing." For the currently pending order, the system uses it as the center and searches for all other orders within its "neighborhood radius," collecting these orders to form its "neighbor set." If the number of neighbors in this order's neighbor set is not less than the "minimum target number," it is considered a "core point," and the system will use it as the core to begin building an order cluster. The current core point is added to a new cluster, and all its neighbors are added to a pending processing queue. An order is taken from the queue and checked if it has been processed. If not, it is marked as processed, and all its neighbors are searched. If the number of neighbors of this neighbor also reaches the "minimum target number," it is also a core point, and the system adds all its neighbors to the pending processing queue to continue expanding the cluster. Regardless of whether a neighbor is a core point, as long as it is within the neighborhood of the current cluster, it is added to the current cluster. The above steps are repeated until the pending processing queue is empty, at which point a complete order cluster is formed.

[0044] The system continues processing the next unprocessed order. If it doesn't belong to any existing cluster, the process is repeated to generate a new cluster. If an order is never covered by the neighborhood of any core point, it's considered isolated from any dense area, and the system marks it as a "noise point," temporarily excluding it from the current scheduling. After all orders are processed, the system obtains several order clusters, where orders within each cluster are geographically close and sufficiently dense.

[0045] Step S210: Based on the timeliness requirements and temperature zone types of each order, the multiple order clusters are divided to obtain multiple task clusters.

[0046] In this step, after completing the initial order clustering based on spatial distribution, the initial order cluster can be further re-divided into multi-dimensional attributes by combining the time requirements of each order (e.g., urgent ≤ 2 hours, standard ≤ 6 hours) and temperature zone type (e.g., deep cold ≤ -18℃, refrigerated 2–8℃, constant temperature 15–25℃) to generate the final task cluster.

[0047] Specifically, the system performs attribute consistency checks on each initial order cluster: if the temperature zone types of orders within the cluster are inconsistent (such as a mixture of cryogenic and refrigerated orders), the cluster is split according to the independent temperature control compartment configuration of the refrigerated vehicle, ensuring that each task cluster contains only orders of the same type of temperature zone, thus avoiding overloading of the refrigeration system or temperature runaway due to mixed loading; if the timeliness requirements of orders within the cluster differ too much (such as the coexistence of urgent and regular orders), the urgent orders are separated into high-priority task sub-clusters, ensuring that they can be preferentially assigned to vehicles with fast response times and optimal routes, thus preventing cargo damage due to delays.

[0048] At the same time, the system will comprehensively evaluate the total service time, total mileage, number of orders and volume / weight of goods for each sub-cluster to ensure that the load of each task cluster is within the carrying capacity of a single vehicle (such as controlling the daily mileage to 100-500km and the single service time to no more than 8 hours), so as to avoid vehicle overloading, temperature control failure or failure to complete delivery on time due to excessive task load.

[0049] Ultimately, the system outputs multiple spatiotemporally and attribute-consistent, load-balanced, and scheduleable task clusters. These task clusters become the smallest scheduling units for subsequent vehicle-task matching using the Clustering Whale Optimization Algorithm (CWOA), realizing the intelligent encapsulation from "point-to-point orders" to "executable task packages," and providing a clear and business-compliant basic input for cluster-level collaborative scheduling.

[0050] Step S212: Determine the target scheduling scheme based on the respective operating status of multiple task clusters and multiple cold chain vehicles.

[0051] In this step, based on the respective operating states of multiple task clusters and multiple cold chain vehicles, a global collaborative matching algorithm (CWOA) can be used to determine the optimal target scheduling scheme. The specific process is as follows:

[0052] The system first treats each task cluster generated in the previous stage as an independent scheduling unit, which includes key attributes such as geographical distribution, temperature zone type, timeliness requirements, total cargo volume, estimated service time and total driving mileage; at the same time, the system obtains the real-time operating status of each cold chain vehicle, including the current hydrogen balance, battery charge, compartment temperature, current location, assigned tasks, fuel cell degradation level, maximum load capacity and driving range.

[0053] Subsequently, the system constructs a multi-dimensional matching space by combining task clusters and vehicle status. With "minimizing total operating costs" as the core objective, it comprehensively considers hydrogen / electric energy consumption costs, time window default penalties, cargo damage risks, vehicle empty driving distances, and scheduling balance to build a multi-dimensional optimization function.

[0054] At the algorithm execution level, the system uses the geometric center of the task cluster as the initial population. Through an adaptive prey encirclement mechanism, combined with sine-cosine operators, the search step size is dynamically adjusted to balance global exploration and local fine-grained search. A simulated annealing bubble net attack strategy is introduced to accept suboptimal solutions with a certain probability, effectively escaping local optima. When there is no significant improvement in optimal solutions for several consecutive generations, a scout bee mechanism is triggered to randomly reset part of the population, maintain population diversity, and prevent premature convergence.

[0055] The algorithm iterates round by round, intelligently matching the most suitable vehicle combination for each task cluster, while meeting the following conditions:

[0056] Temperature range requirements must be fully matched with the vehicle's cooling capacity (e.g., for cryogenic orders, only vehicles equipped with cryogenic compartments will be allocated).

[0057] Clusters with high timeliness priority will be allocated high-response vehicles first;

[0058] The daily mileage of a single vehicle should be controlled within the range of 100-500km to avoid overloading.

[0059] Vehicles with sufficient hydrogen reserves will be given priority for long-distance missions, while vehicles with low hydrogen reserves will be prioritized for allocation to nearby clusters.

[0060] The vehicle workload is balanced to avoid some vehicles being overloaded and others being idle.

[0061] Ultimately, the algorithm outputs an optimal binding relationship between a set of vehicle-task clusters, forming an executable scheduling scheme. Each vehicle is assigned a continuous, connected, and spatiotemporally consistent task cluster, along with recommended routes and hydrogen refueling suggestions. This scheme ensures both the reliability and timeliness of cold chain transportation, while synergistically minimizing the overall energy consumption and operating costs of the vehicle cluster, providing a decision-making basis for real-time guidance at the edge layer and dynamic updates on the cloud platform.

[0062] The above steps significantly improve the accuracy of cluster partitioning and the matching efficiency of scheduling resources by using density-adaptive clustering and multi-dimensional feature fusion partitioning. This effectively solves the problem in existing technologies that cannot achieve efficient cluster partitioning and vehicle task collaborative allocation based on the spatial distribution density and multi-dimensional features of cold chain orders. It achieves the comprehensive effect of improving the intelligence level of cold chain transportation scheduling, reducing empty running rate and energy consumption, and ensuring the timeliness of temperature control.

[0063] As an optional embodiment, a target scheduling scheme is determined based on the respective operating states of multiple task clusters and multiple cold chain vehicles, including: determining the geometric center of each of the multiple task clusters; determining an initial scheduling scheme based on the respective geometric centers of the multiple task clusters and the respective operating states of the multiple cold chain vehicles; iteratively adjusting the initial scheduling scheme based on preset sine and cosine operators and calculating the target value corresponding to the new scheduling scheme based on a preset objective function, wherein the preset objective function is determined based on energy cost, cargo loss cost, and time-delay penalty cost; and determining the target scheduling scheme based on the target value corresponding to the new scheduling scheme.

[0064] Optionally, based on the respective operating states of multiple task clusters and multiple cold chain vehicles, the system gradually constructs and optimizes the target scheduling scheme through an improved clustering whale optimization algorithm. The specific steps are as follows:

[0065] First, for each task cluster that has completed spatial clustering, the system calculates the center point of the geographical locations of all its orders, which serves as the geometric center of that task cluster. This geometric center represents the spatial cluster location of the cluster on the map and is used for subsequent spatial matching and path estimation between vehicles and tasks.

[0066] Next, the real-time operating status of all refrigerated vehicles is acquired, including their current location, remaining hydrogen supply, cabin temperature, whether they are currently performing a task, maximum load capacity, driving range, fuel cell health status, and historical energy consumption. The system performs an initial matching between the geometric center of each task cluster and the current location and status of each refrigerated vehicle, prioritizing vehicles with closer geometric center distances, sufficient hydrogen reserves, and suitable temperature control capabilities to be assigned to that task cluster, forming an initial, executable scheduling scheme. This initial scheme satisfies basic geographical accessibility and capability adaptability, but does not achieve global optimization.

[0067] Subsequently, the optimization iteration process is initiated. In each iteration, the system dynamically adjusts the current scheduling scheme based on an intelligent search mechanism simulating the hunting behavior of whale groups, combined with preset sine and cosine factors. The sine and cosine operators are a dynamic search control mechanism used in optimization algorithms. They do not rely on complex mathematical formulas but intelligently adjust the "exploration" and "development" capabilities of the search process by simulating the behavior of sine and cosine functions in periodic fluctuations. Specifically, the system simulates the spiral approach behavior of whales around their prey, locally perturbing the allocation relationship between some vehicles and task clusters. Simultaneously, by controlling the magnitude of the perturbation through the periodic fluctuations of sine and cosine functions, the algorithm broadly explores different scheduling combinations in the early stages and gradually focuses on high-optimal solution regions in the later stages, achieving a balance between global search and local fine-tuning.

[0068] After each disturbance, the system calculates the target value corresponding to the new scheduling plan. This target value consists of three parts: first, the energy cost of hydrogen and electricity consumed by the vehicle operation; second, the cargo damage cost corresponding to the risk of cargo loss due to abnormal compartment temperature or transportation delays; and third, the time penalty cost incurred for failing to complete delivery within the customer's agreed time window. The three factors are weighted together to form a unified evaluation index reflecting the overall operational efficiency and economy.

[0069] The objective value of the new solution is compared with that of the previous optimal solution. If the objective value of the new solution is lower, it indicates that its overall cost is better, and it is retained as the current optimal solution. If the objective value is higher, the system may still accept the solution with a certain probability based on the simulated annealing mechanism to avoid getting trapped in local optima. When no significant improvement is observed in several consecutive iterations, the system triggers the "scout bee" mechanism, randomly reinitializing the allocation relationship between some vehicles and tasks to reactivate the search space and improve global optimization capabilities.

[0070] The above iterative process continues until preset convergence conditions are met, such as reaching the maximum number of iterations, the target value change falling below a threshold, or the time window being exhausted. Finally, the system outputs a set of optimal correspondences between vehicles and task clusters, forming the final target scheduling scheme.

[0071] Specifically, lightweight decision-making models can be deployed at edge nodes. These models can employ mixed-integer linear programming (MILP) based on queuing information from integrated energy stations, or proactive early warning and peak-shaving guidance strategies based on vehicle hydrogen storage capacity information, enabling rapid real-time scheduling. To address the replenishment cost issue, a MILP model is constructed, simultaneously optimizing constraints such as hydrogen price, electricity price, queuing time, and vehicle energy reserves, resulting in a MILP model with the objective of minimizing total operating costs. The decision variables are set as follows: , indicating that vehicle i is assigned to station j, and the continuous variable is , representing the load factor of station j, with the key constraint being the differentiated energy replenishment time window, as shown in formulas (1) and (2). The differentiated energy replenishment time window guides priority queuing, minimizing the cost objective function in formula (3). Queue time penalties and The model is related to site load rate; sites with high load rates are more prone to queuing. The model can adjust the weighting coefficients accordingly. and By weighing operating costs (electricity / hydrogen costs), operational efficiency (queue time), and service reliability (risk of cargo damage), the solver will ultimately find a set of decision variables. Vehicle charging station allocation and The site load rate is assigned to minimize the overall cost Z.

[0072] hydrogen fuel cell vehicles ( =1 hour): Service start time + 5~10 minutes ≥ Service end time; (1)

[0073] pure electric vehicles ( =1 hour): Service start time + 45~120 minutes ≥ Service end time; (2)

[0074] (3)

[0075] Where Z is the total cost, C_h is the unit price of hydrogen (yuan / kg), C_e is the unit price of electricity (yuan / kWh), H_i is the amount of hydrogen refueling for vehicle i (kg), and E_i is the amount of electricity charged for vehicle i (kWh). This is the penalty coefficient for queuing time. The queuing time penalty represents the time cost incurred by vehicles waiting in line at supply stations. This is the penalty coefficient for cargo damage risk. The risk coefficient for cargo damage can be calculated by multiplying the probability of temperature exceeding the limit by the total value of the batch of goods. The probability of temperature exceeding the limit can be calculated based on historical temperature control data and the estimated time of the current path.

[0076] This solution ensures that the tasks undertaken by each vehicle are spatially consistent, can be completed within a time window, are matched in terms of energy capacity, and are optimal in terms of cost structure, providing reliable and executable scheduling instructions for real-time navigation of edge nodes and digital twin updates in the cloud.

[0077] As an optional embodiment, when a new order is detected in the target area, the status data corresponding to the new order is obtained; based on the status data corresponding to the new order, the influencing task cluster is determined from multiple task clusters; the influencing cold chain vehicles corresponding to the influencing task cluster in the target scheduling scheme are determined; and the status data corresponding to the new order is sent to the influencing cold chain vehicles.

[0078] Optionally, if a new order is detected in the target area, the system activates a dynamic response mechanism to perform partial optimization and real-time updates to the original scheduling plan. The specific steps are as follows:

[0079] First, the system monitors new delivery requests from warehouse outlets or order platforms in real time. After confirming that a new order has entered the dispatch system, it immediately extracts the status data corresponding to the order, including its geographical coordinates, the temperature zone of the goods (such as deep cold or refrigerated), time requirements (such as urgent or standard), the volume and weight of the goods, and the expected loading and unloading time window, etc.

[0080] Subsequently, the system rapidly compares the new order with multiple existing task clusters based on its spatial location and business characteristics. By calculating the spatial distance between the new order and the geometric center of each task cluster, and combining this with temperature zone consistency, timeliness priority, and load balancing, the system identifies several task clusters most significantly affected by the new order, termed "affected task clusters." This identification process does not rely on global recalculation; it performs efficient matching only within a local scope, ensuring that the response speed meets real-time requirements.

[0081] Next, based on the current target scheduling plan, the system determines the cold chain vehicles allocated to the aforementioned "affected task clusters," which are called "affected cold chain vehicles." These vehicles are physical carriers that have already undertaken relevant cluster tasks in the original plan and may require adjustments in spatial paths or time windows due to new orders.

[0082] Finally, the system pushes the status data of the new order to the onboard terminal of each vehicle "affecting the cold chain," and simultaneously displays auxiliary information on the vehicle, such as the new task notification, route adjustment suggestions, changes in estimated service time, and the impact of energy replenishment. Vehicle drivers or dispatchers can choose to accept the new task, request task reorganization, or have the system automatically trigger a lightweight local rescheduling mechanism to dynamically insert the new order and generate an updated route and hydrogen refueling plan without interrupting the original task.

[0083] This mechanism enables "incremental scheduling updates," avoiding full replanning triggered by changes in a single order, significantly reducing computational load and improving system response efficiency. In weak network environments, edge nodes can independently complete the above judgments and information pushes based on locally cached data, ensuring timely issuance of scheduling instructions and guaranteeing the continuity and service quality of cold chain transportation.

[0084] As an optional embodiment, when scheduling multiple cold chain vehicles based on a target scheduling scheme, the method includes: collecting real-time data of multiple cold chain vehicles during order execution, wherein the real-time data includes real-time location and real-time remaining hydrogen; obtaining road conditions and energy station operation information in the target area, wherein the energy station operation information includes current hydrogen price, electricity price, and number of vehicles in queue; and constructing a scheduling decision model based on road conditions, energy station operation information, and real-time data of each of the multiple cold chain vehicles, wherein the scheduling decision model is used to plan order execution routes, charging routes, and hydrogen refueling routes.

[0085] Optionally, during the real-time scheduling of multiple cold chain vehicles based on the target scheduling scheme, the system integrates dynamic environmental information and vehicle operating status to construct an adaptive scheduling decision model, thereby dynamically optimizing vehicle route execution and energy replenishment behavior. The specific steps are as follows:

[0086] First, the system continuously collects real-time operational data from each refrigerated vehicle during its mission, including key parameters such as the vehicle's precise geographical location, real-time remaining hydrogen supply, interior temperature, speed, load status, and battery charge. This data is periodically uploaded by the onboard terminal via a 5G / V2X communication link to ensure the timeliness and accuracy of the information.

[0087] Simultaneously, the system acquires dynamic environmental information within the target area, covering two aspects: first, the road operation status provided by high-precision real-time maps, including the degree of traffic congestion on each road segment, changes in road slope, traffic light phases and countdowns, construction restriction areas, and weather impacts; second, the real-time operation information of the integrated electric-hydrogen energy station, including the current hydrogen sales price, electricity price, number of available hydrogen refueling stations, number of vehicles queuing, estimated waiting time, and station load rate for each station.

[0088] Based on this, the system performs spatiotemporal alignment and semantic fusion on the three types of data mentioned above—real-time vehicle status, road dynamic information, and energy station operating status—to construct a scheduling decision model oriented towards multi-objective collaborative optimization. This model does not rely on global replanning but instead evaluates the comprehensive cost of the current path in real time on a single vehicle or local fleet basis.

[0089] The scheduling decision model dynamically adjusts the next stage of delivery routes based on current road congestion and gradient, prioritizing routes with lower energy consumption, shorter delivery times, and higher temperature control stability to avoid overloading the refrigeration system due to frequent starts and stops or long uphill climbs. Simultaneously, when the vehicle's remaining hydrogen level falls below a preset threshold, the system comprehensively evaluates the total "hydrogen refueling cost" of multiple available refueling stations based on the vehicle's current location, range, and real-time energy station information. This cost is calculated by weighting hydrogen prices, waiting times, arrival distance, and cargo damage risk. The system recommends the optimal refueling station and generates a navigation plan including detours. If the charging cost is significantly lower than the hydrogen refueling cost and the time window allows, the system proactively recommends inserting the vehicle into a charging station to achieve synergistic hydrogen and electric power utilization. When multiple vehicles are simultaneously guided to the same energy station, the system can push "off-peak suggestions" to some vehicles based on predicted arrival times and station load trends, such as delaying refueling with price discounts, guiding vehicles to disperse their refueling behavior and avoid localized congestion.

[0090] This scheduling decision model is deployed and runs on edge nodes, supporting millisecond-level response. When communication is interrupted, the edge nodes can call up the latest road conditions and energy station data cached locally, and make autonomous decisions based on a preset rule base (such as "hydrogen level below 20% must be refueled at the nearest hydrogen station") to ensure that vehicles can still complete tasks safely and efficiently in a cloud-free environment.

[0091] As an optional embodiment, if the real-time remaining hydrogen quantity corresponding to the first cold chain vehicle is lower than the first hydrogen storage threshold, based on the real-time location of the first cold chain vehicle, it is determined whether there is a hydrogen refueling station within a preset distance from the first cold chain vehicle; if there is a hydrogen refueling station within a preset distance from the first cold chain vehicle, a route to the hydrogen refueling station is generated based on the scheduling decision model; and the route is pushed to the first cold chain vehicle.

[0092] Optionally, if the real-time remaining hydrogen quantity corresponding to the first cold chain vehicle is lower than the first hydrogen storage threshold, the system will automatically trigger the energy security response mechanism, and the specific execution steps are as follows:

[0093] First, the system monitors the hydrogen remaining data uploaded by the first cold chain vehicle in real time. When it detects that the remaining hydrogen is lower than the preset first hydrogen storage threshold (for example, 20% of the vehicle's full hydrogen capacity), it determines that the vehicle has entered an energy emergency warning state, and the system immediately starts the path optimization process.

[0094] Subsequently, the system defines a dynamic search area centered on the current real-time location of the first cold chain vehicle and with the maximum driving radius of the vehicle under the current load and road conditions as the radius. This search radius is adaptively adjusted based on real-time operating conditions such as ambient temperature, road slope, and refrigeration load of the vehicle compartment to ensure that the calculated results match the actual driving range and avoid misjudgments due to environmental fluctuations.

[0095] Within the search area, the system quickly retrieves high-precision maps and an energy station database to verify the existence of at least one available hydrogen refueling station. The criteria for judgment include: whether the hydrogen refueling station is operational, whether it has hydrogen supply capacity, whether it supports the vehicle model, and whether there are any temporary closures or malfunction alarms. If no available hydrogen refueling station is found in the area, the system will escalate the warning level and activate emergency backup plans (such as driving to the nearest charging station or requesting manual intervention from the dispatch center).

[0096] If at least one qualified hydrogen refueling station is detected within a preset distance range, the system will invoke the locally deployed scheduling decision model to comprehensively evaluate all available hydrogen refueling stations. Evaluation dimensions may include the expected driving distance and time to reach the refueling station, the number of vehicles currently queuing at the station and the expected waiting time, the real-time hydrogen price and station load rate, the additional impact of congestion and gradient on energy consumption along the route, and whether it may affect the original order delivery time window.

[0097] Based on the aforementioned multi-dimensional parameters, the system calculates the "comprehensive refueling cost" of each candidate hydrogen refueling station and selects the optimal station with the lowest cost as the target. Subsequently, the system generates an optimal navigation route from the current location of the first cold chain vehicle to the target hydrogen refueling station. This route not only considers the shortest path but also comprehensively optimizes energy consumption and time efficiency, such as avoiding congested sections, selecting roads with gentle slopes, and prioritizing access to traffic light priority zones to minimize hydrogen consumption along the way.

[0098] Finally, the system pushes the navigation route to the first cold chain vehicle in real time in the form of pictures and text through the vehicle terminal or intelligent cockpit interface. It also displays key information such as the name of the target hydrogen refueling station, the estimated arrival time, the estimated queuing time, the current hydrogen price and the recommended refueling amount, and triggers voice prompts and visual warnings to remind the driver to respond in time.

[0099] This mechanism achieves a four-step closed loop of "perception-judgment-decision-guidance," ensuring that vehicles can receive safe, efficient, and actionable refueling guidance as soon as possible when hydrogen is running low, preventing transportation interruptions caused by energy depletion, and serving as a core safety barrier to ensure the continuous and stable operation of cold chain fleets.

[0100] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that the present invention is not limited to the described order of actions, because according to the present invention, some steps can be performed in other orders or simultaneously. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0101] Through the above description of the embodiments, those skilled in the art can clearly understand that the cold chain vehicle scheduling method according to the above embodiments can be implemented by means of software plus necessary general-purpose hardware platform. Of course, it can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.

[0102] According to another aspect of the present invention, a cold chain vehicle dispatching system is also provided. Figure 3 This is an architecture diagram of a cold chain vehicle dispatching system provided according to an embodiment of the present invention, such as... Figure 3As shown, it includes: multiple terminal devices, deployed in multiple cold chain vehicles and order outlets in the target area, used to obtain the status data corresponding to each of the multiple orders in the target area and the current status corresponding to each of the multiple cold chain vehicles, and send them to multiple edge computing nodes and a cloud platform; a cloud platform, which is communicatively connected to multiple edge computing nodes and multiple terminal devices, used to execute any of the above-mentioned cold chain vehicle scheduling methods and distribute the target scheduling plan to multiple edge computing nodes; and multiple edge computing nodes, which are communicatively connected to multiple terminal devices, respectively used to distribute the target scheduling plan to the cold chain vehicles within their respective preset communication distances.

[0103] Multiple terminal devices, deployed across various cold chain vehicles and order distribution points within the target area, serve as the system's perception layer and data acquisition terminals, enabling real-time acquisition and transmission of critical operational data. Each terminal device performs differentiated data acquisition functions based on its deployment location: Terminal devices deployed on cold chain vehicles integrate a BeiDou / GPS dual-mode positioning module, a 5G / V2X communication unit, a CAN bus interface, and a temperature control sensor, continuously collecting real-time vehicle location, remaining hydrogen supply, compartment temperature, load status, driving speed, battery charge, and fuel cell operating parameters; terminal devices deployed at order distribution points (such as warehousing centers, convenience stores, and delivery stations) act as industrial-grade intelligent gateways, connecting to the Warehouse Management System (WMS) and order scheduling system to obtain order geographical coordinates, cargo temperature zone type (e.g., cryogenic ≤-18℃, refrigerated 2–8℃), cargo volume / weight, estimated loading / unloading time window, and timeliness level (urgent / standard), among other status information. All terminal devices possess local caching and network outage recovery capabilities, temporarily storing data during communication failures and automatically retransmitting it after network recovery to ensure data integrity.

[0104] The cloud platform establishes stable communication connections with multiple edge computing nodes and all end devices, serving as the system's global decision-making and intelligent hub, undertaking non-real-time, highly complex computational and collaborative optimization tasks. The cloud platform adopts a distributed microservice architecture, supporting concurrent access for fleets of thousands of vehicles and massive data processing. Its core functions include: receiving and integrating order status data and vehicle operation data uploaded from all end devices; constructing a five-dimensional coupled information model of "warehouse-station-vehicle-map-cloud" to achieve unified modeling and digital twin construction of all-domain data; executing global scheduling optimization algorithms, including task cluster partitioning and vehicle allocation based on the clustering whale optimization algorithm (CWOA), generating energy replenishment strategies for the multi-energy collaborative scheduling model (MILP-Hybrid), and city-level logistics route simulation and model training; distributing the generated target scheduling scheme (including vehicle-task allocation relationships, hydrogen refueling / charging guidance strategies, route optimization instructions, etc.) to each corresponding edge computing node, and synchronously updating the global scheduling knowledge base and optimization model parameters to achieve continuous system evolution.

[0105] Multiple edge computing nodes are deployed in large-scale integrated electric-hydrogen energy stations, core storage hubs, or transportation nodes within the target area. They form local communication clusters with multiple end devices deployed within their service radius, undertaking low-latency, high-reliability local response tasks. Each edge computing node has edge inference capabilities and lightweight computing resources. Its core functions include: receiving vehicle and order data uploaded from local end devices in real time, performing local preprocessing and data cleaning; receiving target scheduling plans issued by the cloud platform and accurately distributing scheduling instructions to cold chain vehicles within its service range (within a preset communication distance, typically 3–10 kilometers) based on geographical proximity; autonomously initiating local autonomous mode when communication with the cloud platform is interrupted or network latency exceeds limits, performing key decisions such as lightweight path planning, hydrogen refueling guidance, and task reassignment based on locally cached historical data and a preset rule base, ensuring the system's continuous service capability in weak network or network outage environments; and providing feedback on local operating status and optimization suggestions to the cloud platform, supporting incremental model training and strategy iteration.

[0106] Figure 4 This is a schematic diagram of the "warehouse-station-vehicle-map-cloud" information structure provided by an optional embodiment of the present invention, such as... Figure 4 As shown, a vehicle cluster cloud control platform is constructed. This platform serves as the core hub, integrating and coordinating the following five key elements through standardized data interface protocols:

[0107] A) Warehouse and store outlets: Provide order information, cargo temperature requirements, and estimated loading and unloading time windows;

[0108] B) Integrated Electric-Hydrogen Energy Station: Provides real-time energy supply capabilities (such as hydrogen storage capacity, available workstations, queuing time, and real-time prices).

[0109] C) Intelligent connected vehicles: Report real-time status (such as location, hydrogen level, load, and cabin temperature).

[0110] D) High-precision real-time map: Provides dynamic traffic conditions (such as congestion, gradient, and traffic light sequence);

[0111] E) Cloud Platform: Performs global data fusion, optimization calculation, and decision distribution.

[0112] Develop a three-tiered cloud-edge-device computing architecture to enhance local decision-making capabilities in weak network environments. The device layer is the data source, mainly including vehicle intelligent terminals and warehouse intelligent dashboards. Its core task is to reliably collect and temporarily store data, typically through distributed data collection and lightweight preprocessing, and data temporary storage and retransmission in weak network environments. The edge layer, namely regional edge nodes, is the core of achieving rapid response and autonomous operation in weak network environments. It typically uses open-source edge computing frameworks and stream processing engines to deploy lightweight machine learning models to edge nodes. When network latency with the cloud times out, the edge nodes immediately switch to autonomous mode, such as using high-precision real-time map cache data for localized route planning, using integrated energy station cached guidance information for hydrogen refueling decisions, and adjusting vehicle operating status through vehicle energy management strategies. The cloud focuses on non-real-time, computationally intensive tasks. Through global optimization and model training, it builds a digital twin of city-level logistics transportation, providing optimization and updates to the edge side, and achieving rapid fault location through observability systems such as logs, metrics, and link tracing.

[0113] As an optional embodiment, in the event of a communication interruption between the cloud platform and multiple edge computing nodes, the multiple edge computing nodes schedule the cold chain vehicles within their respective preset communication distances based on the order status data and the current status of the cold chain vehicles.

[0114] Optionally, in the event of a communication interruption between the cloud platform and multiple edge computing nodes, the system automatically enters an edge autonomous operation mode to ensure the continuity, security, and reliability of the cold chain vehicle dispatching service. Without cloud platform command support, each edge computing node independently completes lightweight, real-time dispatching decisions for cold chain vehicles within its jurisdiction based on local caching and real-time sensing data. The specific execution mechanism is as follows:

[0115] When an edge computing node detects a communication link interruption with the cloud platform that persists for more than a preset threshold (e.g., 3 seconds), it immediately triggers a "communication degradation mode activation" mechanism, stops waiting for instructions from the cloud, and switches to local autonomous operation. In this mode, the node no longer relies on a global optimization model, but instead builds a local scheduling decision engine based on a locally persistent historical scheduling knowledge base, high-precision spatiotemporal data cached within the last 5 minutes, and the currently collected real-time order and vehicle status within its jurisdiction. In this mode, each edge computing node only performs scheduling management for cold chain vehicles and orders within its service radius (a preset communication distance, typically 3–10 kilometers).

[0116] Specifically, to ensure system reliability, a three-level degradation strategy can be designed. When communication is interrupted, switch to a local decision-making mode based on historical data; when the algorithm times out, activate the rule engine and allocate tasks according to the proximity principle; in case of an energy emergency, navigate to the nearest hydrogen refueling station, with the specific processing as follows:

[0117] Level 1 Degradation (Communication Interruption > 3 seconds): When an edge node detects a loss of connection with the cloud, it immediately sends a "Communication Degradation Mode Activated" alarm to its affiliated vehicles and dispatchers. The system switches to edge computing-driven mode, stops waiting for cloud instructions, and starts a local decision-making loop. Using pre-synchronized key static data and short-term cached data from the last 5 minutes, combined with prediction algorithms such as Kalman filtering, it makes short-term predictions of road conditions and performs route planning and decision-making based on a preset rule base of safety priority, efficiency priority, and energy security.

[0118] Secondary degradation (algorithm response > 10 seconds): When a complex algorithm (such as CWOA, MILP) fails to output the optimal solution within the specified time, the platform or edge node immediately stops the current complex algorithm and sends a "Algorithm timed out, enable fast solution" notification, triggering the fast response mechanism and enabling the "feasible solution" output by the simplified version of the genetic algorithm to ensure that a feasible solution is output within 1 second.

[0119] Level 3 downgrade (hydrogen level <10%) is the safety baseline response. Its core objective is to ensure vehicle energy safety and avoid breakdowns. When the vehicle's hydrogen level is below 10%, the system triggers the highest level energy alarm and displays a prominent message on the vehicle's screen and the dispatch center's large screen. The system will block all unnecessary task commands and route optimization objectives. The only core command is to navigate to the nearest accessible hydrogen refueling station. At the same time, the platform may attempt to use vehicle-to-infrastructure (V2I) technology to request priority passage through traffic lights for the vehicle and coordinate with the target hydrogen refueling station to prepare for connection.

[0120] According to embodiments of the present invention, a cold chain vehicle dispatching device for implementing the above-described cold chain vehicle dispatching method is also provided. Figure 5 This is a structural block diagram of a cold chain vehicle dispatching device provided according to an embodiment of the present invention, such as... Figure 5 As shown, the cold chain vehicle dispatching device includes: a data acquisition module 502, a first determination module 504, a second determination module 506, a clustering module 508, a partitioning module 510, and a dispatching module 512. The cold chain vehicle dispatching device will be described below.

[0121] The data acquisition module 502 is used to collect the status data corresponding to multiple orders in the target area and the current status of multiple cold chain vehicles. The status data includes at least the order location, time requirement and temperature zone type, and the current status includes the load status, current location and current hydrogen storage.

[0122] The first determining module 504, connected to the acquisition module 502, is used to determine the order distribution density in the target area based on the order locations corresponding to each of the multiple orders.

[0123] The second determining module 506, connected to the first determining module 504, is used to determine clustering parameters based on the order distribution density, wherein the clustering parameters include neighborhood radius and minimum number of points.

[0124] Clustering module 508, connected to second determining module 506, is used to perform density clustering on multiple orders based on clustering parameters to obtain multiple order clusters.

[0125] The partitioning module 510, connected to the clustering module 508, is used to partition multiple order clusters based on the timeliness requirements and temperature zone types corresponding to each order, thereby obtaining multiple task clusters.

[0126] The scheduling module 512, connected to the partitioning module 510, is used to determine the target scheduling scheme based on the respective operating status of multiple task clusters and multiple cold chain vehicles.

[0127] It should be noted that the aforementioned acquisition module 502, first determination module 504, second determination module 506, clustering module 508, partitioning module 510, and scheduling module 512 correspond to steps S202 to S212 in the embodiments. Multiple modules implement the same instances and application scenarios as their corresponding steps, but are not limited to the content disclosed in the above embodiments. It should also be noted that the aforementioned modules, as part of the device, can run on the computer terminal 10 provided in the embodiments.

[0128] Embodiments of the present invention may provide a computer device. Optionally, in this embodiment, the computer device may be located in at least one of a plurality of network devices in a computer network. The computer device includes a memory and a processor.

[0129] The memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the cold chain vehicle scheduling method and apparatus in this embodiment of the invention. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, thereby realizing the aforementioned cold chain vehicle scheduling method. The memory may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include memory remotely located relative to the processor, and these remote memories can be connected to a computer terminal via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0130] The processor can access information and applications stored in the memory via a transmission device to perform the following steps: Collecting status data corresponding to multiple orders in the target area and the current status of multiple cold chain vehicles, wherein the status data includes at least order location, timeliness requirement, and temperature zone type, and the current status includes load status, current location, and current hydrogen storage capacity; determining the order distribution density in the target area based on the order locations corresponding to the multiple orders; determining clustering parameters based on the order distribution density, wherein the clustering parameters include neighborhood radius and target point number parameters; performing density clustering on the multiple orders based on the clustering parameters to obtain multiple order clusters; dividing the multiple order clusters based on the timeliness requirement and temperature zone type corresponding to the multiple orders to obtain multiple task clusters; and determining the target scheduling scheme based on the multiple task clusters and the operating status of the multiple cold chain vehicles.

[0131] Optionally, the processor may also execute program code for the following steps: determining a target scheduling scheme based on the respective operating states of multiple task clusters and multiple cold chain vehicles, including: determining the geometric center of each of the multiple task clusters; determining an initial scheduling scheme based on the respective geometric centers of the multiple task clusters and the respective operating states of the multiple cold chain vehicles; iteratively adjusting the initial scheduling scheme based on a preset sine and cosine operator and calculating the target value corresponding to the new scheduling scheme based on a preset objective function, wherein the preset objective function is determined based on energy cost, cargo loss cost, and time-delay penalty cost; and determining the target scheduling scheme based on the target value corresponding to the new scheduling scheme.

[0132] Optionally, the processor may also execute program code that performs the following steps: when a new order is detected in the target area, obtain the status data corresponding to the new order; based on the status data corresponding to the new order, determine the influencing task cluster from multiple task clusters; determine the influencing cold chain vehicle corresponding to the influencing task cluster in the target scheduling scheme; and send the status data corresponding to the new order to the influencing cold chain vehicle.

[0133] Optionally, the processor may also execute program code for the following steps: in the case of scheduling multiple cold chain vehicles based on a target scheduling scheme, including: collecting real-time data of multiple cold chain vehicles during order execution, wherein the real-time data includes real-time location and real-time remaining hydrogen; obtaining road conditions and energy station operation information in the target area, wherein the energy station operation information includes current hydrogen price, electricity price and number of vehicles in queue; and constructing a scheduling decision model based on road conditions, energy station operation information and real-time data of each of the multiple cold chain vehicles, wherein the scheduling decision model is used to plan order execution routes, charging routes and hydrogen refueling routes.

[0134] Optionally, the processor may also execute program code for the following steps: if the real-time remaining hydrogen quantity corresponding to the first cold chain vehicle is lower than the first hydrogen storage threshold, determine whether there is a hydrogen refueling station within a preset distance from the first cold chain vehicle based on the real-time location of the first cold chain vehicle; if there is a hydrogen refueling station within a preset distance from the first cold chain vehicle, generate a route to the hydrogen refueling station based on the scheduling decision model; and push the route to the first cold chain vehicle.

[0135] This invention provides a method for scheduling cold chain vehicles. It collects status data for multiple orders and the current status of multiple cold chain vehicles in a target area. The status data includes at least order location, timeliness requirement, and temperature zone type; the current status includes load status, current location, and current hydrogen storage. Based on the order locations, the order distribution density in the target area is determined. Clustering parameters, including neighborhood radius and target point number, are determined based on the order distribution density. Density clustering is performed on the multiple orders to obtain multiple order clusters. These order clusters are then divided into multiple task clusters based on the timeliness requirement and temperature zone type of each order. Finally, a target scheduling scheme is determined based on the multiple task clusters and the operating status of each cold chain vehicle. This achieves the goal of intelligently matching the optimal vehicle scheduling scheme, thereby improving scheduling efficiency and reducing transportation costs. Furthermore, it solves the technical problems of current cold chain vehicle cluster energy consumption optimization, which is often limited to local optimization at the single vehicle or single station level, lacking global scheduling capabilities, and traditional optimization algorithms exhibiting slow convergence and susceptibility to local optima.

[0136] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing the hardware related to the terminal device. The program can be stored in a non-volatile storage medium, which may include: flash drive, read-only memory (ROM), random access memory (RAM), disk or optical disk, etc.

[0137] Embodiments of the present invention also provide a non-volatile storage medium. Optionally, in this embodiment, the aforementioned non-volatile storage medium can be used to store the program code executed by the cold chain vehicle scheduling method provided in the above embodiments.

[0138] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any mobile terminal in a group of mobile terminals.

[0139] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: collecting status data corresponding to multiple orders in the target area and the current status corresponding to multiple cold chain vehicles, wherein the status data includes at least order location, timeliness requirement, and temperature zone type, and the current status includes load status, current location, and current hydrogen storage capacity; determining the order distribution density in the target area based on the order locations corresponding to the multiple orders; determining clustering parameters based on the order distribution density, wherein the clustering parameters include neighborhood radius and target point number parameters; performing density clustering on the multiple orders based on the clustering parameters to obtain multiple order clusters; dividing the multiple order clusters based on the timeliness requirement and temperature zone type corresponding to the multiple orders to obtain multiple task clusters; and determining a target scheduling scheme based on the multiple task clusters and the operating status corresponding to the multiple cold chain vehicles.

[0140] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: determining a target scheduling scheme based on the respective operating states of multiple task clusters and multiple cold chain vehicles, including: determining the geometric center corresponding to each of the multiple task clusters; determining an initial scheduling scheme based on the respective geometric centers of the multiple task clusters and the respective operating states of the multiple cold chain vehicles; iteratively adjusting the initial scheduling scheme based on a preset sine and cosine operator and calculating the target value corresponding to the new scheduling scheme based on a preset objective function, wherein the preset objective function is determined based on energy cost, cargo loss cost, and time-delay penalty cost; and determining the target scheduling scheme based on the target value corresponding to the new scheduling scheme.

[0141] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: when a new order is detected in the target area, obtain the status data corresponding to the new order; based on the status data corresponding to the new order, determine the influencing task cluster from multiple task clusters; determine the influencing cold chain vehicle corresponding to the influencing task cluster in the target scheduling scheme; and send the status data corresponding to the new order to the influencing cold chain vehicle.

[0142] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: in the case of scheduling multiple cold chain vehicles based on a target scheduling scheme, the steps include: collecting real-time data of multiple cold chain vehicles during order execution, wherein the real-time data includes real-time location and real-time remaining hydrogen quantity; obtaining road conditions and energy station operation information in the target area, wherein the energy station operation information includes current hydrogen price, electricity price, and number of vehicles in queue; and constructing a scheduling decision model based on road conditions, energy station operation information, and real-time data of each of the multiple cold chain vehicles, wherein the scheduling decision model is used to plan order execution routes, charging routes, and hydrogen refueling routes.

[0143] Optionally, in this embodiment, the non-volatile storage medium is configured to store program code for performing the following steps: when the real-time remaining hydrogen quantity corresponding to the first cold chain vehicle is lower than the first hydrogen storage threshold, based on the real-time location of the first cold chain vehicle, determine whether there is a hydrogen refueling station within a preset distance from the first cold chain vehicle; if there is a hydrogen refueling station within a preset distance from the first cold chain vehicle, generate a route to the hydrogen refueling station based on a scheduling decision model; and push the route to the first cold chain vehicle.

[0144] Embodiments of the present invention also provide a computer program product, including a computer program. Optionally, in this embodiment, when the computer program is executed by a processor, it can: collect status data corresponding to multiple orders in a target area and the current status corresponding to multiple cold chain vehicles, wherein the status data includes at least order location, timeliness requirement, and temperature zone type, and the current status includes load status, current location, and current hydrogen storage capacity; determine the order distribution density in the target area based on the order locations corresponding to the multiple orders; determine clustering parameters based on the order distribution density, wherein the clustering parameters include neighborhood radius and target point number parameters; perform density clustering on the multiple orders based on the clustering parameters to obtain multiple order clusters; divide the multiple order clusters based on the timeliness requirement and temperature zone type corresponding to the multiple orders to obtain multiple task clusters; and determine a target scheduling scheme based on the multiple task clusters and the operating status corresponding to the multiple cold chain vehicles.

[0145] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.

[0146] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0147] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of units can be a logical functional division, and in actual implementation, there may be other division methods. For instance, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the displayed or discussed mutual coupling, direct coupling, or communication connection may be through some interfaces; the indirect coupling or communication connection between units or modules may be electrical or other forms.

[0148] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0149] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0150] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a non-volatile storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0151] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for dispatching cold chain vehicles, characterized in that, include: Collect status data corresponding to multiple orders in the target area and the current status of multiple cold chain vehicles. The status data includes at least the order location, time requirement and temperature zone type, and the current status includes the load status, current location and current hydrogen storage. The order distribution density in the target area is determined based on the order location corresponding to each of the multiple orders. Based on the order distribution density, clustering parameters are determined, wherein the clustering parameters include neighborhood radius and target point number parameters; Based on the clustering parameters, density clustering is performed on the multiple orders to obtain multiple order clusters; Based on the timeliness requirements and temperature zone types corresponding to each of the multiple orders, the multiple order clusters are divided to obtain multiple task clusters; Based on the respective operating status of the multiple task clusters and the multiple cold chain vehicles, a target scheduling scheme is determined.

2. The method according to claim 1, characterized in that, The step of determining the target scheduling scheme based on the respective operating statuses of the multiple task clusters and the multiple cold chain vehicles includes: Determine the geometric center of each of the multiple task clusters; Based on the geometric center of each of the multiple task clusters and the operating status of each of the multiple cold chain vehicles, an initial scheduling scheme is determined. The initial scheduling scheme is iteratively adjusted based on a preset sine and cosine operator, and the target value corresponding to the new scheduling scheme is calculated based on a preset objective function, wherein the preset objective function is determined based on energy cost, cargo loss cost and time penalty cost; Based on the target value corresponding to the new scheduling scheme, the target scheduling scheme is determined.

3. The method according to claim 1, characterized in that, Also includes: If a new order is detected in the target area, obtain the status data corresponding to the new order; Based on the status data corresponding to the new order, the task clusters that affect the order are determined from the multiple task clusters. Determine the affected cold chain vehicles corresponding to the affected task clusters in the target scheduling scheme; The status data corresponding to the new order is sent to the vehicle affecting the cold chain.

4. The method according to claim 1, characterized in that, When scheduling the multiple cold chain vehicles based on the target scheduling scheme, the following applies: Real-time data is collected from the multiple cold chain vehicles during the order execution process, wherein the real-time data includes real-time location and real-time remaining hydrogen quantity; Obtain road conditions and corresponding operational information of energy stations in the target area, wherein the operational information of the energy stations includes the current hydrogen price, electricity price, and number of vehicles in queue; Based on the road conditions, the operation information of the energy station, and the real-time data of each of the multiple cold chain vehicles, a scheduling decision model is constructed. The scheduling decision model is used to plan order execution routes, charging routes, and hydrogen refueling routes.

5. The method according to claim 4, characterized in that, Also includes: If the real-time remaining hydrogen quantity corresponding to the first cold chain vehicle is lower than the first hydrogen storage threshold, based on the real-time location of the first cold chain vehicle, determine whether there is a hydrogen refueling station within a preset distance from the first cold chain vehicle. If a hydrogen refueling station exists within a preset distance from the first cold chain vehicle, a route to the hydrogen refueling station is generated based on the scheduling decision model. The route is pushed to the first cold chain vehicle.

6. A cold chain vehicle dispatching system, characterized in that, include: Multiple terminal devices are deployed in multiple cold chain vehicles and order outlets in the target area to obtain the status data corresponding to each of the multiple orders in the target area and the current status corresponding to each of the multiple cold chain vehicles, and send them to multiple edge computing nodes and cloud platforms. The cloud platform is communicatively connected to the plurality of edge computing nodes and the plurality of terminal devices, and is used to execute the cold chain vehicle scheduling method according to any one of claims 1 to 5 and to distribute the target scheduling scheme to the plurality of edge computing nodes; The multiple edge computing nodes are communicatively connected to the multiple end devices and are used to distribute the target scheduling scheme to the cold chain vehicles within their respective preset communication distances.

7. The system according to claim 6, characterized in that, Also includes: In the event of a communication interruption between the cloud platform and the plurality of edge computing nodes, the plurality of edge computing nodes shall schedule the cold chain vehicles within their respective preset communication distances based on the order status data and the current status of the cold chain vehicles.

8. A cold chain vehicle dispatching device, characterized in that, include: The data acquisition module is used to collect status data corresponding to multiple orders in the target area and the current status of multiple cold chain vehicles. The status data includes at least the order location, time requirement and temperature zone type, and the current status includes the load status, current location and current hydrogen storage. The first determining module is used to determine the order distribution density in the target area based on the order locations corresponding to the multiple orders. The second determining module is used to determine clustering parameters based on the order distribution density, wherein the clustering parameters include neighborhood radius and minimum number of points. The clustering module is used to perform density clustering on the multiple orders based on the clustering parameters to obtain multiple order clusters; The partitioning module is used to partition the multiple order clusters based on the timeliness requirements and temperature zone types corresponding to each of the multiple orders, thereby obtaining multiple task clusters; The scheduling module is used to determine the target scheduling scheme based on the respective operating status of the multiple task clusters and the multiple cold chain vehicles.

9. A non-volatile storage medium, characterized in that, The non-volatile storage medium includes a stored program, wherein, when the program is executed, it controls the device containing the non-volatile storage medium to perform the cold chain vehicle scheduling method according to any one of claims 1 to 5.

10. A computer device, characterized in that, include: Memory and processor The memory stores computer programs; The processor is configured to execute a computer program stored in the memory, wherein when the computer program is executed, the processor performs the cold chain vehicle scheduling method according to any one of claims 1 to 5.

11. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the cold chain vehicle scheduling method according to any one of claims 1 to 5.

Images