Vehicle scheduling method and system and readable storage medium

Abstract

The invention relates to the technical field of intelligent logistics, in particular to a vehicle scheduling method and system and a readable storage medium, and the method comprises the steps: obtaining the platform dynamic information of all platforms in a park, the vehicle information of all to-be-scheduled vehicles, and the current weather state of the park; determining the weather influence degree based on the current weather state and the cargo type of each to-be-dispatched vehicle, and obtaining the dispatching emergency degree based on the weather influence degree of each to-be-dispatched vehicle and the vehicle information; based on the platform dynamic information and the platform static information of all platforms and the vehicle information of each to-be-dispatched vehicle, the matching degree of the to-be-dispatched vehicle and the platform is obtained to determine a distribution matrix between the vehicle and the platform, and based on constraint conditions, the distribution matrix is solved to obtain a dispatching strategy corresponding to each to-be-dispatched vehicle. According to the method, automatic scheduling is performed through the real-time data, so that the risk caused by weather and illegal scheduling is reduced while the scheduling efficiency is improved and the resource utilization rate is optimized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent logistics, in particular to a vehicle scheduling method and system and a readable storage medium. BACKGROUND

[0002] With the rapid development of e-commerce and logistics industry, large logistics parks are increasing, and how to improve the efficiency of vehicle scheduling and ensure the efficient and orderly operation of park vehicles is the top priority of logistics park management.

[0003] Most of the current parks use manual scheduling or rule-based scheduling. In manual scheduling mode, a large number of personnel are arranged to manually schedule at the park post, road, loading and unloading port, which requires a large amount of manual work to achieve. This way is difficult to manage and has low efficiency, which can cause long waiting time for vehicles, low platform utilization, and other problems. The scheduling method based on fixed rules is to automatically lift the lever when the vehicle enters the park, in a first-come-first-served manner, and only through the static attributes of the platform (the loading and unloading port, position, size, purpose, and supported cargo type) to schedule the vehicle to the corresponding loading and unloading port according to the preset rules. This way cannot respond to sudden orders in real time, and the fixed scheduling period leads to long empty waiting time and low running efficiency. SUMMARY

[0004] (I) Technical problems to be solved In view of the above-mentioned shortcomings and deficiencies of the prior art, the present application provides a vehicle scheduling method, system and readable storage medium, which solves the technical problems of high management difficulty and low running efficiency of the scheduling method using manual scheduling or fixed rule-based scheduling.

[0005] (II) Technical solutions In order to achieve the above-mentioned purpose, the main technical solutions adopted by the present application include: In a first aspect, the present application provides a vehicle scheduling method, comprising: Obtaining the platform dynamic information of all platforms in the park, the vehicle information of all vehicles to be scheduled, and the current weather state of the park; the vehicle information includes cargo type and cargo quantity; Based on the current weather state and the cargo type of each vehicle to be scheduled, determining the weather influence degree of each vehicle to be scheduled, and based on the weather influence degree, waiting time, cargo type and cargo quantity of each vehicle to be scheduled, obtaining the scheduling urgency of each vehicle to be scheduled; Based on the platform dynamic information of all platforms, the pre-set platform static information of each platform, and the vehicle information of each vehicle to be scheduled, obtaining the matching degree of each vehicle to be scheduled and each platform; An allocation matrix between the vehicles and the platforms is determined based on the dispatching urgency of each vehicle to be dispatched and the matching degree between each vehicle to be dispatched and each platform, and the allocation matrix is solved based on preset constraints to obtain a dispatching strategy corresponding to each vehicle to be dispatched, so as to dispatch all the vehicles to be dispatched in the park.

[0006] Optionally, in a specific embodiment, the dispatching urgency of each vehicle to be dispatched is obtained based on the weather influence degree, the waiting time, the type of goods and the quantity of goods of each vehicle to be dispatched, and the dispatching urgency of each vehicle to be dispatched is obtained by: determining the type code corresponding to the type of goods of each vehicle; normalizing the weather influence degree, the waiting time, the type code and the quantity of goods of the type of goods of each vehicle to be dispatched to obtain a feature sequence of each vehicle to be dispatched; determining the difference between each feature in the feature sequence of each vehicle to be dispatched and the corresponding feature in the standard sequence, and screening out the maximum difference and the minimum difference; determining the correlation coefficient corresponding to each feature in the feature sequence of each vehicle to be dispatched based on the difference between each feature in the feature sequence of each vehicle to be dispatched and the corresponding feature in the standard sequence, and the maximum difference and the minimum difference, and obtaining the dispatching urgency of each vehicle to be dispatched by weighted average of all correlation coefficients corresponding to the feature sequence of each vehicle to be dispatched.

[0007] Optionally, in a specific embodiment, the dynamic information of the platform includes the use state and the predicted idle time, and the static information of the platform includes the platform type; wherein the matching degree between each vehicle to be dispatched and each platform is obtained based on the dynamic information of all platforms, the pre-set static information of each platform and the vehicle information of each vehicle to be dispatched, and the matching degree between each vehicle to be dispatched and each platform is obtained by: preliminarily screening out the platforms with the use state as idle and the platforms with the predicted idle time less than the preset time window based on the dynamic information of the platform; performing type matching based on the type of goods of each vehicle to be dispatched and the platform type corresponding to each platform after preliminary screening to determine all the platforms matching each vehicle to be dispatched; determining the matching degree between each vehicle to be dispatched and each platform based on the vehicle information of any vehicle to be dispatched and the dynamic information and the static information of the platform corresponding to all the platforms matching the vehicle to be dispatched.

[0008] Optionally, in one specific embodiment, the vehicle information further comprises vehicle position and waiting time, and the platform static information further comprises turnover rate and platform position; wherein, based on the vehicle information of any to-be-scheduled vehicle, and the platform dynamic information and the platform static information corresponding to all platforms matching the to-be-scheduled vehicle, the matching degree between each to-be-scheduled vehicle and each platform is determined, comprising: determining the type matching score between each to-be-scheduled vehicle and the matching platform based on the cargo type of any to-be-scheduled vehicle and the platform type corresponding to all platforms matching the to-be-scheduled vehicle; determining the position score between each to-be-scheduled vehicle and the matching platform based on the vehicle position of any to-be-scheduled vehicle and the platform position corresponding to all platforms matching the to-be-scheduled vehicle; determining the time score between each to-be-scheduled vehicle and the matching platform based on the waiting time of any to-be-scheduled vehicle and the expected idle time corresponding to all platforms matching the to-be-scheduled vehicle; determining the matching degree between each to-be-scheduled vehicle and each platform based on the type matching score, the position score and the time score, and the turnover rate corresponding to all platforms matching the to-be-scheduled vehicle.

[0009] Optionally, in one specific embodiment, the method further comprises: receiving reservation information of a user, and based on the reservation information, screening and releasing the vehicle at the park entrance, and guiding the reserved vehicle into the to-be-scheduled area; the reservation information comprises vehicle information and vehicle license plate information; all reserved vehicles in the to-be-scheduled area are to-be-scheduled vehicles.

[0010] Optionally, in one specific embodiment, before scheduling all to-be-scheduled vehicles in the park, the method further comprises: based on the pre-established park visualization model, simulating the scheduling operation of each to-be-scheduled vehicle to determine whether there is a running conflict in the scheduling process.

[0011] Optionally, in one specific embodiment, the park visualization model is established by the following steps: obtaining model files required for building the park, real geographical position information in the park, and virtual coordinate information of park reference points; the model files comprise at least one 3D model and model elements corresponding to each 3D model; obtaining real coordinate information corresponding to the park reference points according to the virtual coordinate information and the real geographical position information; establishing a first coordinate conversion relationship according to the virtual coordinate information and the real coordinate information; the first coordinate conversion relationship is the conversion relationship between virtual coordinates and real coordinates; According to the first coordinate conversion relationship and the real geographical position information, a virtual coordinate of each 3D model is configured, and all the 3D models are adjusted to corresponding positions; Based on the scene light source and the scene camera, the park scene is built, and the park visualization model is obtained.

[0012] Optionally, in a specific embodiment, the method further comprises: After any reserved vehicle enters the park, all the vehicles in the park are mapped to the park visualization model based on a pre-set vehicle general model; The vehicle position of each reserved vehicle is obtained, so that a visual warning is performed based on the park visualization model and the vehicle position of each reserved vehicle.

[0013] In a second aspect, an embodiment of the present application provides a vehicle scheduling system, comprising an Internet of Things module and a scheduling module which are communicatively connected; The Internet of Things module is configured to obtain platform dynamic information of all platforms in the park, vehicle information of all vehicles to be scheduled, and a current weather state of the park every preset time; the vehicle information comprises a cargo type and a cargo quantity; The scheduling module is configured to determine a weather influence degree of each vehicle to be scheduled based on the current weather state and the cargo type of each vehicle to be scheduled, obtain a scheduling urgency degree of each vehicle to be scheduled based on the weather influence degree of each vehicle to be scheduled, a waiting time, the cargo type and the cargo quantity, obtain a matching degree between each vehicle to be scheduled and each platform based on the platform dynamic information of all platforms, the platform static information pre-set for each platform and the vehicle information of each vehicle to be scheduled, determine an allocation matrix between the vehicles and the platforms based on the scheduling urgency degree of each vehicle to be scheduled and the matching degree between each vehicle to be scheduled and each platform, and solve the allocation matrix based on a pre-set constraint condition to obtain a scheduling strategy corresponding to each vehicle to be scheduled, and schedule all the vehicles to be scheduled in the park.

[0014] In a third aspect, an embodiment of the present application provides a computer readable storage medium, which stores computer instructions, and the computer instructions are run by a processor to execute the vehicle scheduling method described above.

[0015] (Three) beneficial effects The vehicle scheduling method provided by the present application takes the weather influence degree, the waiting time, the cargo type, the cargo quantity and the like into consideration in the urgency degree calculation to perform automatic scheduling, thereby improving the scheduling efficiency, optimizing the resource utilization rate, and reducing the risks caused by the weather and illegal scheduling. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1A vehicle scheduling method flowchart provided for an embodiment of the present application; Figure 2 A vehicle scheduling system structure schematic diagram provided for an embodiment of the present application; Figure 3 A vehicle scheduling system operation flowchart provided for an embodiment of the present application; Figure 4 A vehicle scheduling model training flowchart provided for an embodiment of the present application. DETAILED DESCRIPTION

[0017] In order to better explain the present application, so as to be understood, the present application is described in detail below by specific embodiments, combined with the accompanying drawings.

[0018] With the rapid development of e-commerce and logistics industry, the number of large logistics parks increases, and improving vehicle scheduling efficiency and ensuring orderly operation become the core of park management. At present, most parks use manual scheduling or fixed rule scheduling: manual scheduling requires a large number of personnel to participate, which is difficult to manage and low in efficiency, and is prone to cause long vehicle waiting and low platform utilization; the fixed rule scheduling is mostly "first come, first served", which is only based on the static attribute of the platform for distribution, and cannot respond to sudden orders, and the fixed period causes long empty waiting and low efficiency. The vehicle scheduling method provided by the embodiment of the present application takes into account the weather influence degree, waiting time, cargo type and quantity in the calculation of the urgency of vehicle scheduling, realizes automatic scheduling, and further improves the scheduling efficiency and optimizes the resource utilization rate, while reducing the risks brought by weather and illegal scheduling.

[0019] In order to better understand the above technical solutions, the exemplary embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments described herein. On the contrary, these embodiments are provided in order to enable a clearer, more thorough understanding of the present application and to convey the complete scope of the present application to those skilled in the art.

[0020] The embodiment of the present application provides a vehicle scheduling method, as shown in Figure 1 , comprising: S1, obtaining the platform dynamic information of all platforms in the park, the vehicle information of all vehicles to be scheduled and the current weather state of the park; the vehicle information includes cargo type and cargo quantity.

[0021] Specifically, the vehicle information of any vehicle to be dispatched includes vehicle type (small train / large train / cold chain vehicle, etc.), cargo type (general merchandise / fresh food / emergency medicine / fragile goods, etc.), cargo quantity (weight / volume, etc.), appointment time, waiting time and emergency level, etc.; the platform dynamic information includes usage status (whether it can be used and whether it is running), expected idle time (when it can be idle, when the platform is not in use, the current time), etc.; the external attributes include the current weather state, holidays, etc.

[0022] Further, in the embodiment, the platform dynamic information of all platforms in the park, the vehicle information of all vehicles to be dispatched and the current weather state of the park are acquired, including real-time pushing (initiative sending by the vehicle to be dispatched) and acquisition every preset time.

[0023] S2, based on the current weather state and the cargo type of each vehicle to be dispatched, determine the weather influence degree of each vehicle to be dispatched, and based on the weather influence degree, the waiting time, the cargo type and the cargo quantity of each vehicle to be dispatched, acquire the dispatch emergency level of each vehicle to be dispatched.

[0024] Optionally, in a specific embodiment, based on the weather influence degree, the waiting time, the cargo type and the cargo quantity of each vehicle to be dispatched, the dispatch emergency level of each vehicle to be dispatched is acquired, including: determining the type code corresponding to the cargo type of each cargo; normalizing the weather influence degree, the waiting time, the type code and the cargo quantity of the cargo type of each vehicle to be dispatched to obtain the feature sequence of each vehicle to be dispatched; determining the difference between each feature in the feature sequence of each vehicle to be dispatched and the corresponding feature in the standard sequence, and screening out the maximum difference and the minimum difference; based on the difference between each feature in the feature sequence of each vehicle to be dispatched and the corresponding feature in the standard sequence, and the maximum difference and the minimum difference, determining the correlation coefficient corresponding to each feature in the feature sequence of each vehicle to be dispatched, and weighting and averaging all correlation coefficients corresponding to the feature sequence of each vehicle to be dispatched to obtain the dispatch emergency level of each vehicle to be dispatched.

[0025] Specifically, for example: The real-time collected current weather state (such as rain, high temperature, strong wind, etc.) is divided into multiple levels (usually 5 levels) according to the danger degree to the logistics operation, and the judgment criteria of each level is clear, so that the weather damage can be quantified.

[0026] Then, according to the cargo type of the cargo, the common cargo type is divided into multiple levels (usually 5 levels) according to the degree of vulnerability to weather, and the characteristics of each level are defined to ensure that the vulnerability of the cargo can be quantified.

[0027] The quantified weather hazard and vulnerability of the cargo are substituted into the preset two-dimensional risk matrix, and the value of each cell in the matrix is the score of the weather impact degree.

[0028] For example, when the weather hazard is ≤2 and the vulnerability of the cargo is ≤2, the weather impact degree is ≤2, i.e. light rain (level 2) + sealed general merchandise (level 2) = 2 points; when the weather hazard is ≥4 and the vulnerability of the cargo is ≥4, the weather impact degree is ≥5, i.e. heavy rain (level 4) + fresh food (level 4) = 5 points.

[0029] If there are additional factors such as special packaging of the cargo, protective measures for the operation, etc., the matrix calculation result can be fine-tuned (±1 point, not exceeding the range of 1-5 points) to avoid deviation caused by mechanical application of the matrix.

[0030] An example of the process of obtaining the dispatch urgency degree of each vehicle to be dispatched is as follows: Vehicle A, emergency medicine (cargo type), 2 tons (cargo quantity), 45 minutes (waiting time), 3 (weather impact degree, heavy rain has a slight impact on medicine packaging); Vehicle B, fresh fruits and vegetables (cargo type), 5 tons (cargo quantity), 30 minutes (waiting time), 5 (weather impact degree, heavy rain is easy to cause fruits and vegetables to be damp and deteriorate); Vehicle B, emergency medicine (cargo type), 2 tons (cargo quantity), 45 minutes (waiting time), 3 (weather impact degree, heavy rain has little effect on sealed general merchandise).

[0031] The cargo type is coded based on the pre-set cargo type coding rule, such as emergency medicine = 5, fresh fruits and vegetables = 3, and general merchandise = 1. This rule is based on the urgency and value of the cargo type.

[0032] The weather impact degree, waiting time, type code, and cargo quantity of each vehicle to be dispatched are normalized to map all data to the interval [0, 1] to eliminate the magnitude difference of different units of data. That is, the linear normalization formula is used to process each of the four features (weather impact degree, waiting time, type code, and cargo quantity) of each vehicle to be dispatched, and the formula is: normalized value = (actual value - minimum value of the feature) / (maximum value of the feature - minimum value of the feature), where the minimum value and the maximum value of each feature are pre-set values.

[0033] According to the priority of the park operation, the allocation weight of each feature (weather impact degree, waiting time, type code, and cargo quantity) is set.

[0034] The standard sequence is set based on the historical optimal scheduling data of the yard, representing the ideal characteristic state with the highest emergency degree, wherein the weather influence degree = 5 (normalized to 1), the waiting time = 60 min (normalized to 1), the type code = 5 (normalized to 1), and the cargo quantity ≥ 10 tons (normalized to 1). That is, the standardized sequence is [1, 1, 1, 1].

[0035] Suppose the characteristic sequence of vehicle A is [0.33, 1, 1, 0], the characteristic sequence of vehicle B is [1, 0.5, 0.5, 0.5], and the characteristic sequence of vehicle C is [0, 0, 0, 0]; The difference between the characteristic sequence and the standard sequence is calculated for each vehicle (difference = standard sequence value - characteristic sequence value), and the maximum and minimum differences for each vehicle are filtered. The calculation formula of the correlation coefficient corresponding to any feature is: (minimum difference + 0.5 × maximum difference) / (feature difference + 0.5 × maximum difference), and 0.5 is a pre-set resolution coefficient for reducing the influence of the maximum difference on the result. Taking vehicle A as an example, the differences between it and the standard sequence are 0.67, 0, 0, and 0, and the minimum difference of the characteristic sequence is 0 and the maximum difference is 0.67. The weather influence correlation coefficient = (0.00 + 0.5) / (0.67 + 0.5) ≈ 0.5 / 1.17 ≈ 0.43, and the remaining features are calculated in the same way.

[0036] The correlation coefficients are weighted and summed according to the pre-set weights (weather influence 0.3, waiting time 0.25, type code 0.3, and cargo quantity 0.15), and the result is the scheduling emergency degree (the closer the value is to 1, the higher the emergency degree).

[0037] The embodiments of the present application calculate the scheduling emergency degree based on the weather influence degree, waiting time, cargo type, and cargo quantity. Through the process of quantization, normalization, correlation analysis, and weighted decision-making, the multi-dimensional factors are quantitatively processed to avoid subjective experience bias. By calculating the difference between the characteristic sequence and the standard sequence and the correlation coefficient, the closeness of each vehicle to the ideal emergency state (standard sequence) is measured, so that the attention is more focused and the accuracy of subsequent decision-making is ensured.

[0038] S3, based on the platform dynamic information of all platforms, the pre-set platform static information of each platform, and the vehicle information of each to-be-scheduled vehicle, the matching degree of each to-be-scheduled vehicle and each platform is obtained.

[0039] Optionally, in a specific embodiment, based on the platform dynamic information of all platforms, the pre-set platform static information of each platform, and the vehicle information of each to-be-scheduled vehicle, the matching degree of each to-be-scheduled vehicle and each platform is obtained, including: Based on the platform dynamic information, the idle platforms and the platforms with an idle time less than a preset time window are preliminarily screened out; Based on the cargo type of each vehicle to be dispatched and the platform type corresponding to each platform after the preliminary screening, type matching is performed to determine all platforms matched with each vehicle to be dispatched. Based on the vehicle information of any vehicle to be dispatched and the platform dynamic information and the platform static information corresponding to all platforms matched with the vehicle to be dispatched, the matching degree of each vehicle to be dispatched and each platform is determined.

[0040] Further, based on the vehicle information of any vehicle to be dispatched and the platform dynamic information and the platform static information corresponding to all platforms matched with the vehicle to be dispatched, the matching degree of each vehicle to be dispatched and each platform is determined, including: Based on the cargo type of any vehicle to be dispatched and the platform type corresponding to all platforms matched with the vehicle to be dispatched, the type matching score between each vehicle to be dispatched and the matched platform is determined. Based on the vehicle position of any vehicle to be dispatched and the platform position corresponding to all platforms matched with the vehicle to be dispatched, the position score between each vehicle to be dispatched and the matched platform is determined. Based on the waiting time of any vehicle to be dispatched and the estimated idle time corresponding to all platforms matched with the vehicle to be dispatched, the time score between each vehicle to be dispatched and the matched platform is determined. Based on the type matching score, the position score and the time score, and the turnover rate corresponding to all platforms matched with the vehicle to be dispatched, the matching degree of each vehicle to be dispatched and each platform is determined.

[0041] Specifically, there is a hard match and a soft match between the platform and the vehicle to be dispatched, wherein the hard match is as follows: The cargo type and the platform type are incompatible (such as a vehicle transporting cold chain goods but the platform is not a cold chain), the size is out of limit (the length of the vehicle > the maximum length allowed by the platform), the use state of the platform is busy, and the platform will not be idle for a period of time (i.e., the estimated idle time is greater than the preset time window), etc.

[0042] The soft match: first, type matching is performed, and the type matching score S1 ∈ [0, 1], such as complete matching: cold chain vehicle → cold chain platform S1 = 1.0, partial matching: refrigerated vehicle → ordinary platform S1 = 0.3, and mismatch: dangerous chemical vehicle → ordinary platform S1 = 0, etc.

[0043] Based on the vehicle position of any vehicle to be dispatched and the platform position corresponding to all platforms matched with the vehicle to be dispatched, the position score between each vehicle to be dispatched and the matched platform is determined. The calculation formula of the position score is: ; wherein, is a position score between the vehicle to be dispatched indexed i and the platform indexed j, d i，j is a distance between the vehicle to be dispatched indexed i and the platform indexed j, is a maximum distance between the vehicle to be dispatched indexed i and each platform matching the type of the vehicle to be dispatched; based on the dispatch urgency of each vehicle to be dispatched, the matching degree between each vehicle to be dispatched and each platform, an allocation matrix between the vehicle and the platform is determined, and the allocation matrix is solved based on a pre-set constraint condition to obtain a dispatch strategy corresponding to each vehicle to be dispatched, and all vehicles to be dispatched in the park are dispatched.

[0044] based on the waiting time of any vehicle to be dispatched and the predicted idle time corresponding to all platforms matching the vehicle to be dispatched, a time score between each vehicle to be dispatched and the matching platform is determined; the calculation formula of the time score is: ; wherein, is a time score between the vehicle to be dispatched indexed i and the platform indexed j, is a predicted idle time of the platform indexed j, t now is a current time, t max is a pre-set maximum waiting time, is a waiting time of the vehicle to be dispatched indexed i.

[0045] The turnover rate of each platform is standardized to obtain an efficiency score of each platform.

[0046] based on the type matching score, the position score and the time score, and the turnover rate corresponding to all platforms matching the vehicle to be dispatched, the matching degree between each vehicle to be dispatched and each platform is determined, which is usually weighted and summed to achieve.

[0047] The vehicle-platform matching is replaced from an experience-driven random allocation to a data-driven accurate decision, through the whole-link logic of screening-scoring-solving, it is ensured that each vehicle can be matched to a platform with function adaptation, appropriate distance and time coincidence, and it is also ensured that the dispatch scheme meets the actual constraints of the park (equipment, route, load), which ensures the platform utilization rate and improves the dispatch efficiency.

[0048] S4, based on the scheduling urgency of each vehicle to be dispatched, the matching degree of each vehicle to be dispatched and each platform, determine the allocation matrix between the vehicle and the platform, and solve the allocation matrix based on the pre-set constraint conditions to obtain the scheduling strategy corresponding to each vehicle to be dispatched, and dispatch all vehicles to be dispatched in the park.

[0049] Specifically, there are m vehicles to be dispatched and n available platforms, the scheduling urgency of vehicle i is E i , and the matching degree of vehicle i and platform j is M i,j At this time, the comprehensive scheduling value matrix Z i,j corresponding to each vehicle to be dispatched can be obtained. i + M i,j ; A maximum weight matching problem of a bipartite graph to be constrained (allocation matrix) is constructed, the left node is a set of vehicles to be dispatched, the right node is a set of available platforms, the edge weight is the comprehensive scheduling value matrix Z i,j =E i +M i,j , the target is to maximize the total scheduling value ∑ i, j Z i,j ·x i,j , i is the index of the vehicle to be dispatched, j is the index of the platform, x i,j =1 / 0, equal to 1 indicates that vehicle i is allocated to platform j, and equal to 0 indicates that it is not allocated.

[0050] The following constraint conditions are introduced, each vehicle can only be allocated to one platform, and each platform can serve at most one vehicle at the same time (which can be extended to multiple vehicles).

[0051] The Hungarian algorithm, integer linear programming algorithm or greedy algorithm is used to solve the optimal solution.

[0052] In this embodiment, by constructing the comprehensive scheduling value matrix of scheduling urgency + matching degree, the allocation problem of m vehicles to be dispatched and n available platforms is converted into a maximum weight matching problem of a bipartite graph, the target is to maximize the total scheduling value, rigid constraints such as each vehicle being allocated to only one platform and each platform serving at most one vehicle at the same time (which can be extended) are introduced, and the Hungarian algorithm (small-scale precise solution), the greedy algorithm (large-scale fast response) or the integer linear programming algorithm (complex constraint adaptation) is used to solve the optimal solution, and finally the scheduling strategy is generated. This process not only balances the vehicle priority and platform adaptability to achieve the optimal global scheduling benefit, but also ensures that the scheme is compliant and can be implemented, and can adapt to different park sizes, operation rhythms and long-term iteration requirements through various algorithms and flexible architecture, effectively reducing the disorder and loss of traditional scheduling, and significantly improving the vehicle scheduling efficiency, platform resource utilization and scheme execution reliability.

[0053] Further, the method further comprises: receiving reservation information of the user, and screening and releasing the vehicle at the park entrance based on the reservation information, and guiding the reserved vehicle to enter a waiting dispatch area; the reservation information comprises vehicle information and vehicle license information; all the reserved vehicles in the waiting dispatch area are the vehicles to be dispatched.

[0054] Further, when the reserved vehicle is late, a late penalty is given to the vehicle to be dispatched in the waiting area, that is, the same waiting time, the on-time arrival is prior to the late arrival, the on-time arrival can be +0.5 points, and the late arrival can be-0.5 points.

[0055] Further, before dispatching all the vehicles to be dispatched in the park, the method further comprises: based on the pre-established visual model of the park, simulating dispatching operation of each vehicle to be dispatched to determine whether a running conflict occurs in the dispatching process.

[0056] The visual model of the park is established by the following steps: obtaining model files required for building the park, real geographical position information in the park, and virtual coordinate information of park reference points; the model files comprise at least one 3D model and model elements corresponding to each 3D model; obtaining real coordinate information corresponding to the park reference points according to the virtual coordinate information and the real geographical position information; establishing a first coordinate conversion relationship according to the virtual coordinate information and the real coordinate information; the first coordinate conversion relationship is a conversion relationship between the virtual coordinate and the real coordinate; configuring virtual coordinates of each 3D model according to the first coordinate conversion relationship and the real geographical position information, and adjusting all the 3D models to corresponding positions; building a park scene based on scene light sources and scene cameras to obtain the visual model of the park.

[0057] Further, based on current weather state information, the visual model of the park is given corresponding weather characteristics, that is, the visual model automatically switches between sunny, rainy, snowy and other scene modes.

[0058] At this time, the method further comprises: after any reserved vehicle enters the park, mapping all the vehicles in the park to the visual model of the park based on a pre-set general vehicle model; obtaining a vehicle position of each reserved vehicle, so as to perform visual early warning based on the visual model of the park and the vehicle position of each reserved vehicle.

[0059] The vehicle scheduling method provided by the embodiment realizes intelligent matching of vehicles and platforms, reduces empty waiting, mis-stopping and detours, improves platform turnover rate, and avoids dependence on manual experience.

[0060] In addition, the embodiment of the application provides a vehicle scheduling system, which comprises an Internet of Things module and a scheduling module in communication with each other. The Internet of Things module is configured to acquire platform dynamic information of all platforms in the park, vehicle information of all vehicles to be scheduled, and a current weather state of the park every preset time interval, and the vehicle information comprises a cargo type and a cargo quantity. The scheduling module is configured to determine a weather influence degree of each vehicle to be scheduled based on the current weather state and the cargo type of each vehicle to be scheduled, acquire a scheduling urgency degree of each vehicle to be scheduled based on the weather influence degree of each vehicle to be scheduled, a waiting time length, the cargo type and the cargo quantity of each vehicle to be scheduled, acquire a matching degree between each vehicle to be scheduled and each platform based on the platform dynamic information of all platforms, platform static information pre-set for each platform and the vehicle information of each vehicle to be scheduled, determine an allocation matrix between the vehicles to be scheduled and the platforms based on the scheduling urgency degree of each vehicle to be scheduled and the matching degree between each vehicle to be scheduled and each platform, and solve the allocation matrix based on a pre-set constraint condition to obtain a scheduling strategy corresponding to each vehicle to be scheduled and schedule all vehicles to be scheduled in the park.

[0061] Specifically, the vehicle scheduling system provided by the embodiment comprises a reservation module, a twin module, a scheduling module and an Internet of Things module connected with each other through a gateway, as shown in Figure 2 Specifically, the vehicle scheduling system provided by the embodiment comprises a reservation module, a twin module, a scheduling module and an Internet of Things module connected with each other through a gateway, as shown in

[0062] Specifically, the operation process of the vehicle scheduling system comprises the following steps, as shown in Figure 3 Randomly sample the data set, extract 80% as a training set, 10% as a validation set, and the remaining 10% as a test set. The model training process comprises the following steps, as shown in Figure 4 Training set: used to train the model. During training, the model repeatedly looks at the data in the training set, adjusts its parameters, and makes the prediction result as close to the correct answer as possible.

[0063] Validation set: used for parameter tuning. During the training process, the model does not contact the results of the validation set and only makes predictions.

[0064] Test set: used to evaluate the accuracy of the model. It is completely independent data that the model has never contacted. After training and parameter tuning are completed, the test set is used to test the final ability of the model.​​

[0065] Based on TensorFlow, a machine learning model is built using a Transformer architecture to train a park vehicle scheduling model.

[0066] The steps are: feature input, including vehicle information, platform dynamic information, platform static information, external attributes (weather state information).

[0067] A loss function is constructed to calculate the deviation between the predicted value and the actual value. Gradient descent algorithm is used when the loss function is minimized (r' represents the predicted loading and unloading time, r represents the actual result value, and n represents the number of training sets) to determine the weight values in the above method.

[0068] To reduce the randomness of data division and initialization, after completing one repeat training, resample until repeat = 10 times, stop training and get stable results. Compare the minimum loss function value of 10 times of training, take the minimum training result as the final model. If the output result is close to the actual result, the verification is passed, otherwise we adjust the iteration times, learning rate and other hyperparameters according to the result of the validation set to optimize the performance of the model on the validation set. Finally, use the test set data to simulate the application effect of the model in the actual situation. Through the performance on the test set, it can be judged whether the model has the ability to process new data. If the accuracy is not up to standard, increase the data sample size, adjust the input features, increase the training times, and repeat the above steps until the predicted value meets the expectations.

[0069] The Internet of Things module accesses the park's Internet of Things devices through http protocol, 485 serial port, and MQTT protocol, including park gate barrier, top-mounted snapshot machine, waiting area LED information display screen, waiting area broadcast equipment, and park road along the way camera. Remote viewing and control functions of Internet of Things devices are realized.

[0070] A twin module is established by collecting park building space information, park roads, platform, camera position information, and other related data, including building CAD drawings, on-site photos, map data, coordinates, etc. Based on the collected data, a three-dimensional visual model of the park is established, and a virtual world of the park digital twin (park visualization model) corresponding to the physical entity is established.

[0071] The logistics vehicle driver makes an appointment for the time of entering the park through the appointment module. The appointment module inputs the vehicle information, the planned time of entering the park and the weather characteristics to the scheduling module. The scheduling module outputs the suggested time of entering the park according to the loading and unloading time of the vehicles that have been reserved for the day, so as to avoid the crowded time period. The Internet module issues the license plate and the time of entering the park to the gate system, so that the gate can automatically lift the lever to release the vehicle when the vehicle enters the park at the reserved time.

[0072] The logistics vehicle driver arrives at the park entrance before and after the reserved time. The gate automatically lifts the lever to release the vehicle, and the vehicle stops in the waiting area. At this time, the Internet module sends the vehicle has arrived instruction to the vehicle scheduling model after monitoring the passing record of the gate system. The vehicle scheduling model adds the vehicle to the queue of vehicles waiting for scheduling.

[0073] The real-time state of the platform is obtained. A snapshot machine is installed on each platform. The snapshot machine can obtain the license plate of the current platform operation through image recognition. When there is no vehicle parked, the obtained license plate is empty. The Internet module obtains the idle or working state of the platform and pushes it as an input feature to the scheduling module.

[0074] There are usually multiple vehicles waiting for scheduling in the waiting area. The scheduling module inputs the vehicle information, the static information of the platform, the dynamic information of the platform and the current weather state information to the vehicle scheduling model to select which vehicle to schedule to the platform loading and unloading port with the highest efficiency of the park vehicle operation as the goal. The specific scheduling algorithm in the vehicle scheduling model is described in the foregoing vehicle scheduling method. For example, a vehicle that arrives on time (reserved time) at the waiting area will be prioritized over a vehicle that arrives late. A vehicle carrying emergency rescue medicine will be prioritized over a vehicle carrying ordinary goods. The probability of vehicle scheduling to the underground platform is different because the operation time of the underground platform is usually longer than that of the ground platform according to historical data.

[0075] The vehicle scheduling model interfaces with the Internet module and notifies the scheduled vehicle to drive into the specified platform through the waiting area LED information release screen and the broadcasting equipment.

[0076] The park visual display and early warning are realized through the digital twin module to display the vehicle position distribution, vehicle queuing condition, platform vehicle operation condition and monitor whether the park road is smooth. The digital twin platform is connected with the Internet module to find the vehicle illegal parking or the park road vehicle congestion and visually warn in the first time. The historical track route of the vehicle is clearly drawn, the incident location is located one key, and the surrounding camera monitoring screen is linked and called. The park vehicle operation state can be more directly mastered and the corresponding emergency handling measures can be executed to help the park management personnel to make correct and timely handling measures and ensure the smooth operation of the park vehicle operation.

[0077] The embodiment provided in the application improves the vehicle operation efficiency of the park, improves the operation efficiency, reduces the manual factors, reduces the labor cost and the operation cost, and avoids vehicle congestion and collision by using the digital twinning, records the complete operation track, and improves the safety management and control of the park.

[0078] In a third aspect, the embodiment provides a computer readable storage medium, which stores computer instructions. When the computer instructions are run by a processor, the vehicle scheduling method described above is executed.

[0079] In the description of the present application, it should be understood that the terms "first", "second" are only used for descriptive purpose, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Therefore, the features defined with "first", "second" can explicitly or implicitly include one or more of the features. In the description of the present application, the meaning of "multiple" is two or more, unless otherwise specifically limited.

[0080] In the present application, unless otherwise specifically defined and limited, the terms "mounting", "connection", "connection", "fixing" and the like should be understood in a broad sense, for example, it can be fixed connection, or detachable connection, or integral; it can be mechanical connection, or electrical connection; it can be directly connected, or indirectly connected through intermediate medium; it can be the internal communication of two elements or the interaction relationship between two elements. For those skilled in the art, the specific meaning of the above terms in the present application can be understood according to the specific circumstances.

[0081] In the present application, unless otherwise specifically defined and limited, the first feature is "on" or "under" the second feature, which can be direct contact between the first and second features, or indirect contact between the first and second features through intermediate medium. Moreover, the first feature is "above", "over" and "on" the second feature, which can be directly above or obliquely above the first feature, or only indicates that the horizontal height of the first feature is higher than that of the second feature. The first feature is "below", "under" and "under" the second feature, which can be directly below or obliquely below the first feature, or only indicates that the horizontal height of the first feature is lower than that of the second feature.

[0082] In the description of the specification, the description of the terms "one embodiment", "some embodiments", "an embodiment", "example", "specific example" or "some examples" etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are contained in at least one embodiment or example of the present application. In the specification, the illustrative description of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any appropriate manner in any one or more embodiments or examples. In addition, the person skilled in the art can combine and combine the different embodiments or examples described in the specification and the features of the different embodiments or examples without contradiction.

[0083] Although the embodiments of the present application have been shown and described above, it is understood that the above embodiments are exemplary and are not to be construed as limiting the present application, and the person skilled in the art can modify, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A vehicle dispatching method characterized by comprising: The method comprises the following steps: acquiring platform dynamic information of all platforms in the park, vehicle information of all vehicles to be dispatched, and a current weather state of the park; the vehicle information comprises a cargo type and a cargo quantity; determining a weather influence degree of each vehicle to be dispatched based on the current weather state and the cargo type of each vehicle to be dispatched, and acquiring a dispatch urgency degree of each vehicle to be dispatched based on the weather influence degree of each vehicle to be dispatched, a waiting time, the cargo type and the cargo quantity; acquiring a matching degree of each vehicle to be dispatched and each platform based on the platform dynamic information of all platforms, platform static information of each platform set in advance and the vehicle information of each vehicle to be dispatched; determining an allocation matrix between the vehicles and the platforms based on the dispatch urgency degree of each vehicle to be dispatched and the matching degree of each vehicle to be dispatched and each platform, and solving the allocation matrix based on a constraint condition set in advance to obtain a dispatch strategy corresponding to each vehicle to be dispatched, and dispatching all vehicles to be dispatched in the park.

2. The vehicle dispatching method according to claim 1, characterized by, The method for acquiring the dispatch urgency degree of each vehicle to be dispatched based on the weather influence degree of each vehicle to be dispatched, the waiting time, the cargo type and the cargo quantity comprises the following steps: determining a type code corresponding to the cargo type of each cargo; normalizing the weather influence degree, the waiting time, the type code and the cargo quantity of the cargo type of each vehicle to be dispatched to obtain a feature sequence of each vehicle to be dispatched; determining a difference value between each feature in the feature sequence of each vehicle to be dispatched and a corresponding feature in a standard sequence, and screening out a maximum difference value and a minimum difference value; determining a correlation coefficient corresponding to each feature in the feature sequence of each vehicle to be dispatched based on the difference value between each feature in the feature sequence of each vehicle to be dispatched and the corresponding feature in the standard sequence, and the maximum difference value and the minimum difference value, and obtaining the dispatch urgency degree of each vehicle to be dispatched by weighted average of all correlation coefficients corresponding to the feature sequence of each vehicle to be dispatched.

3. The vehicle dispatching method according to claim 1, characterized by, The platform dynamic information comprises a use state and a predicted idle time, and the platform static information comprises a platform type; wherein the method for acquiring the matching degree of each vehicle to be dispatched and each platform based on the platform dynamic information of all platforms, the platform static information of each platform set in advance and the vehicle information of each vehicle to be dispatched comprises the following steps: preliminarily screening out platforms with an idle use state and platforms with a predicted idle time less than a preset time window based on the platform dynamic information; determining all platforms matched with each vehicle to be dispatched based on the cargo type of each vehicle to be dispatched and the platform type of each platform after the preliminary screening; determining the matching degree of each vehicle to be dispatched and each platform based on the vehicle information of any vehicle to be dispatched and the platform dynamic information and the platform static information of all platforms matched with the vehicle to be dispatched.

4. The vehicle dispatching method according to claim 3, characterized by, The vehicle information further includes vehicle position and waiting time, and the platform static information further includes turnover rate and platform position; wherein, based on the vehicle information of any to-be-scheduled vehicle, and the platform dynamic information and the platform static information corresponding to all platforms matching the to-be-scheduled vehicle, the matching degree between each to-be-scheduled vehicle and each platform is determined, including: Based on the cargo type of any to-be-scheduled vehicle, and the platform type corresponding to all platforms matching the to-be-scheduled vehicle, the type matching score between each to-be-scheduled vehicle and the matching platform is determined; Based on the vehicle position of any to-be-scheduled vehicle, and the platform position corresponding to all platforms matching the to-be-scheduled vehicle, the position score between each to-be-scheduled vehicle and the matching platform is determined; Based on the waiting time of any to-be-scheduled vehicle, and the expected idle time corresponding to all platforms matching the to-be-scheduled vehicle, the time score between each to-be-scheduled vehicle and the matching platform is determined; Based on the type matching score, the position score and the time score, and the turnover rate corresponding to all platforms matching the to-be-scheduled vehicle, the matching degree between each to-be-scheduled vehicle and each platform is determined.

5. The vehicle dispatching method according to any one of claims 1 to 4, characterized by, The method further includes: Receiving reservation information of a user, and based on the reservation information, screening and releasing the vehicle at the entrance of the park, and guiding the reserved vehicle into the to-be-scheduled area; the reservation information includes vehicle information and vehicle license plate information; all reserved vehicles in the to-be-scheduled area are to-be-scheduled vehicles.

6. The vehicle dispatching method according to any one of claims 1 to 4, characterized by, Before scheduling all to-be-scheduled vehicles in the park, the method further includes: Based on the pre-established visual model of the park, simulating the scheduling operation of each to-be-scheduled vehicle to determine whether there is a running conflict in the scheduling process.

7. The vehicle dispatching method according to claim 6, characterized by, The visual model of the park is established by the following steps: Obtaining model files required for building the park, real geographical position information in the park, and virtual coordinate information of park reference points; the model files include at least one 3D model and model elements corresponding to each 3D model; According to the virtual coordinate information and the real geographical position information, obtaining real coordinate information corresponding to the park reference points; According to the virtual coordinate information and the real coordinate information, establishing a first coordinate conversion relationship; the first coordinate conversion relationship is the conversion relationship between virtual coordinates and real coordinates; According to the first coordinate conversion relationship and the real geographical position information, configuring the virtual coordinates of each 3D model, and adjusting all 3D models to the corresponding positions; Based on scene light sources and scene cameras, the park scene is built to obtain the visual model of the park.

8. The vehicle dispatching method according to claim 7, characterized by, The method further includes: After any reserved vehicle enters the park, based on the pre-set vehicle general model, all vehicles in the park are mapped to the visual model of the park; Obtaining the vehicle position of each reserved vehicle, so as to perform visual early warning based on the visual model of the park and the vehicle position of each reserved vehicle.

9. A vehicle dispatching system characterized by comprising: Including an Internet of Things module and a scheduling module connected in communication with each other; The Internet-of-things module is configured to acquire platform dynamic information of all platforms in the park, vehicle information of all to-be-scheduled vehicles, and a current weather state of the park every preset time interval, and the vehicle information includes a cargo type and a cargo quantity; The scheduling module is configured to determine a weather influence degree of each to-be-scheduled vehicle based on the current weather state and the cargo type of each to-be-scheduled vehicle, acquire a scheduling urgency degree of each to-be-scheduled vehicle based on the weather influence degree, a waiting time length, the cargo type, and the cargo quantity of each to-be-scheduled vehicle, acquire a matching degree between each to-be-scheduled vehicle and each platform based on the platform dynamic information of all platforms, platform static information pre-set for each platform, and the vehicle information of each to-be-scheduled vehicle, determine an allocation matrix between vehicles and platforms based on the scheduling urgency degree of each to-be-scheduled vehicle and the matching degree between each to-be-scheduled vehicle and each platform, and solve the allocation matrix based on a pre-set constraint condition to obtain a scheduling strategy corresponding to each to-be-scheduled vehicle, and schedule all to-be-scheduled vehicles in the park.

10. A computer readable storage medium storing computer instructions, characterized in that, The computer instructions, when executed by the processor, perform the vehicle scheduling method of any one of claims 1-8.

Images