Freight return smoothness quantification method and system based on space-time utility calculation

Abstract

The invention provides a freight return smoothness quantification method and system based on space-time utility calculation, and the method comprises the following steps: 1, obtaining driver state data, candidate order data, dynamic road network data and driver preference weight parameters, and carrying out the preprocessing of the driver state data, the candidate order data and the dynamic road network data; 2, planning a reference return path and a detour path through the obtained data, and calculating reference return cost and detour cost; 3, calculating the net bypassing cost through the reference return cost and the bypassing cost; 4, calculating estimated arrival time through the detour path, and calculating a time penalty factor in combination with the expected arrival time of the driver; 5, calculating the smoothness through the driver preference weight parameter, the net detour cost and the time penalty factor, and carrying out the decision of a candidate order through the smoothness; according to the method, the driver-order dynamic model is constructed, so that accurate quantification of the on-the-way willingness of the driver is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of management system, and in particular to a freight return route degree of convenience quantification method and system based on space-time utility calculation. BACKGROUND

[0002] With the increasing demand for individualization in the logistics market, the appeal of drivers for non-monetary utility (such as returning home along the way and controllable time) is increasingly prominent. Traditional vehicle and cargo matching technology simplifies drivers as homogenized transport units, and only uses freight or path efficiency as the basis for decision-making, resulting in high empty return rate, rising order rejection rate and idle transport resources. Therefore, the platform urgently needs to convert the subjective "convenient route" appeal of drivers into a calculable parameter, but the existing technology lacks scientific quantification means.

[0003] In recent years, China's policy has clearly required the logistics system to transform towards both driver experience and efficiency. However, the existing technology has the defect that the static model is detached from the complex decision-making scenario, and has the problems of missing preference modeling, insufficient space-time coupling and weak expansion capability. Therefore, how to build a real-time quantitative mapping of driver preference-order feature-road network state has become a key obstacle to improving matching efficiency. SUMMARY

[0004] Therefore, the purpose of the present application is to provide a freight return route degree of convenience quantification method and system based on space-time utility calculation to at least solve the above problems.

[0005] The technical solution adopted by the present application is as follows: The present application provides a freight return route degree of convenience quantification method based on space-time utility calculation in the first aspect, which comprises the following steps: Step 1: Obtain driver state data, candidate order data, dynamic road network data and driver preference weight parameters, and preprocess the driver state data, candidate order data and dynamic road network data; Step 2: Plan the benchmark return path and detour path through the obtained data, and calculate the benchmark return cost and detour cost; Step 3: Calculate the net detour cost through the benchmark return cost and detour cost; Step 4: Calculate the estimated arrival time through the detour path, and calculate the time penalty factor in combination with the driver's expected arrival time; Step 5: Calculate the degree of convenience through the driver preference weight parameters, net detour cost and time penalty factor, and make a decision on the candidate order through the degree of convenience.

[0006] Further, the driver state data, candidate order data and dynamic road network data obtained in step 1 are as follows: Obtain the driver state data: obtain the current position of the driver, the return destination of the driver, and the expected arrival time of the driver; Obtain candidate order data: Obtain the pickup point location, delivery point location, and shipping revenue of candidate orders; Obtain dynamic road network data: Through the API interface of third-party map service providers, obtain real-time road network information covering the area where the driver's current location, the driver's return destination, the pickup point location, and the delivery point location are located.

[0007] Furthermore, the preprocessing in step 1 includes outlier filtering, road network data smoothing, missing data filling, noise removal, and data standardization.

[0008] Furthermore, the specific steps for obtaining the driver preference weight parameters in step 1 are as follows: For drivers with abundant historical data, all their historical order decision data are collected to construct a dataset. After feature learning and normalization of the dataset, the driver's preference weights are generated. For drivers with insufficient historical data, default preference weights are generated by filling out a pre-defined questionnaire.

[0009] Furthermore, step 2 specifically involves: The optimal route is planned based on the driver's current location and return destination from the driver status data, serving as the baseline return route. The baseline return cost is then calculated using the baseline equation path.

[0010] in, The estimated time cost for the driver to travel from the current location to the destination. Estimated economic cost for the driver to travel from the current location to the destination; Based on the return trip cost; The optimal route is planned as a detour route using driver status data and candidate order data, including the driver's current location, pickup point location, delivery point location, and driver's return destination. The detour cost is then calculated based on the detour route.

[0011] in, The estimated time cost for the entire detour route. The estimated economic cost of the entire detour route; Cost of detour.

[0012] Furthermore, step 3 specifically involves: The net detour cost of this candidate order is obtained by comparing the baseline return cost and the detour cost:

[0013] in, These represent the additional time and money the driver needs to spend to complete this candidate order.

[0014] Furthermore, step 4 specifically involves: The estimated arrival time is calculated by taking the detour route, and a time penalty factor is calculated by combining this with the driver's expected arrival time.

[0015] in, As a time penalty factor, The preset sensitivity coefficient, The estimated arrival time is calculated using the detour route. This represents the driver's expected arrival time.

[0016] Furthermore, step 5 specifically involves: The net utility of the candidate order for the driver is calculated using driver preference weighting parameters, net detour cost, and time penalty factor:

[0017] in, The net utility of this candidate order for the driver is... For the shipping revenue of this candidate order, The driver preference weights, obtained through analysis, reflect the drivers' relative sensitivity to time and economic costs, respectively. The final path length is obtained by mapping the net utility using the Sigmoid function:

[0018] in, For the sake of convenience and , This is the scaling factor.

[0019] A second aspect of the present invention provides a freight return route quantification system based on spatiotemporal utility calculation. The system is used to execute any of the freight return route quantification methods based on spatiotemporal utility calculation described in the first aspect. The system includes: a multi-source data acquisition module, a personalized parameter learning module, and a spatiotemporal utility calculation module. The multi-source data acquisition module is used to acquire driver status data, candidate order data, and dynamic road network data and perform preprocessing. The personalized parameter learning module is used to collect historical driver data or generate preference weights through questionnaires. The spatiotemporal utility calculation module is used to calculate the driver's route affinity to candidate orders.

[0020] Compared with the prior art, the beneficial effects of the present invention are: This invention provides a method and system for quantifying freight return trip route equivalence based on spatiotemporal utility calculation. By constructing a dynamic, multi-dimensional spatiotemporal utility model of driver-order, it achieves accurate quantification of drivers' willingness to travel along routes. The route equivalence S-value calculated by this invention is a high-information-density indicator that integrates real-time road network dynamics, detour economy and time costs, time window constraints, and drivers' personalized preferences. Matching decisions based on this S-value can significantly improve the potential order acceptance rate from the source, reduce the platform's empty mileage rate and communication costs, and ultimately enhance driver retention and platform loyalty. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only preferred embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart of a method for quantifying the return route of freight transportation based on spatiotemporal utility calculation, provided by an embodiment of the present invention. Figure 2 This is a schematic diagram of a freight return route quantification system based on spatiotemporal utility calculation, provided by another embodiment of the present invention. Detailed Implementation

[0023] The principles and features of the present invention are described below with reference to the accompanying drawings. The listed embodiments are only used to explain the present invention and are not intended to limit the scope of the present invention.

[0024] Reference Figure 1 An embodiment of the present invention provides a method for quantifying the return route of freight transportation based on spatiotemporal utility calculation. The method includes the following steps: Step 1: Obtain driver status data, candidate order data, dynamic road network data, and driver preference weight parameters, and preprocess the driver status data, candidate order data, and dynamic road network data; The acquisition of driver status data, candidate order data, and dynamic road network data specifically involves: acquiring driver status data: obtaining the driver's current location, return destination, and expected arrival time; acquiring candidate order data: obtaining the pickup point location, delivery point location, and freight revenue for candidate orders; and acquiring dynamic road network data: through the API interface of a third-party map service provider, obtaining real-time road network information covering the areas where the driver's current location, return destination, pickup point location, and delivery point location are located, including traffic congestion index, estimated travel time for each road segment, and communication costs for highways or bridges, etc. Preprocessing includes outlier filtering, road network data smoothing, missing data imputation, noise removal, and data standardization; The specific steps to obtain the driver preference weight parameters are as follows: For drivers with abundant historical data, all their historical order decision data are collected to construct a dataset. A logistic regression model is then used to learn features from this dataset, and after normalization, preference weights are generated. and ,in Given the driver's relative sensitivity to time costs, The driver's relative sensitivity to economic costs; For newly registered drivers with no or insufficient historical data, initial, default preference weights based on group statistics are generated by filling out a pre-defined questionnaire. Once the driver has generated enough historical data, the system switches to feature learning to generate preference weights.

[0025] For example, the dataset includes the driver's "accept" or "reject" behavior when an order is recommended during each return trip, as well as features such as the net detour time cost, net detour economic cost, and order revenue corresponding to that order at that time; the preference weights represent the driver's trade-off between time cost, economic cost, and order revenue. Step 2: Using the acquired data, plan the baseline return route and detour route, and calculate the baseline return cost and detour cost; specifically: The optimal route is planned based on the driver's current location and return destination from the driver status data, serving as the baseline return route. The baseline return cost is then calculated using the baseline equation path.

[0026] in, The estimated time cost for the driver to travel from the current location to the destination. Estimated economic cost for the driver to travel from the current location to the destination; Based on the return trip cost; The optimal route is planned as a detour route using driver status data and candidate order data, including the driver's current location, pickup point location, delivery point location, and driver's return destination. The detour cost is then calculated based on the detour route.

[0027] in, The estimated time cost for the entire detour route. The estimated economic cost of the entire detour route; Cost of detour.

[0028] For example, the optimal path is the path with the shortest time or lowest overall cost calculated by the map API from dynamic road network data based on real-time traffic conditions.

[0029] Step 3: Calculate the net detour cost using the baseline return cost and the detour cost, specifically as follows: The net detour cost of this candidate order is obtained by comparing the baseline return cost and the detour cost:

[0030] in, These represent the additional time and money the driver needs to spend to complete this candidate order.

[0031] Step 4: Calculate the estimated arrival time using the detour route, and combine this with the driver's expected arrival time to calculate the time penalty factor, specifically: The estimated arrival time is calculated by taking the detour route, and a time penalty factor is calculated by combining this with the driver's expected arrival time.

[0032] in, As a time penalty factor, The preset sensitivity coefficient, The estimated arrival time is calculated using the detour route. This represents the driver's expected arrival time.

[0033] For example, the penalty factor employs a non-linear function to enhance the intensity of the escalation of delays; sensitivity coefficient This is used to control the rate at which the penalty increases with the time of lateness, ensuring that minor delays result in small penalties, while severe delays result in penalties that increase exponentially, thus creating a sufficient negative impact in the final routeability calculation.

[0034] Step 5: Calculate the route affinity using driver preference weight parameters, net detour cost, and time penalty factor. Then, use the route affinity to make decisions regarding candidate orders. Specifically: The net utility of the candidate order for the driver is calculated using driver preference weighting parameters, net detour cost, and time penalty factor:

[0035] in, The net utility of this candidate order for the driver is... For the shipping revenue of this candidate order, The driver preference weights, obtained through analysis, reflect the drivers' relative sensitivity to time and economic costs, respectively. The final path length is obtained by mapping the net utility using the Sigmoid function:

[0036] in, For the sake of convenience and , This is the scaling factor.

[0037] For example, scaling factor Used to adjust the route alignment The steepness of the distribution is close to 1. The value indicates that the candidate order has extremely high spatiotemporal utility for the driver, close to 0. A value of 0 indicates extremely low utility. The calculated routeability... It can provide a high-precision, high-reference feature when ranking, filtering, and recommending vehicle-cargo matching systems.

[0038] Another embodiment of the present invention provides a freight return route quantification system based on spatiotemporal utility calculation. The system is used to execute any of the freight return route quantification methods based on spatiotemporal utility calculation described in one embodiment. The system includes: a multi-source data acquisition module, a personalized parameter learning module, and a spatiotemporal utility calculation module. The multi-source data acquisition module is used to acquire driver status data, candidate order data, and dynamic road network data and perform preprocessing. The personalized parameter learning module is used to collect driver historical data or generate preference weights through questionnaires. The spatiotemporal utility calculation module is used to calculate the driver's route affinity to candidate orders.

[0039] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A quantitative method for freight return route measurement based on spatiotemporal utility calculation, characterized in that, The method includes the following steps: Step 1: Obtain driver status data, candidate order data, dynamic road network data, and driver preference weight parameters, and preprocess the driver status data, candidate order data, and dynamic road network data; Step 2: Plan the baseline return route and detour route using the acquired data, and calculate the baseline return cost and detour cost; Step 3: Calculate the net detour cost using the baseline return cost and the detour cost; Step 4: Calculate the estimated arrival time using the detour route, and combine it with the driver's expected arrival time to calculate the time penalty factor; Step 5: Calculate the route affinity using driver preference weight parameters, net detour cost, and time penalty factor, and use the route affinity to make decisions on candidate orders.

2. The quantitative method for freight return route measurement based on spatiotemporal utility calculation according to claim 1, characterized in that, Step 1 involves obtaining driver status data, candidate order data, and dynamic road network data as follows: Obtain driver status data: obtain the driver's current location, the driver's return destination, and the driver's expected arrival time; Obtain candidate order data: Obtain the pickup point location, delivery point location, and shipping revenue of candidate orders; Obtain dynamic road network data: Through the API interface of third-party map service providers, obtain real-time road network information covering the area where the driver's current location, the driver's return destination, the pickup point location, and the delivery point location are located.

3. A quantitative method for freight return route measurement based on spatiotemporal utility calculation according to claim 2, characterized in that, Step 1 preprocessing includes outlier filtering, road network data smoothing, missing data filling, noise removal, and data standardization.

4. A method for quantifying the return route metric of freight transportation based on spatiotemporal utility calculation as described in claim 3, characterized in that, The specific steps for obtaining the driver preference weight parameters in step 1 are as follows: For drivers with abundant historical data, all their historical order decision data are collected to construct a dataset. After feature learning and normalization of the dataset, the driver's preference weights are generated. For drivers with insufficient historical data, default preference weights are generated by having them fill out a pre-defined questionnaire.

5. A quantitative method for freight return route measurement based on spatiotemporal utility calculation according to claim 4, characterized in that, Step 2 is as follows: The optimal route is planned based on the driver's current location and return destination from the driver status data, serving as the baseline return route. The baseline return cost is then calculated using the baseline equation path. in, The estimated time cost for the driver to travel from the current location to the destination. Estimated economic cost for the driver to travel from the current location to the destination; Based on the return trip cost; The optimal route is planned as a detour route using driver status data and candidate order data, including the driver's current location, pickup point location, delivery point location, and driver's return destination. The detour cost is then calculated based on the detour route. in, The estimated time cost for the entire detour route. The estimated economic cost of the entire detour route; Cost of detour.

6. A quantitative method for freight return route measurement based on spatiotemporal utility calculation according to claim 5, characterized in that, Step 3 specifically involves: The net detour cost of this candidate order is obtained by comparing the baseline return cost and the detour cost: in, These represent the additional time and money the driver needs to spend to complete this candidate order.

7. A quantitative method for freight return route measurement based on spatiotemporal utility calculation according to claim 6, characterized in that, Step 4 is as follows: The estimated arrival time is calculated by taking the detour route, and a time penalty factor is calculated by combining this with the driver's expected arrival time. in, As a time penalty factor, The preset sensitivity coefficient, The estimated arrival time is calculated using the detour route. This represents the driver's expected arrival time.

8. A quantitative method for freight return route measurement based on spatiotemporal utility calculation according to claim 7, characterized in that, Step 5 specifically involves: The net utility of the candidate order for the driver is calculated using driver preference weighting parameters, net detour cost, and time penalty factor: in, The net utility of this candidate order for the driver is... For the shipping revenue of this candidate order, The driver preference weights, obtained through analysis, reflect the drivers' relative sensitivity to time and economic costs, respectively. The final path length is obtained by mapping the net utility using the Sigmoid function: in, For the sake of convenience and , This is the scaling factor.

9. A quantitative system for freight return route measurement based on spatiotemporal utility calculation, characterized in that, The system is used to execute the freight return route quantification method based on spatiotemporal utility calculation as described in any one of claims 1-8. The system includes: a multi-source data acquisition module, a personalized parameter learning module, and a spatiotemporal utility calculation module. The multi-source data acquisition module is used to acquire driver status data, candidate order data, and dynamic road network data and perform preprocessing. The personalized parameter learning module is used to collect driver historical data or generate preference weights through questionnaires. The spatiotemporal utility calculation module is used to calculate the driver's route affinity to candidate orders.

Images