Parcel logistics multimodal transport cooperative scheduling method based on big data

Abstract

The invention discloses a parcel logistics multimodal transport cooperative scheduling method based on big data, which relates to the technical field of big data, and comprises the following steps: when an information monitoring terminal monitors an abnormal condition, triggering dynamic correction of a transport efficiency coefficient, a cost fluctuation coefficient and a service risk coefficient; and when the absolute value of the difference value between the updated evaluation index and the original evaluation index exceeds a preset threshold value, the scheduling decision terminal re-executes the scheme generation process to generate an emergency scheme. Through a built-in evaluation model, according to scheduling modes such as cost priority, timeliness priority and safety priority selected by clients, weights are dynamically allocated, evaluation indexes are output, a transportation mode combination is generated in combination with a genetic algorithm, time connection verification and equipment matching verification are carried out, multi-target collaborative optimization is realized, different client requirements can be flexibly adapted, and the method is suitable for large-scale popularization and application. And the transportation efficiency, the cost and the safety are balanced, and the multimodal transport resource allocation efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the field of big data technology, and particularly to a method for collaborative scheduling of multimodal transport of parcel logistics based on big data. Background Art

[0002] With the rapid development of global economic integration and e-commerce, the logistics industry is facing unprecedented opportunities and challenges. Consumers' demands for logistics services are becoming increasingly diverse, expanding from traditional timeliness and security to multi-dimensional requirements such as cost transparency and green and low-carbon. At the same time, the complexity of the supply chain and the breadth of the logistics network continue to expand. Parcel transportation involves multiple links such as collection, sorting, transportation, warehousing, and distribution, and there is an urgent need to achieve efficient integration of resources through multimodal transport such as railways, roads, aviation, and waterways. However, the collaborative scheduling of multimodal transport involves real-time processing of massive heterogeneous data, dynamic resource allocation, and multi-objective optimization, and the traditional logistics management mode has been difficult to meet the requirements. In this context, the application of technologies such as big data and artificial intelligence has become the key driving force for promoting the intelligent upgrading of the logistics industry. How to use these technologies to achieve efficient collaborative scheduling of multimodal transport has become the focus of the industry.

[0003] Currently, certain progress has been made in the field of logistics in multimodal transport scheduling. For example, some enterprises use GPS positioning and Internet of Things technology to achieve real-time monitoring of transportation tools, or use historical data analysis to optimize route planning. In academic research, various scheduling models have also been proposed, such as static resource allocation methods based on linear programming and time window optimization strategies based on heuristic algorithms. However, the existing technologies still have significant limitations. First, the data collection and analysis capabilities are insufficient. Most systems rely on single-source data (such as the status of transportation tools), lacking comprehensive collection and integration of full-chain data such as parcel characteristics, environmental parameters, and equipment compatibility. Second, the existing scheduling models are mostly static or semi-dynamic and cannot respond in real time to emergencies such as traffic congestion and sudden weather changes, resulting in large fluctuations in transportation efficiency. In addition, the dynamic trade-off mechanism for multiple objectives such as cost, timeliness, and risk in the resource allocation process is imperfect and often relies on fixed weights, making it difficult to meet the personalized needs of customers. For example, although the genetic algorithm proposed in some studies can generate transportation combination plans, it does not consider carbon emission constraints and lacks refined design for verifying equipment matching and time connection.

[0004] However, the existing technologies still face many problems that need to be solved urgently: First, the transportation efficiency and resource utilization rate are low. Due to the isolation of data of various transportation modes in multimodal transportation and the lack of a real-time collaboration mechanism, the loading rate of transportation tools is low, the transfer connection is not smooth, and the phenomenon of resource waste is common. For example, road transportation dominates in short-distance distribution, but the cost advantages of railway or water transportation in long-distance trunk lines have not been fully explored. Second, there is insufficient dynamic balance between cost control and risk management. There is a lack of a quantitative model for the impact of variables such as fuel price fluctuations, parcel damage rates, and environmental factors (such as temperature and humidity) on transportation costs and service risks. Existing methods mostly use empirical values or fixed coefficients, making it difficult to achieve accurate prediction and dynamic adjustment. Third, the ability to respond to anomalies is weak. When encountering traffic congestion, equipment failures, or sudden weather changes, traditional systems often rely on manual intervention, with a lag in response and low efficiency in generating emergency plans, which easily leads to default of customer time windows or loss of goods. Fourth, the requirements of green logistics have not been effectively incorporated into the decision-making system. Although low-carbon transportation has become an industry trend, existing scheduling models rarely take carbon emissions as a hard constraint, resulting in insufficient environmental friendliness of the solutions.

[0005] Therefore, it is necessary to invent a collaborative scheduling method for multimodal parcel logistics based on big data to solve the above problems. Summary of the Invention

[0006] The purpose of the present invention is to provide a collaborative scheduling method for multimodal parcel logistics based on big data to solve the problems raised in the above background technology.

[0007] To achieve the above purpose, the present invention provides the following technical solution: A collaborative scheduling method for multimodal parcel logistics based on big data, including a parcel data collection terminal, an environmental data collection terminal, a loading equipment data collection terminal, a scheduling decision terminal, an information monitoring terminal, a logistics terminal, and a multimodal transportation resource pool. The specific steps are as follows: S1. The parcel data collection terminal, the environmental data collection terminal, and the loading equipment data collection terminal respectively collect the full-chain data of the parcel from receipt, sorting, transportation, warehousing to distribution, obtaining a parcel data set, an environmental data set, and a loading equipment data set; S2. The logistics terminal processes and analyzes the parcel data set, the environmental data set, and the loading equipment data set to obtain a transportation efficiency coefficient, a cost fluctuation coefficient, and a service risk coefficient; S3. The logistics terminal imports the transportation efficiency coefficient, the cost fluctuation coefficient, and the service risk coefficient into a built-in evaluation model, dynamically assigns the weights of the evaluation model according to the scheduling mode selected by the customer, and outputs an evaluation index; S4. The scheduling decision terminal calls the multimodal transport resource pool to obtain the real-time available resources of railways, highways, aviation, and waterways, generates all possible combinations of transportation modes through the genetic algorithm, calculates the evaluation indices of each combination, sorts them in descending order, selects the top five as candidate solutions, then conducts time connection verification and equipment matching verification, and finally selects the solution with the highest evaluation index and carbon emissions lower than the industry average from the feasible candidate solutions; S5. When the information monitoring terminal detects an abnormal situation, it triggers the dynamic correction of the transportation efficiency coefficient, cost fluctuation coefficient, and service risk coefficient. When the absolute value of the difference between the updated evaluation index and the original evaluation index exceeds the preset threshold, the scheduling decision terminal re-executes the solution generation process to generate an emergency plan.

[0008] Preferably, the package data set includes: package weight, package volume, actual package density, package required equipment code, package status, starting point of the customer-specified time window, time window length, lowest transportation cost of historical same-type packages, package breakage rate, complaint rate, historical on-time delivery rate, and historical goods integrity rate; the environment data set includes: weather status, weather level, environmental parameter safety threshold, departure congestion coefficient, destination congestion coefficient, set of high-risk congestion sections, section identifiers in the transportation route, and environmental parameters, where the environmental parameters include: environmental temperature, environmental humidity, and environmental wind force; the loading equipment data set includes: available equipment type code, maximum equipment type code, maximum load-bearing capacity of the logistics node, maximum density allowed by the transportation tool, optimal loading volume of the transportation tool, maximum volume of the logistics node, equipment failure status, and historical average delay time.

[0009] Preferably, the transportation efficiency coefficient is specifically: , where is the transportation mode adaptation coefficient, specifically , where k is the transportation mode, Type(k) is the binary identifier of the transportation mode, and λ k is the basic efficiency parameter of each transportation mode; is the weather impact coefficient, specifically , where t is the weather status, Weather(t) is the weather level, and μ t is the weather impact weight; ε is the equipment compatibility coefficient, specifically , where E is the available equipment type code, E req is the package required equipment code, and E max is the maximum equipment type code; η is the global scaling coefficient, specifically , where e is the natural constant and Z is the evaluation index; , where ω1, ω2, ω3, and ω4 are weight coefficients, ω1, ω2, ω3, and ω4 ∈ [0, 1], and ω1 + ω2 + ω3 + ω4 = 1; f1 is the weight utilization factor, specifically: , where D is the actual density of the package, specifically , W0 is the weight of the package, V0 is the volume of the package, W max is the maximum load capacity of the logistics node, D lim is the maximum density allowed by the transportation vehicle; f2 is the volume utilization factor, specifically: , where V opt is the optimal loading volume of the transportation vehicle, V max is the maximum volume of the logistics node, e is the natural constant; f3 is the time efficiency factor, specifically: , where T min is the shortest time required to complete the transportation task under ideal transportation conditions, △T is the historical average delay time, T w is the starting point of the customer-specified time window, T win is the length of the time window, T p is the actual transportation time predicted based on historical data, T s is the actual start time of transportation; f4 is the congestion impact factor, specifically: , where F avg is the overall congestion coefficient, specifically , where F s is the congestion coefficient at the departure location, F d is the congestion coefficient at the destination; m is the total number of sections in the transportation route, R j is the j-th section identifier in the transportation route, Ω is the set of high-risk congestion sections, I(R j ∈Ω) is the indicator function, specifically , used to determine whether the j-th section belongs to the set of high-risk congestion sections Ω.

[0010] Preferably, the cost fluctuation coefficient is specifically: , where n is the number of optional transportation mode combinations, c i is the unit weight cost of the i-th transportation mode, is the estimated transportation mileage of the i-th transportation mode, p i is the package breakage rate of the i-th transportation mode, For the lowest transportation cost of historical packages of the same type, and θ is the fuel price fluctuation coefficient.

[0011] Preferably, the service risk coefficient is specifically: , where O tr is the historical on-time delivery rate, I cr is the historical package integrity rate, η time is the time sensitivity coefficient, specifically , where T remain is the remaining transportation time, T plan is the planned transportation time; η env is the environmental risk buffer coefficient, specifically , X i is the actual data of the i-th environment, X i，max is the safety threshold of each environmental parameter; Q sec is the safety compliance score, Q cus is the customer complaint rate score, specifically , where the complaint rate = the number of complaints ÷ the number of orders placed, and the number of orders placed is in units of ten thousand orders; m is the number of modes in the transportation mode combination, R risk，k is the specific risk index of the k-th transportation mode, and n is the total number of environmental parameters; δ1, δ2, δ3, and δ4 are weight coefficients, δ1, δ2, δ3, and δ4 ∈ [0, 1], and δ1 + δ2 + δ3 + δ4 = 1.

[0012] Preferably, the evaluation model is specifically: , where k1 is the timeliness weight coefficient, k2 is the cost weight coefficient, k3 is the safety weight coefficient, k1, k2, and k3 ∈ [0, 1], and k1 + k2 + k3 = 1, K is the transportation efficiency coefficient, C is the cost fluctuation coefficient, and S is the service risk coefficient.

[0013] Preferably, the scheduling mode is divided into a cost-priority mode, a timeliness-priority mode, and a safety-priority mode; in the cost-priority mode, the cost weight coefficient is set to the highest; in the timeliness-priority mode, the timeliness weight coefficient is set to the highest; in the safety-priority mode, the safety weight coefficient is set to the highest.

[0014] Preferably, the time connection verification specifically verifies the rationality of the time connection of each transportation mode in the candidate solution, predicts the actual transportation time T p and the actual transportation start time T s , and determines whether T s +T p is within the start point T of the time window specified by the customerw With T w +T win Within the range, combined with the shortest time T under ideal transportation conditions min and the historical average delay time ΔT, verify whether the time efficiency factor f3 meets the preset time efficiency threshold; the device matching verification specifically verifies the matching between the loading device and the package in the candidate solution, and compares the available device type code E with the package required device code E req to ensure that E≥E req and E≤E max At the same time, verify the maximum load-bearing W of the logistics node max whether it is greater than or equal to the package weight W0, and the maximum density D allowed by the transportation tool lim whether it is greater than or equal to the actual density D of the package, and whether the volume utilization factor f2 meets the preset volume utilization threshold.

[0015] Preferably, the dynamic correction is as follows: A1. When there is traffic congestion, correct the overall congestion coefficient F avg and the value of the indicator function I(R j ∈Ω) in the transportation efficiency coefficient K and the service risk coefficient S; A2. When the weather suddenly changes, correct the weather impact coefficient δ and the environmental risk buffer coefficient η env in the transportation efficiency coefficient K and the service risk coefficient S; A3. When there is equipment failure, correct the equipment compatibility coefficient ε and the safety compliance score Q sec in the transportation efficiency coefficient K and the service risk coefficient S; A4. When the package status is abnormal, correct the weight utilization factor f1, the volume utilization factor f2 and the package breakage rate p of the i-th transportation mode i in the transportation efficiency coefficient K, the cost fluctuation coefficient C and the service risk coefficient S.

[0016] The technical effects and advantages of the present invention: Through the package data collection terminal, the environmental data collection terminal and the loading device data collection terminal, the present invention collects the whole-chain data of the package from receipt to distribution, forms a package data set, an environmental data set and a loading device data set, and obtains the transportation efficiency coefficient, the cost fluctuation coefficient and the service risk coefficient through the processing and analysis of the logistics terminal, providing multi-dimensional data support for scheduling decisions, solving the problem of fragmented traditional scheduling data, and improving the scientificity and accuracy of scheduling decisions; Through the built-in evaluation model, the present invention dynamically assigns weights according to scheduling modes such as cost priority, timeliness priority, and safety priority selected by customers, outputs an evaluation index, combines with a genetic algorithm to generate a transportation mode combination, and conducts time connection verification and equipment matching verification, achieving multi-objective collaborative optimization, being able to flexibly adapt to different customer requirements, striking a balance among transportation efficiency, cost, and safety, and improving the resource allocation efficiency of multimodal transport; The present invention monitors abnormal situations in real time through an information monitoring terminal, triggers dynamic correction of the transportation efficiency coefficient, cost fluctuation coefficient, and service risk coefficient. When the change in the evaluation index exceeds a preset threshold, the scheduling decision terminal re-executes the scheme generation process to generate an emergency plan, enhancing the system's dynamic response ability to emergencies such as traffic congestion and sudden weather changes, ensuring the stability and reliability of the transportation process, reducing transportation risks, and improving the quality of logistics services. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a system framework diagram of the present invention.

[0018] Figure 2 It is a flowchart of the method implementation steps of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0019] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0020] The present invention provides a Figure 1 package logistics multimodal transport collaborative scheduling method based on big data as shown, including a package data collection terminal, an environmental data collection terminal, a loading equipment data collection terminal, a scheduling decision terminal, an information monitoring terminal, a logistics terminal, and a multimodal transport resource pool; The package data collection terminal consists of an electronic weighing device, a laser scanning or visual recognition device, an RFID tag reader, a logistics camera, and a logistics platform system; The environmental data collection terminal consists of a meteorological API interface, an environmental sensor, a traffic condition API, and a map service platform or a traffic management system; The loading equipment data collection terminal consists of built-in sensors of the equipment, such as a vibration sensor and a current sensor; The scheduling decision terminal consists of a high-performance computer or server and a data interaction interface. The high-performance computer or server is used to run the genetic algorithm. The data interaction interface obtains the real-time available resource data of railways, highways, aviation, and waterways by connecting to the multimodal transport resource pool. The information monitoring terminal consists of a monitoring software system, an alarm device, and a display screen. The monitoring software system is used to analyze data in real time to identify events such as traffic congestion, sudden weather changes, equipment failures, and abnormal package status. The alarm device is used to alert the operator when an abnormal event is triggered. The display screen is used to display the real-time transportation status and abnormal warning information. The logistics terminal consists of a high-performance computer or server with an evaluation model built in. The high-performance computer or server is used to calculate the transportation efficiency coefficient, cost fluctuation coefficient, and service risk coefficient. The multimodal transport resource pool consists of a distributed database server and a cloud computing platform, which stores and manages the real-time available resource data of railways, highways, aviation, and waterways.

[0021] The specific steps are as Figure 2 shown: S1. The package data acquisition terminal, environmental data acquisition terminal, and loading equipment data acquisition terminal respectively collect the full-chain data of the package from receipt, sorting, transportation, warehousing to distribution, obtaining the package data set, environmental data set, and loading equipment data set. Furthermore, in the above technical solution, the package data set includes: package weight, package volume, actual package density, package required equipment code, package status, start point of the customer-specified time window, time window length, lowest transportation cost of historical same-type packages, package breakage rate, complaint rate, historical on-time delivery rate, and historical goods integrity rate; the environmental data set includes: weather status, weather grade, safety threshold of environmental parameters, congestion coefficient of the departure place, congestion coefficient of the destination, set of high-risk congestion sections, section identification in the transportation route, and environmental parameters, where the environmental parameters include: environmental temperature, environmental humidity, and environmental wind force; the loading equipment data set includes: available equipment type code, maximum value of equipment type code, maximum load-bearing of the logistics node, maximum density allowed by the transportation tool, optimal loading volume of the transportation tool, maximum volume of the logistics node, equipment failure status, and historical average delay time.

[0022] It should be noted that the weight of the package is automatically collected by electronic weighing devices at logistics nodes such as sorting centers and warehouses; the volume of the package is automatically measured by laser scanning or visual recognition devices such as 3D vision measuring instruments to measure the length, width, and height of the package and calculate the volume; the actual density of the package is automatically calculated from the values of the package weight and the package volume; the required device code for the package is manually selected by the staff according to the package characteristics such as refrigeration and fragility during collection and delivery, or automatically identified through intelligent tags such as RFID; the status of the package is identified and collected through methods such as manual spot checks and visual inspections by logistics cameras, such as package deformation identification and collection; the starting point of the customer-specified time window and the length of the time window are collected through the logistics platform; the lowest transportation cost of historical similar packages is calculated based on historical data; the package damage rate is obtained by calculating the ratio of the number of damaged packages of the same type in historical transportation to the total transportation volume; the complaint rate is calculated by collecting the number of complaints through the customer service system and combining it with the number of orders placed in the order system, where the number of orders placed is in units of ten thousand orders; the historical on-time delivery rate is obtained by calculating the ratio of the number of orders delivered on time in historical orders to the total number of orders; the historical goods integrity rate is obtained by calculating the ratio of the number of orders with undamaged goods in historical orders to the total number of orders; the weather status and weather grade are obtained by parsing the weather data along the transportation route in real time through the API of the meteorological department into status and grade; the environmental parameters are collected by sensors deployed at logistics nodes and transportation tools such as trucks and containers, including temperature sensors, humidity sensors, anemometers, rain gauges, air quality sensors, etc.; the safety thresholds of the environmental parameters are preset by industry standards or enterprise internal specifications; the congestion coefficients at the departure and destination are obtained through the real-time traffic condition API of the traffic management department, such as the open platforms of Gaode Map and Baidu Map, to obtain the road congestion index; the set of high-risk congested road sections is automatically marked based on historical traffic data; the section identifiers in the transportation route are obtained through the map service platform or the traffic management system; the available device type codes are preset, and they use 8-bit binary codes, with the first four bits identifying the major device categories and the last four bits identifying the detailed models; the maximum value of the device type code is preset, and its maximum value is 256 types; the maximum load-bearing capacity of the logistics node is obtained from the design drawings or equipment specifications of the logistics node such as the warehouse and freight station; the maximum allowable density of the transportation tool is calculated based on the maximum load and volume of the transportation tool; the optimal loading volume of the transportation tool is obtained through statistical analysis of historical loading data; the maximum volume of the logistics node is preset; the device failure status is monitored in real time through built-in sensors in the device such as vibration sensors and current sensors, and an alarm is automatically triggered in case of a failure; the historical average delay time is calculated based on historical data.

[0023] S2. The logistics terminal processes and analyzes the package dataset, environment dataset, and loading equipment dataset to obtain the transportation efficiency coefficient, cost fluctuation coefficient, and service risk coefficient; Further, in the above technical solution, the transportation efficiency coefficient is specifically: , Among them, is the transportation mode adaptation coefficient, specifically , where k is the transportation mode, Type(k) is the binary identifier of the transportation mode, and λ k is the basic efficiency parameter of each transportation mode; is the weather impact coefficient, specifically , where t is the weather condition, Weather(t) is the weather grade, and μ t is the weather impact weight; ε is the equipment compatibility coefficient, specifically , where E is the available equipment type code, E req is the package demand equipment code, and E max is the maximum value of the equipment type code; η is the global scaling coefficient, specifically , where e is the natural constant and Z is the evaluation index; , where ω1, ω2, ω3, and ω4 are weight coefficients, ω1, ω2, ω3, and ω4 ∈ [0, 1], and ω1 + ω2 + ω + ω4 = 1; f1 is the weight utilization factor, specifically: , Among them, D is the actual density of the package, specifically , W0 is the package weight, V0 is the package volume, and W max is the maximum load-bearing of the logistics node, and D lim is the maximum density allowed by the transportation tool; f2 is the volume utilization factor, specifically: , Among them, V opt is the optimal loading volume of the transportation tool, and V max is the maximum volume of the logistics node, and e is the natural constant; f3 is the time efficiency factor, specifically: , Among them, T min is the shortest time required to complete the transportation task under ideal transportation conditions, △T is the historical average delay time, and T w is the starting point of the customer-specified time window, and T win is the time window length, and T p is the actual transportation time predicted based on historical data, and T sis the actual start time of transportation; f4 is the congestion impact factor, specifically: , where F avg is the overall congestion coefficient, specifically , where F s is the origin congestion coefficient, F d is the destination congestion coefficient; m is the total number of road segments in the transportation route, R j is the j-th road segment identifier in the transportation route, Ω is the set of high-risk congested road segments, I(R j ∈Ω) is an indicator function, specifically , used to determine whether the j-th road segment belongs to the set of high-risk congested road segments Ω; It should be noted that the values of ω1, ω2, ω3, and ω4 in the transportation efficiency coefficient are dynamically updated according to the scheduling mode. If it is the time-efficiency priority mode, then ω1 = 0.4, ω2 = 0.3, ω = 0.2, ω4 = 0.1; if it is the cost-priority mode, then ω1 = 0.35, ω2 = 0.35, ω = 0.2, ω4 = 0.1; if it is the safety-priority mode, then ω1 = 0.3, ω2 = 0.3, ω = 0.2, ω4 = 0.2; The value of k in γ being 1 represents water transportation, 2 represents road transportation, 3 represents railway transportation, 4 represents air transportation, λ1 = 0.7, λ2 = 0.7, λ3 = 0.8, λ4 = 0.9, and Type(k) is specifically that when only road transportation is used, then Type(1) = 0, Type(2) = 1, Type(3) = 0, Type(4) = 0; The value of t in δ being 1 represents sunny, 2 represents cloudy, 3 represents light rain or light snow, 4 represents moderate rain or moderate snow, 5 represents heavy rain or heavy snow or typhoon. Substituting the k value into Weather(k), Weather(1) = 1, Weather(2) = 2, Weather(3) = 3, Weather(4) = 4, Weather(5) = 5, where the weather grade Weather(k) can be adjusted according to the climate characteristics of the business area, and the weather impact weight μ t can be set as μ1 ∈ [0.1, 0.2], μ2 ∈ [0.2, 0.3], μ3 ∈ [0.4, 0.5], μ4 ∈ [0.6, 0.7], μ5 ∈ [0.8, 1.0], where the weather impact weight μ t is set according to the weather grade and historical transportation data; The initial value of η for Z is set to 0.5. The constant "-5" determines the sensitivity of η to changes in the evaluation index Z. When Z is slightly higher than 0.5, η approaches 0, significantly amplifying the weight of the transport efficiency coefficient. When Z is slightly lower than 0.5, η rapidly approaches 1, greatly weakening the weight of the transport efficiency coefficient.

[0024] The cost fluctuation coefficient is specifically: , where n is the number of optional transport mode combinations, c i is the unit weight cost of the i-th transport mode, is the estimated transport mileage of the i-th transport mode, p i is the package breakage rate of the i-th transport mode, is the lowest transport cost of historical packages of the same type, and θ is the fuel price fluctuation coefficient.

[0025] The service risk coefficient is specifically: , where O tr is the historical on-time delivery rate, I cr is the historical package integrity rate, η time is the time sensitivity coefficient, specifically , where T remain is the remaining transport time, T plan is the planned transport time; η env is the environmental risk buffer coefficient, specifically , X i is the actual data of the i-th environment, X i，max is the safety threshold of each environmental parameter; Q sec is the safety compliance score, Q cus is the customer complaint rate score, specifically , where the complaint rate = number of complaints ÷ number of orders placed, and the number of orders placed is in units of ten thousand orders; m is the number of modes in the transport mode combination, R risk，k is the specific risk index of the k-th transport mode, and n is the total number of environmental parameters; δ1, δ2, δ3, and δ4 are weight coefficients, δ1, δ2, δ3, and δ4 ∈ [0, 1], and δ1 + δ2 + δ3 + δ4 = 1; It should be noted that the safety threshold X i，max of each environmental parameter is preset by industry standards or enterprise specifications, such as the temperature upper limit for cold chain transportation; in the service risk coefficient , when k = 1, it is water transportation, R risk，1 = port operation delay duration ÷ standard operation duration; when k = 2, it is road transportation, R risk，2=Actual congestion mileage÷Total mileage; When k = 3, it is for railway, R risk，3 =1 - Current train punctuality rate÷100, where the current train punctuality rate = Actual on-time train trips÷Total train trips; When k = 4, it is for aviation, R risk，4 =Airport delay index÷5, where the delay index ranges from 0 to 5, and 5 represents severe delay.

[0026] S3. The logistics terminal imports the transportation efficiency coefficient, cost fluctuation coefficient, and service risk coefficient into the built-in evaluation model, dynamically assigns the weights of the evaluation model according to the scheduling mode selected by the customer, and outputs the evaluation index; Further, in the above technical solution, the evaluation model is specifically: , where k1 is the timeliness weight coefficient, k2 is the cost weight coefficient, k3 is the safety weight coefficient, k1, k2, and k3 ∈ [0, 1], and k1 + k2 + k3 = 1, K is the transportation efficiency coefficient, C is the cost fluctuation coefficient, and S is the service risk coefficient.

[0027] The scheduling mode is divided into cost - priority mode, timeliness - priority mode, and safety - priority mode; In the cost - priority mode, the cost weight coefficient is set to the highest; In the timeliness - priority mode, the timeliness weight coefficient is set to the highest; In the safety - priority mode, the safety weight coefficient is set to the highest; It should be noted that the weights are automatically assigned according to the scheduling mode selected by the customer, and the specific rules are as follows: Timeliness - priority mode: k1 = 0.6, k2 = 0.2, k3 = 0.2; Cost - priority mode: k1 = 0.5, k2 = 0.3, k3 = 0.2; Safety - priority mode: k1 = 0.5, k2 = 0.3, k3 = 0.2.

[0028] S4. The scheduling decision terminal calls the multimodal transport resource pool to obtain the real - time available resources of railway, highway, aviation, and waterway, generates all possible combinations of transportation modes through the genetic algorithm, calculates the evaluation index of each combination, arranges them in descending order, selects the top five as candidate solutions, then conducts time connection verification and equipment matching verification, and finally selects the solution with the highest evaluation index and carbon emissions lower than the industry average from the feasible candidate solutions; It should be noted that according to the genetic algorithm rules, the combination of "railway + highway + aviation + waterway" is represented in binary coding as 1111, where the 1st bit is for railway, the 2nd bit is for highway, the 3rd bit is for aviation, the 4th bit is for waterway, 1 indicates that the transportation mode is selected, and 0 indicates that it is not selected; The carbon emissions are specifically , where i represents the i-th transportation mode. For example, when i = 1, it is water transportation; when i = 2, it is road transportation, etc. The carbon emission coefficient can be obtained by referring to industry standards, such as the national greenhouse gas accounting guidelines and the emission factor database of industry associations; Further, in the above technical solution, the time connection verification specifically verifies the rationality of the time connection of each transportation mode in the candidate solution, and predicts the actual transportation time T based on historical data p and the actual transportation start time T s , and determines whether T s +T p is within the starting point T of the time window specified by the customer w and T w +T win range. At the same time, in combination with the shortest time T under ideal transportation conditions min and the historical average delay time ΔT, verify whether the time efficiency factor f3 meets the preset time efficiency threshold; the equipment matching verification specifically verifies the matching between the loading equipment and the package in the candidate solution, and compares the available equipment type code E with the package required equipment code E req , ensuring that E≥E req and E≤E max . At the same time, verify whether the maximum load-bearing capacity W of the logistics node max is greater than or equal to the package weight W0, whether the maximum density D allowed by the transportation tool lim is greater than or equal to the actual density D of the package, and whether the volume utilization factor f2 meets the preset volume utilization threshold.

[0029] It should be noted that the preset threshold of f2 is shown in Table 1: Table 1 Threshold type Threshold value Coverage scenario Efficient ≥0.6 High-volume efficiency scenarios for air or container transportation Conservative ≥0.5 High-value goods or volume-sensitive transportation Conventional ≥0.45 Standard volume requirements for general goods Lowest viable ≥0.1 Emergency dispatch or oversized cargo transportation The preset threshold of f3 is shown in Table 2: Table 2 Threshold type Threshold value Coverage scenario Extreme punctuality ≥0.97 Extreme time-sensitive requirements for medical emergencies or fresh food cold chain Conservative ≥0.9 Time-limited express delivery Conventional ≥0.8 Time limit bottom line for general e-commerce packages Lowest tolerance ≥0.2 Bulk materials or non-emergency transportation In a specific embodiment, assuming that the customer selects the time efficiency priority mode, the scheduling decision terminal generates the following 5 candidate solutions through the genetic algorithm, as shown in Table 3: Table 3 Candidate solution Transportation mode Evaluation index Z Time connection verification Equipment matching verification Carbon emissions (t) 1 Highway + Air 0.85 Pass Pass 1.2 2 Railway + Air 0.82 Pass Pass 1.0 3 Highway + Railway 0.78 Pass Equipment code mismatch 1.3 4 Air 0.75 Time window exceeded - 1.1 5 Highway + Waterway 0.70 Pass Pass 1.6 Judge the carbon emissions according to industry standards and select Solution 1; S5. When the information monitoring terminal detects an abnormal situation, it triggers the dynamic correction of the transportation efficiency coefficient, cost fluctuation coefficient and service risk coefficient. When the absolute value of the difference between the updated evaluation index and the original evaluation index exceeds the preset threshold, the scheduling decision terminal re-executes the solution generation process to generate an emergency solution.

[0030] Further, in the above technical solution, the dynamic correction is as follows: A1. When there is traffic congestion, correct the values of the whole - journey congestion coefficient F in the transportation efficiency coefficient K and the service risk coefficient S avg and the indicator function I(R j ∈Ω); A2. When the weather suddenly changes, correct the values of the weather - influence coefficient δ and the environmental - risk buffer coefficient η in the transportation efficiency coefficient K and the service risk coefficient S env ; A3. When there is equipment failure, correct the values of the equipment - compatibility coefficient ε and the safety - compliance score Q in the transportation efficiency coefficient K and the service risk coefficient S sec ; A4. When the package status is abnormal, correct the values of the weight - utilization factor f1, the volume - utilization factor f2, and the package - breakage rate p of the i - th transportation mode in the transportation efficiency coefficient K, the cost - fluctuation coefficient C, and the service risk coefficient S i .

[0031] In a specific embodiment, assume that the information monitoring terminal monitors through the meteorological API that the weather along the transportation route suddenly changes from cloudy to heavy rain, the environmental temperature rises from 25 °C to 30 °C, approaching the safety threshold of 35 °C, and the humidity rises from 60% to 90%, exceeding the safety threshold of 80%, triggering the correction of the efficiency coefficient K. Specifically, the weather - influence coefficient δ rises from 0.5 to 4.5. If the package - required equipment is the refrigerated box E req = 0100, the currently available equipment E = 0100, but the heavy rain may cause the risk of temporary equipment failure, and the equipment - compatibility coefficient ε drops from 1 to 0.8. At the same time, it triggers the correction of the service risk coefficient S. Specifically, the environmental - risk buffer coefficient η env rises from 0.3 to 0.7. Because the heavy rain may violate the transportation safety regulations, the safety - compliance score Q sec drops from 90 points to 75 points. After the dynamic correction, the evaluation index Z′ = 0.48. Based on the evaluation index Z = 0.6 before correction, the calculated difference is 0.12, which is greater than the preset threshold of 0.1, triggering the re - execution of the scheme - generation process. The scheduling decision - making terminal then generates a combination of "air + road" through the genetic algorithm and passes the time - connection verification and equipment - matching verification. This scheme is feasible. Although the cost increases, it avoids the delay and cargo - damage risks caused by the heavy rain.

[0032] Finally, it should be noted that the above are only preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A collaborative scheduling method for multimodal transport of parcel logistics based on big data, characterized in that, It includes a parcel data collection terminal, an environmental data collection terminal, a loading equipment data collection terminal, a scheduling decision-making terminal, an information monitoring terminal, a logistics terminal, and a multimodal transport resource pool. The specific steps are as follows: S1. The parcel data collection terminal, the environmental data collection terminal, and the loading equipment data collection terminal respectively collect the full-chain data of the parcel from receipt, sorting, transportation, warehousing to distribution, obtaining a parcel data set, an environmental data set, and a loading equipment data set; S2. The logistics terminal processes and analyzes the parcel data set, the environmental data set, and the loading equipment data set to obtain a transportation efficiency coefficient, a cost fluctuation coefficient, and a service risk coefficient; S3. The logistics terminal imports the transportation efficiency coefficient, the cost fluctuation coefficient, and the service risk coefficient into the built-in evaluation model, dynamically assigns the weights of the evaluation model according to the scheduling mode selected by the customer, and outputs an evaluation index; S4. The scheduling decision-making terminal calls the multimodal transport resource pool to obtain the real-time available resources of railways, highways, aviation, and waterways, generates all possible combinations of transportation modes through a genetic algorithm, calculates the evaluation indexes of each combination, arranges them in descending order, selects the top five as candidate solutions, then conducts time connection verification and equipment matching verification, and finally selects the solution with the highest evaluation index and a carbon emission lower than the industry average from the feasible candidate solutions; S5. When the information monitoring terminal detects an abnormal situation, it triggers the dynamic correction of the transportation efficiency coefficient, the cost fluctuation coefficient, and the service risk coefficient. When the absolute value of the difference between the updated evaluation index and the original evaluation index exceeds the preset threshold, the scheduling decision-making terminal re-executes the solution generation process to generate an emergency plan.

2. The multi-modal intermodal collaborative scheduling method for parcel logistics based on big data according to claim 1, wherein The parcel data set includes: parcel weight, parcel volume, actual density of the parcel, parcel required equipment code, parcel status, starting point of the customer-specified time window, time window length, lowest transportation cost of historical same-type parcels, parcel damage rate, complaint rate, historical on-time delivery rate, and historical goods integrity rate; the environmental data set includes: weather status, weather grade, environmental parameter safety threshold, departure congestion coefficient, destination congestion coefficient, set of high-risk congestion sections, section identification in the transportation route, and environmental parameters, where the environmental parameters include: environmental temperature, environmental humidity, and environmental wind force; the loading equipment data set includes: available equipment type code, maximum value of the equipment type code, maximum load-bearing of the logistics node, maximum density allowed by the transportation tool, optimal loading volume of the transportation tool, maximum volume of the logistics node, equipment failure status, and historical average delay time.

3. A collaborative scheduling method for multimodal transport of parcels based on big data according to claim 1, characterized in that, The transportation efficiency coefficient is specifically: ， Among them, is the transportation mode adaptation coefficient, specifically , where k is the transportation mode, Type(k) is the binary identifier of the transportation mode, and λ k is the basic efficiency parameter of each transportation mode; is the weather impact coefficient, specifically , where t is the weather condition, Weather(t) is the weather grade, and μ t is the weather impact weight; ε is the equipment compatibility coefficient, specifically , where E is the available equipment type code, E req is the package demand equipment code, and E max is the maximum value of the equipment type code; η is the global scaling coefficient, specifically , where e is the natural constant and Z is the evaluation index; , where ω1, ω2, ω3, and ω4 are weight coefficients, ω1, ω2, ω3, and ω4 ∈ [0, 1], and ω1 + ω2 + ω3 + ω4 = 1; f1 is the weight utilization factor, specifically: ， where D is the actual density of the package, specifically , W0 is the weight of the package, V0 is the volume of the package, W max is the maximum load-bearing capacity of the logistics node, D lim is the maximum density allowed by the transportation vehicle; f2 is the volume utilization factor, specifically: ， Among them, V opt is the optimal loading volume of the transportation vehicle, V max is the maximum volume of the logistics node, e is the natural constant; f3 is the time efficiency factor, specifically: ， Among them, T min is the shortest time required to complete the transportation task under ideal transportation conditions, △T is the historical average delay time, T w is the starting point of the time window specified by the customer, T win is the length of the time window, T p is the actual transportation time predicted based on historical data, T s is the actual start time of transportation; f4 is the congestion impact factor, specifically: ， Among them, F avg is the overall congestion coefficient, specifically , where F s is the departure point congestion coefficient, and F d is the destination congestion coefficient; m is the total number of road segments in the transportation route, and R j is the identification of the j-th road segment in the transportation route, Ω is the set of high-risk congested road segments, and I(R j ∈Ω) is an indicator function, specifically , which is used to determine whether the j-th road segment belongs to the set Ω of high-risk congested road segments.

4. A collaborative scheduling method for multimodal transport of parcels based on big data according to claim 1, characterized in that The cost fluctuation coefficient is specifically: ， Among them, n is the number of optional transportation mode combinations, and c i is the unit weight cost of the i-th transportation mode, is the estimated transportation mileage of the i-th transportation mode, and p i is the package breakage rate of the i-th transportation mode, is the lowest transportation cost of historical packages of the same type, and θ is the fuel price fluctuation coefficient.

5. A collaborative scheduling method for multimodal transport of parcels based on big data according to claim 1, characterized in that, The service risk coefficient is specifically: ， Among them, O tr is the historical on-time delivery rate, I cr is the historical package integrity rate, η time is the time sensitivity coefficient, specifically , where T remain is the remaining transportation time, T plan is the planned transportation time; η env is the environmental risk buffer coefficient, specifically , X i is the actual data of the i-th environment, X i，max is the safety threshold of each environmental parameter; Q sec is the safety compliance score, Q cus is the customer complaint rate score, specifically , where the complaint rate = the number of complaints ÷ the number of orders placed, and the number of orders placed is in units of ten thousand orders; m is the number of modes in the transportation mode combination, R risk，k is the specific risk index of the k-th transportation mode, and n is the total number of environmental parameters; δ1, δ2, δ3, and δ4 are weight coefficients, δ1, δ2, δ3, and δ4 ∈ [0, 1], and δ1 + δ2 + δ3 + δ4 = 1.

6. The method for collaborative scheduling of multimodal transport of parcels based on big data according to claim 1, wherein The evaluation model is specifically: ， Among them, k1 is the timeliness weight coefficient, k2 is the cost weight coefficient, k3 is the safety weight coefficient, k1, k2, and k3 ∈ [0, 1], and k1 + k2 + k3 = 1. K is the transportation efficiency coefficient, C is the cost fluctuation coefficient, and S is the service risk coefficient.

7. A collaborative scheduling method for multimodal transport of parcels based on big data according to claim 1, characterized in that, The scheduling modes are divided into a cost - priority mode, a timeliness - priority mode, and a safety - priority mode; in the cost - priority mode, the cost weight coefficient is set to the highest; in the timeliness - priority mode, the timeliness weight coefficient is set to the highest; in the safety - priority mode, the safety weight coefficient is set to the highest.

8. A collaborative scheduling method for multimodal transport of parcels based on big data according to claim 3, characterized in that The time connection verification specifically verifies the rationality of the time connection of each transportation mode in the candidate plan, and predicts the actual transportation time T based on historical data p and the actual transportation start time T s , and judge whether T s +T p is within the starting point T of the time window specified by the customer w and T w +T win range. At the same time, combined with the shortest time T under ideal transportation conditions min and the historical average delay time ΔT, verify whether the time efficiency factor f3 meets the preset time efficiency threshold; the equipment matching verification specifically verifies the matching of the loading equipment and the package in the candidate plan, and compares the available equipment type code E with the package required equipment code E req , ensure that E≥E req and E≤E max . At the same time, verify whether the maximum load-bearing W of the logistics node max is greater than or equal to the package weight W0, whether the maximum density D allowed by the transportation tool lim is greater than or equal to the actual density D of the package, and whether the volume utilization factor f2 meets the preset volume utilization threshold.

9. A method for collaborative scheduling of multimodal transport of parcel logistics based on big data according to claim 6, characterized in that, The dynamic correction is as follows: A1. When there is traffic congestion, correct the values of the overall congestion coefficient F in the transportation efficiency coefficient K and the service risk coefficient S avg and the indicator function I(R j ∈Ω); A2. When the weather changes suddenly, correct the values of the weather impact coefficient δ and the environmental risk buffer coefficient η in the transportation efficiency coefficient K and the service risk coefficient S env ; A3. When there is a device failure, correct the values of the equipment compatibility coefficient ε and the safety compliance score Q in the transportation efficiency coefficient K and the service risk coefficient S sec ; A4. When the package status is abnormal, correct the values of the weight utilization factor f1, volume utilization factor f2, and the package breakage rate pi of the i-th transportation mode in the transportation efficiency coefficient K, cost fluctuation coefficient C, and service risk coefficient S. i Make corrections to the values.

Images