Flexible logistics distribution method and system based on intelligent distribution model

Abstract

The invention discloses a flexible logistics distribution method based on an intelligent distribution model. The method comprises the following steps: S1, establishing an intelligent distribution model; s2, collecting and processing commodity distribution process data, wherein the processed commodity distribution process data is used for providing a data basis for the intelligent distribution model; s3, inputting the currently processed commodity distribution process data into the intelligent distribution model so as to obtain an optimal logistics distribution scheme with the lowest total transportation cost and service time; and performing intelligent logistics distribution based on the optimal logistics distribution scheme. The invention also discloses a corresponding flexible logistics distribution system, an electronic device and a computer readable storage medium, and through specific setting of each step and module, the cigarette distribution efficiency is improved, the distribution link cost is reduced, and the distribution service quality is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of large models and intelligent logistics distribution, and in particular to a flexible logistics distribution method and system based on an intelligent distribution model. Background Art

[0002] Cigarette delivery is an image window for external services, and meeting the personalized needs of customers is our current main task. As the demand for cigarettes in the market becomes more diversified, immediate, and precise, retail customers are increasingly demanding the quality of cigarette delivery services. The high cost, low efficiency, poor benchmarking, and untimely response of cigarette delivery are prominent problems, and the optimization and integration of delivery elements is imperative.

[0003] First, the current tobacco distribution route optimization algorithm takes the direct distribution area as the main research object. Based on GPS and GIS technology, it breaks through the administrative boundaries and proposes a set of solutions with strong adaptability, fast computing speed and high real-time performance. According to the actual business needs, the functional modules are divided, and various methods such as center clustering, line connection algorithm, and tail car allocation are combined. Relying on the data of the integrated platform of Internet of Things management and control, including order date, retail information, order quantity and other information, according to the city's geographical conditions, the maximum loading capacity of the delivery vehicle, the familiarity of the delivery personnel and other restrictions, the algorithm model is formed through computer mathematical modeling. According to the principle of saving mileage and maximizing efficiency, the data of the original fixed route is selectively used as the reference of the final route, and combined with the delivery vehicle model and delivery operation time, dynamic optimization is achieved, and the delivery route of the order of the day is optimized in the shortest possible time, providing each delivery personnel with a relatively reasonable and scientific delivery route. The algorithm model of the existing technology is shown in formula (1). Taking the delivery operation time as a single-dimensional indicator, a relationship model between the delivery operation time and factors such as one-way mileage and time, the number of delivery households and the arrival handover is established, that is:

[0004] H 单车综合作业时间 =h z +2h c +∑h u ; (1)

[0005] h z : Loading time = unit loading time × total loading volume;

[0006] h c : One-way time = one-way distance × driving speed;

[0007] h u : Delivery time per household = average order quantity per household × unit delivery time + average distance between households × driving speed.

[0008] However, this algorithm model has certain original defects due to insufficient data foundation:

[0009] (1) Unit delivery time, customer distance, and driving speed can only be estimated based on average values;

[0010] (2) The order of customer delivery depends entirely on the experience of the delivery person, and customers not included in the pre-ordered routes cannot be compatible with the algorithm model;

[0011] (3) The peaks and valleys of sales volume lead to an increase in the fluctuation of the capacity constraints, making it difficult to achieve optimal route planning, resulting in circuitous distribution routes and reduced efficiency.

[0012] Second, in the distribution model currently used, the initial model algorithm includes three steps: customer area allocation, area identification, route generation, and inter-area adjustment. Figure 1 As shown in the figure, the overall design process of the distribution model at this stage, in which route generation and inter-regional adjustment are divided into two independent paths, one is calculated using an algorithm based on center clustering, and the other is calculated using route concatenation.

[0013] First, customer area allocation is performed. After obtaining marketing orders, the orders are split into various areas for calculation according to the areas demarcated in the system; then, regional identification is performed. Route optimization uses different algorithms to deal with different customer distribution situations; finally, route generation and inter-regional adjustments are performed. In each block, ordering customers are assigned to each original route, and distance analysis is performed to form several continuous and close customer groups. According to the results of regional analysis, different algorithms (cluster analysis or line concatenation) are used to merge and split the generated customer groups, and customer groups close to each other are merged into one route according to the capacity of the vehicle and the maximum delivery time. At the same time, it is allowed to split those customer groups that are "too much to add as a whole, and too little if not added" to maximize the loading rate, while considering road optimization and tail vehicle allocation.

[0014] The intelligent model uses clustering algorithms to optimize delivery routes based on order information, and uses Beidou navigation technology to refine the coordinate locations of retailers, laying the foundation for real-time optimization and real-time delivery. The key lies in the continuous optimization of algorithms, further optimizing various algorithm models involved in route planning, using GIS technology, artificial intelligence technology, and three algorithm models such as large route splitting algorithms, fixed route reconstruction algorithms, and clustering algorithms. Based on daily order data, retailer locations, and traffic road data, combined with the number of delivery vehicles and vehicle loading rates, the optimal delivery route is globally planned, and vehicles are dispatched, breaking through the internal and small-scale planning restrictions of the route, realizing the business adaptability of the route optimization system, and supporting custom selection of multiple route modes including fixed routes, flexible routes, and mixed routes. Supports comparison of multiple delivery plans and free selection of plans.

[0015] Dynamic optimization of the delivery route can be achieved. The most reasonable delivery route can be calculated based on the four parameters of working time, number of delivery households, loading volume and delivery mileage. The following three algorithms can be used:

[0016] ① Large route splitting algorithm: A large route is set in an area, and routes are generated according to the delivery order of all retailers on the route specified by the dispatcher; n routes are cut out under the following four conditions: a. the maximum load of the vehicle is not exceeded; b. the number of households does not exceed the maximum; c. the labor intensity is basically balanced; d. the delivery time cannot exceed the maximum time.

[0017] The load capacity of the tail truck determines whether the delivery route needs to be re-optimized: if it is below the threshold, the load capacity of the tail truck is too small, and re-optimization is performed to distribute the load of the tail truck to the delivery trucks in front; if it is above the threshold, the tail truck is retained and re-optimization is performed to balance part of the load in the front trucks to the tail truck to form the final route.

[0018] Route planning is not based on a single factor to divide the route. It is necessary to achieve a balance between load capacity, mileage, number of households, and labor intensity. The route planning result is not generated completely in the order of customers. Considering the balance of load capacity, number of households, working hours, and mileage, it may happen that a household in the middle is optimized to the next route.

[0019] ② Fixed route reconstruction algorithm: break up and reconstruct the fixed route into a large route, use taboo search (randomly swap the delivery order of two retailers) or even exhaustive method, try various combinations of retailer delivery orders (n to the power of n-1). Consider influencing factors such as loading rate, number of households, maximum working time, etc., and split the customer sequence under this order to evaluate the actual route. The evaluation parameter is the total working time of all routes, and the order with the shortest total working time is used as the large route order. Balance the loading rate, number of households, maximum working time, and mileage, and cut the large route. The fixed route reconstruction algorithm is a system-generated order of retailers' delivery, and the orders are more matched with the fixed routes in the area.

[0020] ③ Clustering algorithm: Estimate the number of vehicles k required based on the order volume and available vehicles in the area; select k retailers in the area as representative objects; divide the remaining retailers into k groups based on their distance from the representative objects; use non-representative retailers to continuously replace representative retailers.

[0021] a. For each non-representative retailer, recalculate the distance with each representative retailer to form a new k group;

[0022] b. Use the cost function to estimate the clustering quality of the new k groups (if S < 0, then replace and a new group of k center points is formed);

[0023] c. S1 = loading rate - 1; S2 = expected delivery time - maximum delivery time; S3 = number of retailers loaded - maximum number of retailers

[0024] d. Continue replacing until S=0, forming k original circuits.

[0025] Continuously learn algorithm strategies and parameter configurations, optimize routes, and reduce algorithm adaptation and tuning time. Support users to manually set algorithm variables, change parameters such as the number of vehicles, loading capacity, working hours, etc., and compare multiple delivery plans vertically for managers to analyze and select.

[0026] The above model algorithms are all based on the current distribution process data analysis settings, which are basically applicable to the tobacco industry's delivery and order-visit model, and have strict restrictions on customer ordering time. Therefore, the intelligence and digitization of the model need to be further developed and improved to adapt to the market-oriented reform of the tobacco industry and the trend from high-speed development to high-quality development.

[0027] Third, at present, information technology is not fully utilized in the delivery process, and there is a lack of data collection methods in the cigarette delivery process, including customer handover coordinates, handover time, handover scene, handover interval, handover sequence and other important information. The lack of key data will have a significant impact on delivery route planning, making it difficult for the intelligent delivery model algorithm to obtain the optimal solution, resulting in circuitous delivery routes, reduced efficiency and increased costs. Summary of the invention

[0028] The purpose of the present invention is to design a logistics distribution method and system based on an intelligent distribution model in response to the defects of the prior art. By standardizing the supervision of the entire logistics distribution process, relying on the electronic receipt process of goods to collect real-time information data such as the operation status of the finished product distribution link, vehicle in-transit monitoring data, and real-time status of commodity distribution, an intelligent distribution model is established based on sensitive factors such as distribution quantity, distribution distance, distribution tools, distribution links, and operation time. The model is verified through actual operation data, and finally an efficient real-time distribution model for finished products of commercial enterprises is established, which improves the efficiency of commodity distribution, reduces the cost of distribution links, and improves the quality of distribution services.

[0029] The first aspect of the present invention is to provide a flexible logistics distribution method based on an intelligent distribution model, comprising:

[0030] S1, establish intelligent distribution model;

[0031] S2, collecting and processing commodity distribution process data, the processed commodity distribution process data is used to provide a data basis for the intelligent distribution model;

[0032] S3, inputting the currently processed commodity distribution process data into the intelligent distribution model, so as to obtain the optimal logistics distribution plan with the lowest total transportation cost and service time; and performing intelligent logistics distribution based on the optimal logistics distribution plan.

[0033] Preferably, the S1 comprises:

[0034] S11, determining the influencing factors of the distribution process, and establishing a distribution operation time sub-model based on the influencing factors of the distribution process; wherein the influencing factors of the distribution process include fixed influencing factors of the distribution process and boundary condition influencing factors of the distribution process;

[0035] S12, establish route planning sub-model based on GIS computer distribution network planning algorithm;

[0036] S13, integrating the distribution operation time sub-model and the route planning sub-model to form an intelligent distribution model; the intelligent distribution model is based on efficient real-time collection, classification analysis, autonomous learning and iterative evolution of distribution process data, and the final output distribution planning solution aims at the lowest total transportation cost and service time.

[0037] Preferably, the S11 includes:

[0038] (1) The fixed factors affecting the delivery process are determined as loading time, vehicle travel time, delivery time per household, average number of households per vehicle per day, payment time, and delivery ticket submission time;

[0039] (2) Calculating the basic data of the distribution area based on the historical data of a specific area and analyzing and estimating the basic data of the distribution area, and determining the boundary condition influencing factors of the distribution process based on the basic data of the distribution area; the boundary condition influencing factors of the distribution process include the upper limit of the distribution operation time, the upper limit of the number of households that can be delivered by a single delivery vehicle, the upper limit of the delivery time for each household, the average loading preparation time, the average cage truck loading time, the average number of cage trucks for single delivery per day, the average unit cage truck pickup time, the average delivery ticket submission time, the average household order quantity, the average number of households per vehicle per day, the average single package handover time, and the driving speed;

[0040] (3) A distribution operation time sub-model is established based on the fixed influencing factors of the distribution process and the boundary condition influencing factors of the distribution process; the distribution operation time sub-model is decomposed using the distribution operation time as a single-dimensional indicator and is expressed by formula (2):

[0041] H p =h z +2h x +h m ×b h +h j +h g(2);

[0042] Where: H p : Delivery operation time;

[0043] h z : Loading time = unit loading time × number of cage trucks + loading preparation time + picking time in temporary storage area;

[0044] Pick-up time in temporary storage area = pick-up time per cage truck × number of cage trucks;

[0045] h x : Vehicle driving time = one-way distance ÷ driving speed;

[0046] h m :Delivery time per household = single pack of cigarettes handover time × household order volume + average household distance time;

[0047] b h : Average number of bicycles per household per day;

[0048] h j : Payment time;

[0049] h g : Time to hand in delivery receipts.

[0050] The payment time and the delivery note submission time are zero or non-zero (for online payment and electronic uploading of notes, the payment time and the delivery note submission time are set to zero).

[0051] Preferably, the route planning sub-model also includes: a computer distribution network planning algorithm based on GIS, and assists in the approval and adjustment of routes based on the experience formed over a long period of time, and forms a dynamic optimization distribution mode based on the principles of bus planning, dynamic cutting, cycle scheduling and balanced operation;

[0052] The S12 includes:

[0053] (1) Planning the distribution bus, including: based on the mileage saving rule and the principle of maximum efficiency, using GIS geographic information technology, connecting the retail customers in each dynamic area in a reasonable sequence to form a dynamic distribution bus for each area; when special reasons arise, using manual fine-tuning to optimize the bus sequence to make the route division more reasonable and smooth;

[0054] (2) Optimizing area vehicles, including: establishing a minimum vehicle dispatch optimization model based on different vehicle models, and determining the total number of area vehicles based on the sales volume, customer distribution, and vehicle loading of each dynamic distribution area;

[0055] (3) Real-time segmented dispatching, including: combining the sales data of the day with the minimum dispatch optimization model to segment the dynamic distribution bus in each area, generate real-time delivery plans for each vehicle group, ensure the minimum number of vehicles dispatched each day, and reduce the proportion of last-mile dispatch;

[0056] (4) Balancing vehicle dispatching operations, including: This embodiment is implemented using computer software, by setting four control parameters: the number of delivery households, the total amount of deliveries, the operation time and the number of vehicle dispatching times, to ensure that the labor intensity of each delivery person is balanced under the dynamic optimization mode.

[0057] Preferably, the mathematical model of the intelligent distribution model is as shown in formula (3):

[0058]

[0059] The constraint conditions are shown in formula (4):

[0060]

[0061] Among them, t i represents the time when the vehicle arrives at customer i; w i represents the operation time of the vehicle at customer i; c i For customer i, i∈{0,1,2,…,N}; t ij Representing Client i to Client j Time, t ij =c ij ; l0 represents the latest service time of the distribution center; s i represents the delivery service time of customer i, s0 = 0; c ij is the specific time value from customer i to customer j; x ij is the time variable from customer i to customer j; t j is the time variable of customer j; i represents the customer number; j represents another customer number; k represents the vehicle number; N represents the total number of vehicles; Q represents the upper limit of time; e i Represents the initial service time value of customer i.

[0062] Preferably, the specific steps of solving the intelligent distribution model include:

[0063] (1) constructing a single-layer membrane structure comprising m basic membranes;

[0064] (2) Initialization: Randomly assign m cars to m basic membranes and initialize the pheromone matrix t of each basic membrane. k and the heuristic information matrix y k , k = 0, 1, ..., m; at the initial moment, all membranes have the same t k and k ;

[0065] (3) Vehicle construction path within the basic membrane: Vehicles in the basic membrane k use the pheromone matrix t within the membrane k and the heuristic information matrix y k , and use the sub-algorithm to iterate; and in each basic membrane, every time the vehicle moves one step, the pheromone matrix t in the membrane must be updated according to the pheromone update rule k ;

[0066] (4) Communication from the basic membrane to the surface membrane: Based on the co-transportation rule between membranes, each basic membrane outputs a car and the optimal path to the surface membrane, so the surface membrane contains m cars and paths; then the optimal path among the m paths is used to update the pheromone matrix t0 in the surface membrane using the global pheromone update rule;

[0067] (5) Surface membrane ants construct paths: The m vehicles in the surface membrane use t0 and y0, and after iterating using ACA, record their optimal path length;

[0068] (6) The surface membrane communicates with the basic membrane: The surface membrane sends its optimal path to each basic membrane and uses the global pheromone update rule to update the pheromone matrix t in each basic membrane. k , k=0,1,…,m;

[0069] (7) Rewriting the communication rules: Rewriting the communication rules is reflected in the changes of individual information within each membrane;

[0070] Through this intra-membrane sub-algorithm, the vehicle material in the membrane evolves continuously and eventually evolves into a "driving path", i.e., a feasible solution. After all vehicles in the basic membrane complete their evolution process in parallel, they are output to the surface membrane through the co-transportation rule in the communication rule. In the surface membrane, the rewriting rule is used again to evolve these feasible solutions to select high-quality solutions. Finally, the co-transportation rule is used again to input them into the basic membrane respectively. At this point, an iterative calculation is completed.

[0071] (8) Determine the termination condition: If the algorithm termination condition is met, execute step (9); otherwise, return to step (3);

[0072] (9) Optimal path output: The surface membrane outputs its optimal path as the solution to the problem.

[0073] Preferably, S2 includes:

[0074] S21, collecting commodity distribution process data; the commodity distribution process data includes operation information, vehicle condition picture data and abnormal information, wherein the operation information is collected and uploaded by the logistics implementation personnel's handheld electronic terminal, the vehicle condition picture data is collected and reported after inspection by the logistics implementation personnel, and the abnormal information includes abnormal road condition information of the vehicle traveling along the predetermined route, uploaded handover process abnormal information, abnormal picture data of the handover scene, merchant update information and merchant newly added coordinate information. The dispatching personnel interact with the logistics implementation personnel in real time about the abnormal information and correct the abnormal problems, adjust the logistics process and re-upload the handover data.

[0075] S22, classifying the commodity delivery process data according to data analysis purposes; the data analysis purposes include serving as a basis for big data analysis of handover history tracing, serving as basic data for delivery route planning algorithms, and serving as a basis for calculating the workload and performance of logistics personnel;

[0076] For use as a basis for big data analysis of handover history tracing, the classified data includes handover coordinates, handover time point, and handover scene information;

[0077] For use as basic data for delivery route planning algorithms, the classified data include handover coordinates, handover sequence, handover interval, and vehicle travel trajectory;

[0078] For purposes of calculating the workload and performance of logistics personnel, the classified data include the total volume of delivery operations;

[0079] S23, performing the data analysis on the commodity delivery process data; including: obtaining accurate customer handover time, time interval between customers and customer handover point coordinate information through real-time collection, long-term accumulation and big data analysis of delivery process data such as handover coordinates, handover sequence, handover interval and vehicle driving trajectory, and formulating clear customer sorting rules accordingly.

[0080] The second aspect of the present invention is to provide a flexible logistics distribution system based on an intelligent distribution model, comprising:

[0081] Model building module, used to build intelligent distribution models;

[0082] A data collection module is used to collect and process commodity distribution process data, and the processed commodity distribution process data is used to provide a data basis for the intelligent distribution model;

[0083] The logistics distribution module is used to input the currently processed commodity distribution process data into the intelligent distribution model, so as to obtain the optimal logistics distribution plan with the lowest total transportation cost and service time; and perform intelligent logistics distribution based on the optimal logistics distribution plan.

[0084] A third aspect of the present invention provides an electronic device, comprising a processor and a memory, wherein the memory stores a plurality of instructions, and the processor is configured to read the instructions and execute the method described in the first aspect.

[0085] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a plurality of instructions, and the plurality of instructions can be read by a processor and execute the method described in the first aspect.

[0086] Beneficial effects of the method and system of the present invention:

[0087] The present invention studies the methods and means of efficiently collecting data on finished cigarette delivery, such as delivery operation conditions, vehicle on-the-way monitoring data, and real-time status of commodity delivery, and based on this, establishes an intelligent delivery model based on sensitive factors such as delivery quantity, delivery distance, delivery tools, delivery links, and operation time, thereby improving cigarette delivery efficiency, reducing delivery link costs, and improving delivery service quality. Beneficial effects include:

[0088] (1) A dynamic optimization distribution model based on bus planning, dynamic cutting, cyclic scheduling, and balanced operation was designed. A route optimization model with line concatenation and center clustering as the core algorithm was implemented, which included customer area allocation, area identification, route generation, and inter-area adjustment. The algorithm was fast and the results were good.

[0089] (2) A method for efficiently collecting finished cigarette delivery data such as delivery operation status, vehicle in-transit monitoring data, and real-time status of commodity delivery through electronic delivery was established. Each node in the cigarette delivery process was monitored in real time, and data information such as cigarette delivery operation status, vehicle trajectory and in-transit data, and real-time status of commodity delivery (including customer handover coordinates, handover time, handover scene, handover interval, and handover sequence) were collected in real time. Methods for data classification and intelligent sorting and analysis of big data were proposed to provide a data basis for the intelligent delivery model algorithm.

[0090] (3) In order to adapt to the market-oriented reform trend of the industry, it is proposed to introduce the intelligent optimization algorithm based on the ECMC framework into the planning and design of the finished cigarette distribution routes of commercial enterprises, and a useful attempt is made for the real-time distribution scheduling method under the scenario of "real-time sales, real-time sorting, and real-time distribution". BRIEF DESCRIPTION OF THE DRAWINGS

[0091] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in the related technologies, the drawings required for use in the specific embodiments or the related technical descriptions will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0092] Figure 1 It is a schematic diagram of the overall design process of the distribution model according to the prior art;

[0093] Figure 2 A flow chart of a flexible logistics distribution method based on an intelligent distribution model provided according to an embodiment of the present invention;

[0094] Figure 3 A membrane structure diagram of an evolutionary communication (ECMC) type system framework provided according to an embodiment of the present invention;

[0095] Figure 4 A flowchart of solving a mathematical model of an intelligent distribution model provided according to an embodiment of the present invention;

[0096] Figure 5 This is an architecture diagram of a flexible logistics distribution system based on an intelligent distribution model provided according to an embodiment of the present invention;

[0097] Figure 6 An optimized train number trend diagram provided according to an embodiment of the present invention;

[0098] Figure 7 A structural diagram of an electronic device provided according to an embodiment of the present invention. DETAILED DESCRIPTION

[0099] The technical solution of the present invention will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0100] In the description of the present invention, it should be noted that the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc., indicating the orientation or positional relationship, are based on the orientation or positional relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and "third" are used for descriptive purposes only, and cannot be understood as indicating or implying relative importance.

[0101] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0102] Embodiment 1

[0103] See also Figure 1 This embodiment provides a logistics distribution scenario for cigarettes targeted by this embodiment. The method and system standardize the supervision of the entire process of cigarette distribution, rely on the electronic receipt process of cigarettes to collect real-time information data such as the operation status of the finished cigarette distribution link, vehicle in-transit monitoring data, and real-time status of commodity distribution. An intelligent distribution model is established based on sensitive factors such as distribution quantity, distribution distance, distribution tools, distribution links, and operation time. The model is verified through actual operation data, and finally an efficient real-time distribution model for finished cigarettes of commercial enterprises is established, which improves the efficiency of cigarette distribution, reduces the cost of distribution links, and improves the quality of distribution services.

[0104] like Figure 2 As shown, the first aspect of the present invention is to provide a flexible logistics distribution method based on an intelligent distribution model, comprising:

[0105] S1, establish intelligent distribution model;

[0106] As a preferred embodiment, the S1 includes:

[0107] S11, determining the influencing factors of the distribution process, and establishing a distribution operation time sub-model based on the influencing factors of the distribution process; wherein the influencing factors of the distribution process include fixed influencing factors of the distribution process and boundary condition influencing factors of the distribution process;

[0108] In this embodiment, the S11 includes:

[0109] (1) The fixed factors affecting the delivery process are determined as loading time, vehicle travel time, delivery time per household, average number of households per vehicle per day, payment time, and delivery ticket submission time;

[0110] (2) Calculating the basic data of the distribution area based on the historical data of a specific area and analyzing and estimating the basic data of the distribution area, and determining the boundary condition influencing factors of the distribution process based on the basic data of the distribution area; the boundary condition influencing factors of the distribution process include the upper limit of the distribution operation time, the upper limit of the number of households that can be delivered by a single delivery vehicle, the upper limit of the delivery time for each household, the average loading preparation time, the average cage truck loading time, the average number of cage trucks for single delivery per day, the average unit cage truck pickup time, the average delivery ticket submission time, the average household order quantity, the average number of households per vehicle per day, the average single package handover time, and the driving speed;

[0111] (3) A distribution operation time sub-model is established based on the fixed influencing factors of the distribution process and the boundary condition influencing factors of the distribution process; the distribution operation time sub-model is decomposed using the distribution operation time as a single-dimensional indicator and is expressed by formula (2):

[0112] H p =h z +2h x +h m ×b h +h j +h g (2);

[0113] Where: H p : Delivery operation time;

[0114] h z : Loading time = unit loading time × number of cage trucks + loading preparation time + picking time in temporary storage area;

[0115] Pick-up time in temporary storage area = pick-up time per cage truck × number of cage trucks;

[0116] h x : Vehicle driving time = one-way distance ÷ driving speed;

[0117] h m :Delivery time per household = single pack of cigarettes handover time × household order volume + average household distance time;

[0118] b h : Average number of bicycles per household per day;

[0119] h j : Payment time;

[0120] h g : Time to hand in delivery receipts.

[0121] In this embodiment, by counting and analyzing historical data from 2016 to 2023, the basic data of the distribution area is calculated to obtain the following data parameters.

[0122] Set the upper limit of the delivery operation time: 6.5 hours of delivery operation time + 1.5 hours of flexible time for meals, rest, etc. = 8 hours. Set the upper limit of the number of households for delivery single vehicle operation: 50 households

[0123] Set the delivery time limit for each household: 5 minutes

[0124] Comprehensive working time composition:

[0125] A: Loading preparation time (average): 5 minutes per vehicle.

[0126] B: Cage loading time (average): 5 seconds / cage

[0127] C: Number of cages delivered per vehicle per day (average): 3.4 cages

[0128] D: Unit cage truck pickup time (average): 5 minutes / cage

[0129] E: Delivery bill submission time (average value): 10 minutes. Those skilled in the art should know that this time can also be other set values, for example, in other embodiments, by electronically uploading bills, this time is 0 minutes.

[0130] F: Average order quantity per household: 5.6 packs

[0131] G: Daily average number of bicycles per household: 26

[0132] H: Payment time: 30 minutes. Those skilled in the art should know that this time may also be other set values. For example, in other embodiments, if payment is made online, this time is also 0.

[0133] I: Single package handover time (average): 20 seconds / package.

[0134] J: Driving speed: After three consecutive months of investigation, the speed of the entire network in the urban area during the daily (working day) period is measured as shown in Table 1. Table 1 is the speed of the entire network in the urban area.

[0135] Table 1

[0136] time 8:00--10:00 10:00--12:00 12:00--14:00 14:00--16:00 Average speed per time period 30.2km / h 35.6km / h 39.7km / h 32.9km / h

[0137] Therefore, the average driving speed is: 34.6km / h.

[0138] As a preferred implementation, as the real-time collected delivery process data continues to accumulate, the various parameters corresponding to the fixed influencing factors of the delivery process and the boundary condition influencing factors of the delivery process can be replaced by the big data analysis results. The intelligent model can identify the applicability of parameters through autonomous learning, continuously update and iterate, optimize parameter combinations, and actively adapt to the route planning and design of different customer combinations.

[0139] S12, establish route planning sub-model based on GIS computer distribution network planning algorithm;

[0140] As a preferred implementation, the route planning sub-model also includes: a GIS-based computer distribution network planning algorithm, and uses the experience formed over a long period of time to approve and adjust the routes, forming a dynamic optimization distribution model based on the principles of bus planning, dynamic cutting, cyclic scheduling and balanced operation.

[0141] As a preferred implementation, the S12 includes:

[0142] (1) Planning the distribution bus, including: based on the mileage saving rule and the principle of maximum efficiency, using GIS geographic information technology, connecting the retail customers in each dynamic area in a reasonable route sequence to form a dynamic distribution bus for each area; when new retail customers appear or encounter special reasons such as road restrictions or construction, the bus sequence is optimized and adjusted manually to make the route division more reasonable and smooth;

[0143] (2) Optimizing area vehicles, including: establishing a minimum vehicle dispatch optimization model based on different vehicle models, and determining the total number of area vehicles based on the sales volume of commodities (cigarettes in this embodiment), customer distribution, and vehicle loading in each dynamic distribution area. This step avoids the difficulties in vehicle management and personnel deployment caused by excessive vehicle allocation, and fails to take advantage of real-time route optimization;

[0144] (3) Real-time segmented dispatching, including: combining the sales data of the day with the minimum dispatch optimization model to segment the dynamic distribution bus in each area, generate real-time delivery plans for each vehicle group, ensure the minimum number of vehicles dispatched each day, and reduce the proportion of last-mile dispatch;

[0145] (4) Balanced dispatching operations, including: This embodiment is implemented using computer software. By setting four control parameters, namely, the number of delivery households, the total amount of deliveries, the operation time and the number of dispatches, the labor intensity of each deliveryman is balanced under the dynamic optimization mode, thereby solving the problem of uneven workload and unreasonable arrangement among deliverymen on fixed routes.

[0146] S13, integrating the distribution operation time sub-model and the route planning sub-model to form an intelligent distribution model; the intelligent distribution model is based on efficient real-time collection, classification analysis, autonomous learning and iterative evolution of distribution process data, and the final output distribution planning solution aims at the lowest total transportation cost and service time.

[0147] The process and theoretical basis for establishing the intelligent distribution model of this embodiment are as follows:

[0148] Based on the basic theory of path planning with time windows and various data extracted in real time, the ant colony clustering algorithm is combined with the cell membrane immune algorithm to avoid the defect of local optimal pheromone deposition of the ant colony clustering algorithm as much as possible, and the evolutionary communication (ECMC) system is used as the framework (such as Figure 3 The figure shows the membrane structure diagram of the evolutionary communication (ECMC) type system framework. The ant colony system is embedded in the basic membrane as an intra-membrane sub-algorithm, and the algorithm performance is improved through communication rules between membranes to quickly obtain the optimal distribution path at the global level; then it is cut according to the basic constraints, and finally a sales model that can adapt to delivery based on visits is achieved.

[0149] The basic constraints are:

[0150] ①The starting point and end point of each route are logistics centers;

[0151] ②The load of each vehicle shall not exceed the capacity limit;

[0152] ③ The operation time of each vehicle shall not exceed the prescribed time limit;

[0153] ④ Deliver to the customer's location within the specified time window;

[0154] Ultimately achieving the goal of lowest total transportation cost and service time.

[0155] As a preferred implementation, the mathematical model of the intelligent distribution model is shown in formula (3):

[0156]

[0157] The constraint conditions are shown in formula (4):

[0158]

[0159]

[0160] Among them, t i represents the time when the vehicle arrives at customer i; w i represents the operation time of the vehicle at customer i; c i For customer i, i∈{0,1,2,…,N}; t ij Representing Client i to Client j Time, t ij =c ij ; l0 represents the latest service time of the distribution center; s i represents the delivery service time of customer i, s0 = 0; c ij is the specific time value from customer i to customer j; x ij is the time variable from customer i to customer j; t jis the time variable of customer j; i represents the customer number; j represents another customer number; k represents the vehicle number; N represents the total number of vehicles; Q represents the upper limit of time; e i Represents the initial service time value of customer i.

[0161] As a preferred embodiment, Figure 4 The specific steps of solving the mathematical model of the intelligent distribution model are shown. Specifically, the specific steps of solving the intelligent distribution model include:

[0162] (1) constructing a single-layer membrane structure comprising m basic membranes;

[0163] (2) Initialization: Randomly assign m cars to m basic membranes and initialize the pheromone matrix t of each basic membrane. k and the heuristic information matrix y k , k = 0, 1, ..., m; at the initial moment, all membranes have the same t k and k ;

[0164] (3) Vehicle construction path within the basic membrane: Vehicles in the basic membrane k use the pheromone matrix t within the membrane k and the heuristic information matrix y k , and use the sub-algorithm to iterate; and in each basic membrane, every time the vehicle moves one step, the pheromone matrix t in the membrane must be updated according to the pheromone update rule k ;

[0165] (4) Communication from the basic membrane to the surface membrane: Based on the co-transportation rule between membranes, each basic membrane outputs a car and the optimal path to the surface membrane, so the surface membrane contains m cars and paths; then the optimal path among the m paths is used to update the pheromone matrix t0 in the surface membrane using the global pheromone update rule;

[0166] (5) Surface membrane ants construct paths: The m vehicles in the surface membrane use t0 and y0, and after iterating using ACA, record their optimal path length;

[0167] (6) The surface membrane communicates with the basic membrane: The surface membrane sends its optimal path to each basic membrane and uses the global pheromone update rule to update the pheromone matrix t in each basic membrane. k , k=0,1,…,m;

[0168] (7) Rewriting communication rules: Rewriting communication rules is reflected in the change of individual information within each membrane.

[0169] Through this intra-membrane sub-algorithm, the vehicle material in the membrane evolves continuously and eventually evolves into a "driving path", which is a feasible solution. After all vehicles in the basic membrane complete their evolution process in parallel, they are output to the surface membrane through the co-transportation rule in the communication rule. In the surface membrane, the rewrite rule is used again to evolve these feasible solutions to select high-quality solutions. Finally, the co-transportation rule is used again to input them into the basic membrane respectively. At this point, an iterative calculation is completed;

[0170] (8) Determine the termination condition: If the algorithm termination condition is met, execute step (9); otherwise, return to step (3);

[0171] (9) Optimal path output: The surface membrane outputs its optimal path as the solution to the problem.

[0172] S2, collecting and processing commodity distribution process data. For this embodiment, the finished cigarette distribution data is collected and processed; the processed commodity distribution process data is used to provide a data basis for the intelligent distribution model;

[0173] This embodiment intends to adopt electronic means, relying on the electronic handover process to monitor each node of the cigarette distribution process in real time, collect real-time data information such as the operation status of the cigarette distribution link, vehicle trajectory and in-transit data, and real-time status of commodity distribution (including customer handover coordinates, handover time, handover scene, handover interval, handover sequence), and conduct classification and big data analysis, intelligent organization, and provide a data basis for the intelligent distribution model; through sensitive factor analysis, the extracted key data is sorted by sensitivity, the key factors affecting the distribution process are determined and influenced, and the problems existing in the key links are solved in a targeted manner.

[0174] As a preferred embodiment, the S2 includes:

[0175] S21, collecting commodity distribution process data; the commodity distribution process data includes operation information, vehicle condition picture data and abnormal information, wherein the operation information is collected and uploaded by the logistics implementation personnel's handheld electronic terminal, the vehicle condition picture data is collected and reported after inspection by the logistics implementation personnel, and the abnormal information includes abnormal road condition information of the vehicle traveling along the predetermined route, uploaded handover process abnormal information, abnormal picture data of the handover scene, merchant update information and merchant newly added coordinate information. The dispatching personnel interact with the logistics implementation personnel in real time about the abnormal information and correct the abnormal problems, adjust the logistics process and re-upload the handover data.

[0176] In this embodiment, the specific implementation method includes:

[0177] (1) The deliveryman uploads operation information in real time through the handheld electronic terminal. Before the operation begins, the delivery task of the day is downloaded to the terminal to verify key information such as the area, number of delivery households, number of cigarettes, order of delivery to retailers, and delivery route. The deliveryman can apply online to change the delivery route and order of delivery according to the actual situation; if the delivery task of the day cannot be performed for some reason, the downloaded task can be rolled back; the dispatcher can adjust and re-issue the delivery task according to the actual situation of the day.

[0178] (2) Before the deliveryman leaves, he / she will check the vehicle’s condition, collect image data, and report the vehicle’s current mileage and condition. Vehicles with abnormal conditions can be replaced based on the actual situation. If there are no abnormal conditions, the deliveryman will confirm the departure of the vehicle. After the delivery is completed, the vehicle will return to the logistics center and report the vehicle’s mileage again, upload image data, and confirm the return trip. During the delivery process, the electronic terminal will record and upload the vehicle’s driving trajectory in real time.

[0179] (3) The vehicle travels along the scheduled route, and the delivery personnel provide real-time feedback on abnormal road conditions (such as road repairs, traffic restrictions, and traffic jams that affect delivery work); after arriving at the delivery area, the delivery personnel upload the handover process data (including handover coordinates, handover time, handover interval, and handover sequence) and take photos of the handover scene data (including photos of all cigarettes handed over by the household, photos of the consignee with the cigarettes, and the tobacco monopoly license number) in real time. If there is any abnormal situation, the alarm will be triggered in real time; the delivery personnel provide feedback on the updated information of the retailer (changes in contact information, changes in delivery date, etc.) and collect the coordinates of the newly added retailers; the dispatcher interacts with the delivery personnel in real time, abnormal problems are corrected immediately, the delivery process is adjusted in real time, and the handover data is uploaded again.

[0180] S22, classifying the commodity delivery process data according to data analysis purposes; the data analysis purposes include serving as a basis for big data analysis of handover history tracing, serving as basic data for delivery route planning algorithms, and serving as a basis for calculating the workload and performance of logistics personnel;

[0181] For use as a basis for big data analysis of handover history tracing, the classified data includes handover coordinates, handover time point, and handover scene information;

[0182] For use as basic data for delivery route planning algorithms, the classified data include handover coordinates, handover sequence, handover interval, and vehicle travel trajectory;

[0183] For purposes of calculating the workload and performance of logistics personnel, the classified data include total delivery operation data.

[0184] In this embodiment, as the real-time collected data accumulates, it is necessary to classify and manage it for analysis and utilization.

[0185] In the tobacco intelligent distribution scenario of this embodiment, data analysis is classified according to its purpose:

[0186] 1. The handover coordinates, handover time, and handover scene information are used as the classification of cigarette handover traces for real-time review by internal management personnel. If any problems are found during the handover, they will be immediately fed back to the delivery personnel’s terminal and corrected. After the review is passed, they will be archived to the server as a basis for big data analysis of handover history tracing;

[0187] 2. Handover coordinates, handover sequence, handover interval, and vehicle driving trajectory are used as distribution process data classification to form distribution process trajectory information as the basic data for distribution route planning algorithm;

[0188] 3. The total amount of delivery operation data is used as the basis for calculating the monthly workload and performance of the delivery personnel; the present invention focuses on the distribution process data.

[0189] S23, performing data analysis on the commodity distribution process data;

[0190] In this embodiment, the implementation of S23 can obtain accurate information such as customer handover time, time interval between customers, and customer handover point coordinates through real-time collection, long-term accumulation and big data analysis of delivery process data such as handover coordinates, handover sequence, handover interval, and vehicle driving trajectory. Based on this, clear customer sorting rules can be formulated to make the algorithm parameters more reasonable.

[0191] S3, inputting the currently processed commodity distribution process data into the intelligent distribution model, so as to obtain the optimal logistics distribution plan with the lowest total transportation cost and service time; and performing intelligent logistics distribution based on the optimal logistics distribution plan.

[0192] This embodiment uses an intelligent distribution model, relying on the cigarette distribution network layout optimization and one-cage distribution mode currently being implemented by tobacco logistics, to break the distribution mode of fixed time, fixed line, fixed person and fixed vehicle, and relies on determining the methods and means of efficiently collecting the distribution data of finished cigarettes, establishing an intelligent distribution model and verifying the model through actual operation data to establish a bus planning, dynamic cutting, real-time scheduling and balanced operation intelligent cigarette distribution model, thus forming an efficient dynamic optimization route mode for cigarette distribution.

[0193] Embodiment 2

[0194] The other settings of this embodiment are the same as those of the first embodiment, except that: in this embodiment, an electronic receipt mode is adopted in the receipt link, and an electronic receipt module is set up to perform electronic signatures and / or signature photos, confirm whether the delivery option exists, and whether there are damaged cigarettes and perform registration procedures to achieve electronic receipt. Through the electronic receipt mode, the delivery documents and handover scenes are information-based. By retrieving the handover photos of the disputed orders, verifying the recipient of the cigarettes, verifying the receipt of the bills, checking the types and quantities of the handed-over cigarettes, and checking the geographical coordinates of the households, key information can not only be provided to the internal management department for supervision, but also can be used as electronic auxiliary evidence for customer inquiries and the resolution of various disputes.

[0195] Embodiment 3

[0196] like Figure 5 As shown, this embodiment provides a flexible logistics distribution system based on an intelligent distribution model, including:

[0197] Model building module 101, used to build an intelligent distribution model;

[0198] The data collection module 102 is used to collect and process the commodity distribution process data, and the processed commodity distribution process data is used to provide a data basis for the intelligent distribution model;

[0199] The logistics distribution module 103 is used to input the currently processed commodity distribution process data into the intelligent distribution model, so as to obtain the optimal logistics distribution plan with the lowest total transportation cost and service time; and perform intelligent logistics distribution based on the optimal logistics distribution plan.

[0200] Actual operation verification of the model in the tobacco intelligent logistics distribution scenario

[0201] 1. Regional tobacco application

[0202] We have been piloting the application of intelligent distribution models since 2018. Based on the spatial distribution characteristics and sales characteristics of retailers, combined with the particularity of one of the most complex road networks in the world, we have explored efficient, reasonable and practical logistics distribution optimization scheduling solutions under the premise of a delivery-based visit strategy. Through dynamic route scheduling, a distribution model with dynamic optimization routes has been formed. According to factors such as the day's orders, vehicle information, customer information, geographic information, and maximum working hours, the original fixed routes are cut and reorganized, solving the problems of uncoordinated and unbalanced visits and deliveries between areas and differences in delivery volume, improving the loading rate of distribution vehicles, and effectively utilizing transportation resources.

[0203] As of the end of 2020, the tobacco direct distribution area has achieved 100% coverage of dynamic optimization of distribution routes using intelligent distribution model algorithms. In the relay area, the district bureau, as a pilot, officially adopted the intelligent model for distribution route optimization in November 2020. As of the end of 2020, the application of the model algorithm has achieved good results and no obvious problems have been found.

[0204] 2. Analysis of model application effect

[0205] Since the application of the intelligent model between 2017 and 2020, under the premise of basically stable delivery volume, the direct distribution area has streamlined 16 vehicles, the number of households delivered by a single vehicle has increased from 27 to 34, and the average number of households per vehicle has increased by 25%; the loading rate of a single vehicle has exceeded 90%; and the effective utilization rate of transportation capacity has increased by 13%. As shown in Table 2. The trend chart of optimized train times is as follows Figure 6 shown.

[0206] Table 2 Statistics of optimized delivery vehicles in tobacco direct distribution area

[0207] years Number of households Delivery volume Optimize train schedule Improved efficiency 2016 737889 485051.736 26364 0 2017 636744 485736.576 23002 12.75％ 2018 595904 486963.024 21543 18.29％ 2019 659050 482411.316 21035 20.21％ 2020 642179 482918.192 20014 24.09％ Total average 654353.2 394182.1688 22191.6 18.83％

[0208] As the utilization rate of delivery capacity increases, the coverage area of ​​direct delivery areas has gradually expanded. By the end of 2020, Mentougou District's Mencheng area, Changping District's Tiantongyuan, Huilongguan, Huoying, Xi'erqi areas, Shunyi District's Tianzhu and Liqiao areas have been successively designated as direct delivery areas. According to the statistics of delivery volume, Daxing accounts for 21% of the district, Shunyi accounts for 13% of the district, Changping accounts for 31% of the district, and Mentougou accounts for 73% of the district.

[0209] The present invention also provides a memory storing a plurality of instructions, wherein the instructions are used to implement the method as in the first embodiment.

[0210] like Figure 7 As shown, the present invention also provides an electronic device, including a processor 301 and a memory 302 connected to the processor 301, the memory 302 stores a plurality of instructions, and the instructions can be loaded and executed by the processor so that the processor can execute the method as in the first embodiment.

[0211] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A flexible logistics distribution method based on an intelligent distribution model, characterized in that: include: S1, establish intelligent distribution model; S2, collecting and processing commodity distribution process data, the processed commodity distribution process data is used to provide a data basis for the intelligent distribution model; S3, inputting the currently processed commodity distribution process data into the intelligent distribution model, so as to obtain the optimal logistics distribution plan with the lowest total transportation cost and service time; and performing intelligent logistics distribution based on the optimal logistics distribution plan.

2. According to claim 1, a flexible logistics distribution method based on an intelligent distribution model is characterized in that: The S1 includes: S11, determining the influencing factors of the distribution process, and establishing a distribution operation time sub-model based on the influencing factors of the distribution process; wherein the influencing factors of the distribution process include fixed influencing factors of the distribution process and boundary condition influencing factors of the distribution process; S12, establish route planning sub-model based on GIS computer distribution network planning algorithm; S13, integrating the distribution operation time sub-model and the route planning sub-model to form an intelligent distribution model; the intelligent distribution model is based on efficient real-time collection, classification analysis, autonomous learning and iterative evolution of distribution process data, and the final output distribution planning solution aims at the lowest total transportation cost and service time.

3. According to claim 2, a flexible logistics distribution method based on an intelligent distribution model is characterized in that: The S11 includes: (1) The fixed factors affecting the delivery process are determined as loading time, vehicle travel time, delivery time per household, average number of households per vehicle per day, payment time, and delivery ticket submission time; (2) Calculating the basic data of the distribution area based on the historical data of a specific area and analyzing and estimating the basic data of the distribution area, and determining the boundary condition influencing factors of the distribution process based on the basic data of the distribution area; the boundary condition influencing factors of the distribution process include the upper limit of the distribution operation time, the upper limit of the number of households that can be delivered by a single delivery vehicle, the upper limit of the delivery time for each household, the average loading preparation time, the average cage truck loading time, the average number of cage trucks for single delivery per day, the average unit cage truck pickup time, the average delivery ticket submission time, the average household order quantity, the average number of households per vehicle per day, the average single package handover time, and the driving speed; (3) A distribution operation time sub-model is established based on the fixed influencing factors of the distribution process and the boundary condition influencing factors of the distribution process; the distribution operation time sub-model is decomposed using the distribution operation time as a single-dimensional indicator and is expressed by formula (2): H p =h z +2h x +h m ×b h +h j +h g (2); Where: H p : Delivery operation time; h z : Loading time = unit loading time × number of cage trucks + loading preparation time + picking time in temporary storage area; Pick-up time in temporary storage area = pick-up time per cage truck × number of cage trucks; h x : Vehicle driving time = one-way distance ÷ driving speed; h m :Delivery time per household = single pack of cigarettes handover time × household order volume + average household distance time; b h : Average number of bicycles per household per day; h j : Payment time; h g : Delivery bill submission time; The payment time and delivery note submission time are zero or non-zero.

4. According to claim 3, a flexible logistics distribution method based on an intelligent distribution model is characterized in that: The route planning sub-model also includes: a computer distribution network planning algorithm based on GIS, and assists in the approval and adjustment of routes based on the experience formed over a long period of time, and forms a dynamic optimization distribution model based on the principles of bus planning, dynamic cutting, cycle scheduling and balanced operation; The S12 includes: (1) Planning the distribution bus, including: based on the mileage saving rule and the principle of maximum efficiency, using GIS geographic information technology, connecting the retail customers in each dynamic area in a reasonable sequence to form a dynamic distribution bus for each area; when special reasons arise, using manual fine-tuning to optimize the bus sequence to make the route division more reasonable and smooth; (2) Optimizing area vehicles, including: establishing a minimum vehicle dispatch optimization model based on different vehicle models, and determining the total number of area vehicles based on the sales volume, customer distribution, and vehicle loading of each dynamic distribution area; (3) Real-time segmented dispatching, including: combining the sales data of the day with the minimum dispatch optimization model to segment the dynamic distribution bus in each area, generate real-time delivery plans for each vehicle group, ensure the minimum number of vehicles dispatched each day, and reduce the proportion of last-mile dispatch; (4) Balanced dispatch operations, including: using computer software to set four control parameters: number of delivery households, total delivery volume, operation time, and number of dispatches, to ensure that the labor intensity of each delivery person is balanced under the dynamic optimization mode.

5. A flexible logistics distribution method based on an intelligent distribution model according to claim 4, characterized in that: The mathematical model of the intelligent distribution model is shown in formula (3): The constraint conditions are shown in formula (4): have been i ≤t i +w i ≤I i ,i∈{1,...,N} Among them, t i represents the time when the vehicle arrives at customer i; w i represents the operation time of the vehicle at customer i; c i For customer i, i∈{0,1,2,…,N}; t ij represents the time from customer i to customer j, t ij =c ij ; l0 represents the latest service time of the distribution center; s i represents the delivery service time of customer i, s0 = 0; c ij is the specific time value from customer i to customer j; x ij is the time variable from customer i to customer j; t j is the time variable of customer j; i represents the customer number; j represents another customer number; k represents the vehicle number; N represents the total number of vehicles; Q represents the upper limit of time; e i represents the initial service time of customer i.

6. A flexible logistics distribution method based on an intelligent distribution model according to claim 5, characterized in that: The specific steps of solving the intelligent distribution model include: (1) constructing a single-layer membrane structure comprising m basic membranes; (2) Initialization: Randomly assign m cars to m basic membranes and initialize the pheromone matrix t of each basic membrane. k and the heuristic information matrix y k , k = 0, 1, ..., m; at the initial moment, all membranes have the same t k and k ; (3) Vehicle construction path within the basic membrane: Vehicles in the basic membrane k use the pheromone matrix t within the membrane k and the heuristic information matrix y k , and use the sub-algorithm to iterate; and in each basic membrane, every time the vehicle moves one step, the pheromone matrix t in the membrane must be updated according to the pheromone update rule k ; (4) Communication from the basic membrane to the surface membrane: Based on the co-transportation rule between membranes, each basic membrane outputs a car and the optimal path to the surface membrane, so the surface membrane contains m cars and paths; then the optimal path among the m paths is used to update the pheromone matrix t0 in the surface membrane using the global pheromone update rule; (5) Surface membrane ants construct paths: The m vehicles in the surface membrane use t0 and y0, and after iterating using ACA, record their optimal path length; (6) The surface membrane communicates with the basic membrane: The surface membrane sends its optimal path to each basic membrane and uses the global pheromone update rule to update the pheromone matrix t in each basic membrane. k , k=0,1,…,m; (7) Rewriting the communication rules: Rewriting the communication rules is reflected in the changes of individual information within each membrane; Through this intra-membrane sub-algorithm, the vehicle material in the membrane evolves continuously and eventually evolves into a "driving path", which is a feasible solution. After all vehicles in the basic membrane complete their evolution process in parallel, they are output to the surface membrane through the co-transportation rule in the communication rule. In the surface membrane, the rewriting rule is used again to evolve these feasible solutions to select high-quality solutions. Finally, the co-transportation rule is used again to input them into the basic membrane respectively. At this point, an iterative calculation is completed. (8) Determine the termination condition: If the algorithm termination condition is met, execute step (9); otherwise, return to step (3); (9) Optimal path output: The surface membrane outputs its optimal path as the solution to the problem.

7. A flexible logistics distribution method based on an intelligent distribution model according to claim 6, characterized in that: The S2 comprises: S21, collecting commodity distribution process data; the commodity distribution process data includes operation information, vehicle condition picture data and abnormal information, wherein the operation information is collected and uploaded by the logistics implementation personnel's handheld electronic terminal, the vehicle condition picture data is collected and reported after inspection by the logistics implementation personnel, and the abnormal information includes abnormal road condition information of the vehicle traveling along the scheduled route, uploaded handover process abnormal information, abnormal handover scene picture data, merchant update information and merchant newly added coordinate information. The dispatcher interacts with the logistics implementation personnel in real time about the abnormal information and corrects the abnormal problem, adjusts the logistics process and re-uploads the handover data; S22, classifying the commodity delivery process data according to data analysis purposes; the data analysis purposes include serving as a basis for big data analysis of handover history tracing, serving as basic data for delivery route planning algorithms, and serving as a basis for calculating the workload and performance of logistics personnel; For use as a basis for big data analysis of handover history tracing, the classified data includes handover coordinates, handover time point, and handover scene information; For use as basic data for delivery route planning algorithms, the classified data include handover coordinates, handover sequence, handover interval, and vehicle travel trajectory; For purposes of calculating the workload and performance of logistics personnel, the classified data include the total volume of delivery operations; S23, performing the data analysis on the commodity delivery process data; including: obtaining accurate customer handover time, time interval between customers and customer handover point coordinate information through real-time collection, long-term accumulation and big data analysis of delivery process data such as handover coordinates, handover sequence, handover interval and vehicle driving trajectory, and formulating clear customer sorting rules accordingly.

8. A flexible logistics distribution system based on an intelligent distribution model, used to implement any of the methods described in claims 1-7, characterized in that: include: Model building module, used to build intelligent distribution models; A data collection module is used to collect and process commodity distribution process data, and the processed commodity distribution process data is used to provide a data basis for the intelligent distribution model; The logistics distribution module is used to input the currently processed commodity distribution process data into the intelligent distribution model, so as to obtain the optimal logistics distribution plan with the lowest total transportation cost and service time; and perform intelligent logistics distribution based on the optimal logistics distribution plan.

9. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores a plurality of instructions, and the processor is used to read the instructions and execute the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a plurality of instructions, and the plurality of instructions can be read by a processor to execute the method according to any one of claims 1 to 7.

Images