Cross-border e-commerce logistics big data optimization system

Abstract

The present invention relates to the technical field of cargo loading optimization, and in particular to a cross-border e-commerce logistics big data optimization system, which comprises a data acquisition unit, a data storage unit, a loading algorithm unit and a real-time guidance unit. The present invention obtains cargo size, weight and shape data through the data acquisition unit, the storage unit stores cargo and transport tool data, the loading algorithm unit divides three-dimensional grid units according to transport tool space data, classifies cargo by weight using a clustering analysis algorithm, determines the best position of a single cargo through shape matching and a heuristic search algorithm, groups and arranges cargo using a path planning algorithm in combination with information such as a delivery sequence, and the real-time guidance unit receives new data during loading, re-runs the algorithm and generates guidance information. The system effectively solves the problem of loading space utilization and improves space utilization.

Description

Technical Field

[0001] The present invention relates to the technical field of cargo loading optimization, and in particular to a cross-border e-commerce logistics big data optimization system. Background Art

[0002] Cargo loading optimization is an important technology. In the field of cross-border e-commerce logistics, the rationality of cargo loading directly affects transportation efficiency and cost. With the vigorous development of cross-border e-commerce logistics, the rationality of cargo loading has become the focus of industry attention. Traditional cargo loading methods have significant defects in spatial location utilization, which seriously restricts the overall benefits of logistics transportation.

[0003] In actual operations, loaders usually rely on limited experience and extremely simple rules to place goods. They rarely delve into the detailed structure of the interior space of the transport vehicle. They lack systematic analysis and response strategies for the complex structural features inside the transport vehicle, such as the beam-column structure in the truck compartment, the curvature of the corners of the container, and the irregular bulkheads in the cabin of the cargo plane. When faced with regular-shaped cartons or large mechanical parts of various shapes, due to the lack of effective data support and algorithm guidance, loaders cannot accurately judge the adaptability of goods of different shapes to the interior space of the transport vehicle. In order to solve this technical problem, we provide a cross-border e-commerce logistics big data optimization system. Summary of the invention

[0004] The purpose of the present invention is to provide a cross-border e-commerce logistics big data optimization system to solve the problems raised in the above-mentioned background technology.

[0005] The present invention divides the three-dimensional grid units based on the internal space data of the transportation vehicle, initializes the relevant list, classifies the cargo weight using the cluster analysis algorithm according to the load weight limit, determines the optimal placement of a single cargo through shape matching and heuristic search algorithms, and uses a path planning algorithm to group and arrange the cargo in combination with the distribution sequence, logistics network topology and historical distribution time data.

[0006] To achieve the above purpose, a cross-border e-commerce logistics big data optimization system is provided, including a data collection unit, a data storage unit, a loading algorithm unit and a real-time guidance unit;

[0007] The data acquisition unit includes a laser scanning device, a weighing sensor and an image analysis module, wherein the laser scanning device is used to collect the size of the goods, the weighing sensor is used to obtain the weight of the goods, and the image analysis module is used to obtain the shape data of the goods;

[0008] The data storage unit is used to store cargo data and internal space data of transportation tools, the transportation tools include containers, cargo planes and trucks, and the internal space data includes three-dimensional dimensions, load weight limit and internal fixed structure position information;

[0009] The loading algorithm unit divides the internal space data of the transport vehicle into three-dimensional grid units, each unit having coordinate information, accommodating weight and volume attributes, and initializes a cargo list and an occupied space list at the same time. According to the carrying weight limit of the transport vehicle, the cargo is classified by weight using a clustering analysis algorithm to form cargo groups of different weight levels. For each cargo, the shape data is matched with the shape of the three-dimensional grid unit of the unoccupied space in the transport vehicle, and the best placement position is found using a heuristic search algorithm. At the same time, according to the delivery sequence information of cargoes at different destinations, combined with the logistics network topology and historical delivery time data, the path planning algorithm is used to group the cargoes, and the cargoes with the same destination and similar delivery time are placed together, and the cargoes are arranged in sequence according to the expected unloading order.

[0010] During the loading process, the real-time guidance unit continuously receives new cargo data and transportation plan change data, such as temporarily adding cargo or changing destination information, re-runs the loading algorithm, dynamically adjusts the loading plan, and generates cargo loading guidance information based on the algorithm calculation results to guide the loading of cargo in the transport vehicle.

[0011] As a further improvement of the technical solution, when dividing the three-dimensional grid unit, the loading algorithm unit uses a polygonal grid division method to process the irregular space inside the transportation vehicle, as follows:

[0012] Acquire the three-dimensional dimensions of the internal space data of the transport vehicle and the position information of the internal fixed structure from the data storage unit, and use the three-dimensional model-based analysis algorithm to identify the internal space contour of the transport vehicle;

[0013] Selecting basic polygons according to the shape characteristics of the irregular space, wherein the shape characteristics are bevel space and curved surface space, and setting the area range of each polygon mesh and the size of the inner angle of the polygon, and finally determining a mesh generation algorithm according to the selected polygon;

[0014] A three-dimensional coordinate is assigned to each polygon mesh vertex, and the accommodating weight and volume of each polygon mesh are proportionally allocated according to the carrying weight limit of the transportation vehicle and the volume information of the irregular space.

[0015] As a further improvement of the technical solution, when the loading algorithm unit divides the transport tool space into three-dimensional grid units, the side length of the grid unit is determined according to the minimum spatial dimension of the transport tool and the preset precision coefficient, as follows:

[0016] The internal space data of the transportation vehicle is obtained from the data storage unit. The values ​​of the three dimensions of length, width and height are compared to find the minimum value. The minimum value is defined as the minimum space dimension. The preset accuracy coefficient is determined according to the accuracy requirements of logistics loading and the size of the goods. Finally, the minimum space dimension is multiplied by the preset accuracy coefficient to obtain the side length of the grid unit.

[0017] As a further improvement of the technical solution, the cluster analysis algorithm in the loading algorithm unit is a K-means algorithm, and the number of clusters is dynamically determined according to the weight limit of the transport vehicle and the weight distribution characteristics of the cargo, as follows:

[0018] Extracting the carrying weight limit information of the transport vehicle from the data storage unit, and preliminarily dividing the number of theoretical weight intervals according to the carrying weight limit of the transport vehicle and the cargo weight classification granularity in logistics operations, wherein the cargo weight classification granularity is used to determine the size and accuracy of the weight interval division;

[0019] Obtain a list of cargo weight data from the data acquisition unit, perform statistical analysis on the cargo weight data, determine the distribution range and mean of the weight, and use data clustering to identify the clustered areas of cargo weight based on the statistical results;

[0020] Comprehensively analyze the number of theoretical weight intervals under the weight limit of the transport vehicle and the clustering areas in the cargo weight distribution characteristics. If the mean value of the cargo weight distribution is less than or equal to the first threshold, it indicates that the cargo weight distribution is concentrated. When the sum of the absolute values ​​of the differences between the boundary value of the cargo weight clustering area and the boundary value of the theoretical interval division is greater than or equal to the second threshold, it is determined to exceed the theoretical interval division range. At this time, the number of clusters is adjusted mainly based on the cargo weight clustering area. The first threshold is a specific value determined based on the actual logistics transportation situation and statistical analysis, and the second threshold is a specific value determined based on the weight limit of the transport vehicle and logistics operation experience.

[0021] As a further improvement of the technical solution, the heuristic search algorithm in the loading algorithm unit introduces a heuristic function to search for the best position, as follows:

[0022] The heuristic function combines the matching degree between the shape of the goods and the unoccupied space and the distance between the goods and the center of the expected placement area to search for the best placement position, as follows:

[0023] For cargo, the shape data acquired by the data acquisition unit is used to extract shape features. For unoccupied space, it is represented by three-dimensional network units divided according to transportation tools, and each unit has its shape and size information.

[0024] Calculate the volume of the cargo and the volume of the unoccupied space, and calculate the volume matching degree based on the volume of the cargo and the volume of the unoccupied space. Then use the shape similarity measurement method to calculate the distance between the shape contour of the cargo and the shape contour of the unoccupied space, and then convert it into matching degree through a function. Finally, the weighted sum of the volume matching degree and the shape contour matching degree is obtained to obtain the comprehensive shape matching degree.

[0025] According to the loading strategy of the transport vehicle, the delivery order and weight of the goods, the expected placement area of ​​the goods is preliminarily determined, and the distance from the goods to the center of the expected placement area is calculated in the three-dimensional coordinate system of the transport vehicle. Finally, the matching degree of the shape of the goods with the unoccupied space and the distance from the goods to the center of the expected placement area are combined to construct the heuristic function.

[0026] As a further improvement of the technical solution, when the loading algorithm unit uses the path planning algorithm to group the goods, the matching degree between the shape of the goods and the unoccupied space in the heuristic function is used to optimize the placement order within the group of goods at the same destination, as follows:

[0027] For goods with similar delivery times and the same destination, first determine the time proximity threshold based on the fluctuation range of historical delivery time data. If goods A and goods B are within this threshold, calculate the matching degree of goods A and the current unoccupied space. and the matching degree of cargo B with the current unoccupied space ,like , then in the expected unloading sequence, cargo A is placed at the priority unloading position, and the distance between the priority unloading position and the expected unloading channel in the transport vehicle meets the preset shortest distance standard.

[0028] As a further improvement of the technical solution, the loading algorithm unit combines the logistics network topology structure to improve the estimated placement area determination method in the heuristic function, as follows:

[0029] Analyze the congestion probability of each node in the logistics network based on historical data and real-time traffic information For cargo passing through nodes with high congestion probability, when determining the expected placement area, it is offset to the center of gravity position in the transport vehicle, as follows:

[0030] Assume the offset coefficient is ;in is a constant determined according to the type of transport and the characteristics of the cargo, is the set of logistics nodes that the cargo transportation path passes through, according to The value adjusts the center coordinates of the estimated placement area.

[0031] As a further improvement of the technical solution, the loading algorithm unit adjusts the priority of cargo grouping by using the distance factor in the heuristic function during cargo grouping, as follows:

[0032] Construct delivery time intervals for each destination based on historical delivery time data For goods whose delivery time is at the front end of the interval, the front end of the interval is ; Increase the weight of the distance factor in the heuristic function and set the dynamic weight coefficient And modify the heuristic function to Adjust the priority of cargo grouping.

[0033] As a further improvement of the present technical solution, after receiving new cargo data and transport plan change data, the real-time guidance unit re-runs the loading algorithm, assigns priority calculation marks to newly added cargo or cargo with changed destinations, and for temporarily added cargo, determines the priority calculation order by calculating the degree of impact on the subsequent unloading order based on its weight, shape and estimated joining time; for cargo with changed destination information, determines the priority calculation order according to the position of the new destination in the logistics network topology and the estimated delivery time.

[0034] Compared with the prior art, the present invention has the following beneficial effects:

[0035] In the cross-border e-commerce logistics big data optimization system, dividing the three-dimensional grid units based on the internal space data of the transport vehicle helps to accurately describe the details of the internal space of the transport vehicle, initialize related lists, effectively avoid data confusion, ensure that the loading operation is carried out in an orderly manner, and classify the weight of the goods according to the load weight limit using the cluster analysis algorithm to improve the safety and stability of transportation. The shape matching and heuristic search algorithms are used to determine the best placement of a single cargo, reduce the waste caused by unreasonable use of space, and significantly improve space utilization. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is an overall block diagram of the present invention.

[0037] The meaning of each number in the figure is:

[0038] 1. Data acquisition unit; 2. Data storage unit; 3. Loading algorithm unit; 31. Space division module; 32. Weight classification module; 33. Matching placement module; 34. Sequence optimization module; 4. Real-time guidance unit. DETAILED DESCRIPTION

[0039] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. 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.

[0040] The present invention provides a cross-border e-commerce logistics big data optimization system, please refer to Figure 1 As shown, it includes a data acquisition unit 1, a data storage unit 2, a loading algorithm unit 3 and a real-time guidance unit 4;

[0041] The data acquisition unit 1 includes a laser scanning device, a weighing sensor and an image analysis module. The laser scanning device is used to collect the size of the goods, the weighing sensor is used to obtain the weight of the goods, and the image analysis module is used to obtain the shape data of the goods, which provides a comprehensive and accurate data basis for cross-border e-commerce logistics cargo loading, thereby enabling a series of subsequent data-based optimization operations to improve the space utilization of transportation tools, the rationality of cargo weight distribution and delivery efficiency.

[0042] The data storage unit 2 is used to store cargo data and internal space data of transportation tools. The transportation tools include containers, cargo planes and trucks. The internal space data includes three-dimensional dimensions, load weight limits and internal fixed structure position information, which provides comprehensive and accurate data support for cross-border e-commerce logistics cargo loading, so that the loading process can be scientifically planned based on various data, ensuring that the cargo is adapted to the transportation tools, ensuring transportation safety, improving space utilization efficiency, and optimizing the overall logistics process.

[0043] The loading algorithm unit 3 divides the internal space data of the transport vehicle into three-dimensional grid units. Each unit has coordinate information, accommodating weight and volume attributes, thereby realizing accurate quantification and positioning of the transport vehicle space, helping to make full use of every space in the transport vehicle, thereby improving cargo loading efficiency and reducing transportation costs.

[0044] When the loading algorithm unit 3 divides the three-dimensional grid unit, the irregular space inside the transportation tool is processed by a polygonal grid division method, as follows:

[0045] Acquire the length, width, height and internal fixed structure position information of the internal space data of the transport vehicle from the data storage unit 2, and use the three-dimensional model-based analysis algorithm to identify the internal space contour of the transport vehicle;

[0046] Select basic polygons according to the shape characteristics of the irregular space. Select basic polygons for oblique space and curved space. For example, triangles or quadrilaterals can be selected for oblique space, and polygon approximation can be selected for curved space. Set the area range of each polygonal mesh and the range of the inner angle of the polygon. If a triangle is selected as the basic polygon, determine the specific triangle shape based on the sum of the inner angles being 180 degrees and the set inner angle range. Through polygonal mesh division specifically for irregular space, it can better fit the complex shape structure inside the transportation vehicle, make the space division more practical, avoid the space waste and mismatch problems caused by traditional regular mesh division in irregular areas, and effectively utilize the irregular space inside the transportation vehicle.

[0047] Finally, the mesh generation algorithm is determined according to the selected polygon. For example, the Delaunay triangulation algorithm is used for the triangular mesh, and the three-dimensional coordinates are assigned to each polygon mesh vertex. The weight limit of the transport vehicle is set to , the volume of the irregular space is , then each polygonal mesh can accommodate weight ; is the volume of the polygon mesh, and the volume attribute is , the accommodative weight and volume of each polygonal grid are allocated proportionally, ensuring the reasonable distribution of cargo weight and volume in irregular space, taking into account the carrying limitations of transportation vehicles, improving the safety and space utilization of loading, and providing more accurate and detailed spatial information for the loading algorithm.

[0048] A reasonable method for determining the side length of grid cells can reduce the number of unnecessary grids, reduce the amount of calculation in the algorithm during spatial search and matching, and improve calculation efficiency. Accurate cargo weight classification helps to better balance the center of gravity of the transport vehicle during subsequent loading, reduce safety hazards caused by uneven weight distribution, and improve the accuracy of the loading plan.

[0049] The loading algorithm unit 3 initializes the cargo list and the occupied space list while dividing the grid units, and uses the cluster analysis algorithm to classify the cargo by weight according to the load weight limit of the transport vehicle to form cargo groups of different weight levels, as follows:

[0050] Obtain the internal space data of the transportation tool from the data storage unit 2, which is relatively long. ,Width and high Find the minimum value of these three dimensions and define the minimum value as the minimum spatial dimension. , accurately grasp the key limiting factors of the internal space of the transport vehicle, provide the most practical basic size for subsequent grid division, and determine the preset accuracy coefficient according to the accuracy requirements of logistics loading and cargo size Finally, the minimum spatial dimension is multiplied by the preset precision coefficient to obtain the grid unit side length ; By calculating the side length through simple multiplication, we can get a uniform and reasonable grid unit, providing an accurate reference standard for space quantification and cargo placement.

[0051] Initialize the cargo list at the same time ; Clearly sort out the status of goods and space, and facilitate real-time tracking and management during the loading process, including is the quantity of goods, For the List of dimensions, weight, shape and space occupied by the goods ; is the number of occupied space units, Indicates The information of occupied space units is stored in the cargo containers, which avoids confusion of cargo information and repeated occupation of space, and improves the orderliness and accuracy of loading operations.

[0052] According to the weight limit of the transport vehicle , based on the granularity of cargo weight classification in logistics operations , with 100 kg as a classification granularity, the theoretical weight range is preliminarily divided into ; It fully considers the carrying capacity of the means of transport, provides a preliminary framework for reasonable grouping, and makes the weight classification of goods have a scientific basis based on the actual situation of the means of transport, avoiding blind classification.

[0053] Get cargo weight data list from data collection unit ; Calculate the weight mean and distribution range, the distribution range is the maximum value and minimum value The difference between the weight of the cargo and the weight of the cargo are clustered using a data clustering algorithm. ; In order to gather the number of areas, we deeply explored the characteristics of cargo weight distribution, which can more accurately find the weight concentration areas, make the cargo grouping more in line with the actual weight distribution, and improve the accuracy of weight level division.

[0054] Comprehensively analyze the number of theoretical weight intervals under the weight limit of transportation vehicles and the clustering areas in the cargo weight distribution characteristics, and set the first threshold as , the second threshold is ,like and ; The number of clusters is adjusted based on the cargo weight clustering area to adapt to the distribution of different cargo weights and avoid the limitations of fixed classification methods. Then, based on the final number of clusters, the cluster analysis algorithm is used to divide the cargo into cargo groups of different weight levels. , s is the number of clusters, which realizes the scientific stratification of goods by weight and facilitates the reasonable arrangement of positions during loading.

[0055] The internal space of a transport vehicle is complex, and different positions have different effects on the placement of goods. Constructing a heuristic function based on the expected placement area and distance factors can make the search direction more in line with actual loading requirements and increase the probability of finding the best position. For each cargo, its shape data is matched with the shape of the three-dimensional grid unit in the unoccupied space in the transport vehicle, and a heuristic search algorithm is used to find the best placement position.

[0056] The heuristic search algorithm in loading algorithm unit 3 introduces a heuristic function to search for the best position, as follows:

[0057] For each cargo, the shape data is acquired by the data acquisition unit 1. For the unoccupied space, the three-dimensional network unit is represented according to the transportation means. Let the first The shape and size information of each network unit is ;in The length, width and height of the grid unit are respectively used to comprehensively and meticulously grasp the shape information of the cargo, providing a rich data basis for subsequent precise matching and improving the compatibility of the loading plan for different cargoes.

[0058] Calculate cargo volume The volume of unoccupied space , assuming the length, width, and height of the goods are ,but ; For unoccupied space grid cells, their volume It provides an intuitive quantitative indicator to measure the degree of fit between cargo and space, avoids placing oversized or undersized cargo in inappropriate spaces, improves space utilization efficiency, and calculates the volume matching degree based on the cargo volume and the volume of unoccupied space. , The matching degree indicates the degree of volume fit between the cargo and the unoccupied space. The closer it is to 1, the better the volume fit. It can clearly reflect the volume matching relationship and quickly select spaces with more suitable volumes from many possible placement locations, thus reducing unnecessary search calculations.

[0059] Then, the shape similarity measurement method is used to calculate the distance between the shape contour of the cargo and the shape contour of the unoccupied space. Suppose the set of cargo shape contour points is , the set of contour points of the unoccupied space grid unit shape is , calculate the shape contour distance between the two ; It can deeply analyze the differences in shape contours, provide more accurate shape matching evaluation, and accurately find suitable space for irregularly shaped goods, further optimize space utilization, and then convert shape contour distance into matching degree through function conversion ,when hour, Indicates that the shape matches perfectly, when hour, , indicating that the shape difference is the largest, the distance metric is converted into a matching index that is convenient for comprehensive calculation with the volume matching degree, so that the shape contour matching degree and the volume matching degree are weighted and summed under the same standard, which improves the rationality of the comprehensive shape matching degree calculation.

[0060] Assume the weight of volume matching is , the weight of shape contour matching is , then the comprehensive shape matching degree By adjusting The value of can emphasize the importance of volume matching or shape contour matching according to actual needs, make the matching results more in line with actual loading needs, and improve the pertinence and adaptability of the loading plan.

[0061] According to the loading strategy of the transport vehicle, the order of delivery of the goods and the weight of the goods, the estimated placement area of ​​the goods is preliminarily determined as , fully considering various practical needs during transportation, making the cargo placement more reasonable, which is conducive to improving unloading efficiency and reducing the risk of cargo damage during transportation, and setting the coordinates of the cargo in the three-dimensional coordinate system of the transport vehicle as , the coordinates of the center of the expected placement area are , then the distance from the cargo to the center of the expected placement area is:

[0062] ;

[0063] It provides a quantitative basis for measuring the position relationship between the goods and the expected placement area, and can more accurately evaluate the advantages and disadvantages of the location when searching for the best placement location. and the distance from the cargo to the center of the expected placement area Combination construction heuristic function:

[0064] ;

[0065] is the weight coefficient, , by adjusting The value of can control the relative importance of shape matching and distance factors in the heuristic function. When searching for the best placement position, the smaller the heuristic function value, the better the position. It can quickly and effectively find the optimal solution among many possible positions, improve the efficiency and accuracy of the loading algorithm, and achieve efficient and reasonable loading of goods.

[0066] At the same time, based on the delivery sequence information of goods at different destinations, combined with the logistics network topology and historical delivery time data, the path planning algorithm is used to group the goods, place the goods with the same destination and similar delivery time together, and arrange the goods in sequence according to the expected unloading order.

[0067] When the loading algorithm unit 3 uses the path planning algorithm to group the goods, the matching degree between the shape of the goods and the unoccupied space in the heuristic function is used to optimize the placement order within the group of goods at the same destination, as follows:

[0068] Analyze the fluctuation of historical delivery time data. Suppose the historical delivery time data is , calculate its mean and standard deviation, set the time similarity threshold according to the actual situation, and accurately screen out goods with similar delivery times, providing a reliable basis for subsequent reasonable grouping and placement order optimization.

[0069] For goods and goods , respectively calculate the comprehensive shape matching degree between it and the current unoccupied space and , set cargo The shape data is , the unoccupied space information is , calculate the cargo volume The volume of unoccupied space ; Get the volume matching degree ; Then calculate the shape contour matching degree through the shape similarity measurement method , and then get the comprehensive shape matching degree , is the weight coefficient, and the comprehensive shape matching degree of the goods is calculated in the same way , which provides a more scientific reference for determining the placement order, improves the rationality of the layout of goods in the means of transport, and is conducive to making full use of space and ensuring the stability of goods during transportation.

[0070] Compare and ,like > , then in the expected unloading sequence, the cargo Place it in the priority unloading position, which is determined by calculating its distance from the expected unloading channel. ,like If the preset shortest distance standard is met, the goods are confirmed In order to prioritize the unloading location, the above process is repeated for subsequent goods, and the placement order within the cargo group with the same destination is continuously optimized so that the goods can be quickly found and removed during unloading, which greatly improves the unloading efficiency and reduces the operating time of the logistics node.

[0071] The loading algorithm unit 3 combines the logistics network topology structure to improve the estimated placement area determination method in the heuristic function, as follows:

[0072] Collect historical data, including the traffic flow, speed and other information of each logistics node in different time periods in the past. In time period The traffic volume in , the passing speed is , combined with real-time traffic information, it updates the traffic conditions of logistics nodes in real time, calculates the congestion probability of each node in the logistics network, comprehensively and accurately reflects the actual traffic conditions of logistics nodes, and provides a reliable basis for subsequent adjustments to the cargo placement area.

[0073] Determine the threshold of high congestion probability nodes , filter out the congestion probability greater than The nodes are regarded as the set of nodes with high congestion probability , so that subsequent adjustments to cargo placement for special situations are more targeted, and focus on optimizing the placement of related cargo at key nodes that may affect transportation efficiency.

[0074] For goods passing through nodes with high congestion probability, set the offset coefficient ,in ; is the set of logistics nodes that the cargo transportation path passes through, Representing a collection The number of elements in the offset coefficient can reasonably determine the offset degree of cargo placement according to the comprehensive situation of congestion, so that the placement of cargo in the means of transport is more inclined to avoid risks related to congested paths.

[0075] Assume that the coordinates of the center of gravity of the transport vehicle in the three-dimensional coordinate system are , the original estimated placement area center coordinates are , the adjusted center coordinates of the expected placement area The calculation formula is ; ; ; The center coordinates of the expected placement area are adjusted according to the offset coefficient to achieve dynamic optimization of the cargo placement location. Flexible adjustments are made according to the actual situation of the logistics network to facilitate the reasonable distribution of cargo within the means of transport and reduce the risk of transportation delays caused by congestion.

[0076] During the cargo grouping process, the loading algorithm unit 3 uses the distance factor in the heuristic function to adjust the priority of cargo grouping, as follows:

[0077] Collect historical delivery time data and set the destination The historical delivery time data is , is the number of data points, calculates its mean and standard deviation, and constructs the delivery time interval , is a coefficient determined according to the actual situation. For the front end of the interval, it is defined as , This method of constructing delivery time intervals based on historical data is a time interval set according to the time efficiency requirements of delivery. It makes full use of past experience to make the setting of time intervals more in line with the actual delivery situation, and can accurately identify goods with urgent delivery time, providing a reliable basis for subsequent targeted priority adjustments.

[0078] Extract the original heuristic function ; For goods whose delivery time is at the front of the interval, increase the weight of the distance factor and set a new dynamic weight coefficient , ; is the adjustment function determined by the delivery time t. hour, The bigger, The larger the value is, the modified heuristic function is:

[0079] ;

[0080] By dynamically adjusting the weight coefficient to modify the heuristic function, the importance of the distance factor can be flexibly changed according to the urgency of the goods delivery time, so that the goods with high delivery time requirements are grouped closer to the unloading area, thereby improving the loading and unloading efficiency of such goods.

[0081] In the process of grouping goods, when calculating the heuristic function value of goods, the goods whose delivery time is at the front of the interval are used Calculation,determine the priority of cargo grouping according to the value of the heuristic function. The smaller the heuristic function value, the higher the priority. In this way, cargo with more urgent delivery time requirements can give priority to distance factors when grouping and placing them, so that they can be closer to the expected unloading area in the means of transport, improve unloading efficiency, and ensure on-time delivery.

[0082] After receiving new cargo data and transportation plan change data, the real-time guidance unit 4 re-runs the loading algorithm and assigns priority calculation marks to newly added cargo or cargo with changed destination to distinguish it from the original cargo data, ensuring that these special cases are handled with priority when the loading algorithm is re-run, effectively avoiding confusion between new data and original data.

[0083] For temporarily added goods, the priority calculation order is determined by calculating the degree of impact on the subsequent unloading order based on their weight, shape and expected addition time. A comprehensive and in-depth assessment is made of the potential interference of the temporarily added goods on the entire logistics process, and their priority calculation order is accurately determined to avoid unreasonable arrangements caused by considering a single factor.

[0084] For goods with changed destination information, the priority calculation order is determined according to the position of the new destination in the logistics network topology and the estimated delivery time. The priority calculation order is based on the destination location and delivery time, and comprehensive consideration is given to the overall layout of the logistics network and delivery timeliness. This makes the re-planning of goods with changed destinations more scientific and reasonable, optimizes logistics transportation routes, and reduces the extension of transportation time and increase in costs caused by destination changes.

[0085] In the present invention, the data acquisition unit obtains the size, weight and shape data of the goods, the storage unit stores the data of the goods and the means of transport, the loading algorithm unit divides the three-dimensional grid units according to the spatial data of the means of transport, the cluster analysis algorithm is used to classify the goods by weight, the shape matching and heuristic search algorithms are used to determine the best position of a single cargo, the path planning algorithm is used to group and arrange the cargo in combination with the information such as the distribution sequence, the real-time guidance unit receives new data during loading, re-runs the algorithm and generates guidance information, the system effectively solves the problem of loading space utilization and improves space utilization.

[0086] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. Cross-border e-commerce logistics big data optimization system, characterized by: It comprises a data acquisition unit (1), a data storage unit (2), a loading algorithm unit (3) and a real-time guidance unit (4); The data acquisition unit (1) comprises a laser scanning device, a weighing sensor and an image analysis module, wherein the laser scanning device is used to acquire the size of the goods, the weighing sensor is used to acquire the weight of the goods, and the image analysis module is used to acquire the shape data of the goods; The data storage unit (2) is used to store cargo data and internal space data of transportation vehicles, wherein the transportation vehicles include containers, cargo planes and trucks, and the internal space data includes three-dimensional dimensions, load weight limit and internal fixed structure position information; The loading algorithm unit (3) divides the internal space data of the transport vehicle into three-dimensional grid units, each unit having coordinate information, accommodating weight and volume attributes, and initializes a cargo list and an occupied space list. According to the carrying weight limit of the transport vehicle, a cluster analysis algorithm is used to classify the cargo by weight to form cargo groups of different weight levels. For each cargo, the shape data is matched with the shape of the three-dimensional grid unit of the unoccupied space in the transport vehicle, and a heuristic search algorithm is used to find the best placement position. At the same time, according to the delivery sequence information of cargoes at different destinations, combined with the logistics network topology and historical delivery time data, a path planning algorithm is used to group the cargoes, and cargoes with the same destination and similar delivery time are placed together, and the cargoes are arranged in sequence according to the expected unloading order. The loading algorithm unit (3) improves the method for determining the expected placement area in the heuristic function in combination with the logistics network topology, as follows: Analyze the congestion probability P of each node in the logistics network based on historical data and real-time traffic information yd For cargo passing through nodes with high congestion probability, when determining the expected placement area, it is offset to the center of gravity position in the transport vehicle, as follows: Assume the offset coefficient is Where k is a constant determined by the type of transport tool and the characteristics of the goods, jd is the set of logistics nodes that the goods transportation path passes through, and the center coordinates of the expected placement area are adjusted according to the β value; The loading algorithm unit (3) adjusts the priority of cargo grouping by using the distance factor in the heuristic function during cargo grouping, as follows: The heuristic function combines the matching degree between the shape of the goods and the unoccupied space and the distance between the goods and the center of the expected placement area; Construct the delivery time interval [t min , t max ], for goods whose delivery time is at the front end of the interval, the front end of the interval is t<t mid , Increase the weight of the distance factor in the heuristic function and set the dynamic weight coefficient And modify the heuristic function to Adjust the priority of cargo grouping, where d represents the distance from the cargo to the center of the expected placement area, and M represents the comprehensive shape matching degree; During the loading process, the real-time guidance unit (4) continuously receives new cargo data and transportation plan change data, re-runs the loading algorithm, dynamically adjusts the loading plan, and generates cargo loading guidance information based on the algorithm calculation results to guide the loading of cargo in the transportation vehicle.

2. The cross-border e-commerce logistics big data optimization system according to claim 1 is characterized by: The loading algorithm unit (3) comprises a space division module (31). When dividing the three-dimensional grid units, the space division module (31) uses a polygonal grid division method to process the irregular space inside the transportation tool, as follows: Acquire the three-dimensional dimensions and internal fixed structure position information in the internal space data of the transportation tool from the data storage unit (2), and use a three-dimensional model-based analysis algorithm to identify the internal space contour of the transportation tool; Selecting basic polygons according to the shape characteristics of the irregular space, wherein the shape characteristics are bevel space and curved surface space, and setting the area range of each polygon mesh and the size of the inner angle of the polygon, and finally determining a mesh generation algorithm according to the selected polygon; A three-dimensional coordinate is assigned to each polygon mesh vertex, and the accommodating weight and volume of each polygon mesh are proportionally allocated according to the carrying weight limit of the transportation vehicle and the volume information of the irregular space.

3. The cross-border e-commerce logistics big data optimization system according to claim 1 is characterized by: When the loading algorithm unit (3) divides the transport tool space into three-dimensional grid units, the side length of the grid unit is determined according to the minimum spatial dimension of the transport tool and a preset precision coefficient, as follows: The internal space data of the transportation vehicle is obtained from the data storage unit (2), and the values ​​of the three dimensions of length, width and height are compared to find the minimum value, which is defined as the minimum space dimension. The preset accuracy coefficient is determined according to the accuracy requirements of logistics loading and the size of the goods, and finally the minimum space dimension is multiplied by the preset accuracy coefficient to obtain the side length of the grid unit.

4. The cross-border e-commerce logistics big data optimization system according to claim 1 is characterized by: The loading algorithm unit (3) comprises a weight classification module (32), wherein the clustering analysis algorithm in the weight classification module (32) is a K-means algorithm, and the number of clusters is dynamically determined according to the weight limit of the transport vehicle and the weight distribution characteristics of the cargo, as follows: Extracting the weight limit information of the transport vehicle from the data storage unit (2), and preliminarily dividing the number of theoretical weight intervals according to the weight limit of the transport vehicle and the granularity of cargo weight classification in logistics operations, wherein the granularity of cargo weight classification is used to determine the size and accuracy of the weight interval division; Obtaining a cargo weight data list from the data collection unit (1), performing statistical analysis on the cargo weight data, determining the weight distribution range and mean, and using data clustering to identify cargo weight clustering areas based on the statistical results; Comprehensively analyze the number of theoretical weight intervals under the weight limit of the transport vehicle and the clustering areas in the cargo weight distribution characteristics. If the mean value of the cargo weight distribution is less than or equal to the first threshold, it indicates that the cargo weight distribution is concentrated and is judged to be beyond the theoretical interval division range. At this time, the number of clusters is adjusted mainly based on the cargo weight clustering area. The first threshold is a specific value determined based on the actual logistics and transportation situation and statistical analysis.

5. The cross-border e-commerce logistics big data optimization system according to claim 1 is characterized by: The loading algorithm unit (3) includes a matching placement module (33). The heuristic search algorithm in the matching placement module (33) introduces a heuristic function to search for the best position, which is as follows: The heuristic function combines the matching degree between the shape of the goods and the unoccupied space and the distance between the goods and the center of the expected placement area to search for the best placement position, as follows: For cargo, shape data acquired by the data acquisition unit (1) is used to extract shape features. For unoccupied space, a three-dimensional network unit is used to represent it according to the division of transportation tools, and each unit has its shape and size information. Calculate the volume of the cargo and the volume of the unoccupied space, and calculate the volume matching degree based on the volume of the cargo and the volume of the unoccupied space. Then use the shape similarity measurement method to calculate the distance between the shape contour of the cargo and the shape contour of the unoccupied space, and then convert it into matching degree through a function. Finally, the weighted sum of the volume matching degree and the shape contour matching degree is obtained to obtain the comprehensive shape matching degree. According to the loading strategy of the transport vehicle, the delivery order and weight of the goods, the expected placement area of ​​the goods is preliminarily determined, and the distance from the goods to the center of the expected placement area is calculated in the three-dimensional coordinate system of the transport vehicle. Finally, the matching degree of the shape of the goods with the unoccupied space and the distance from the goods to the center of the expected placement area are combined to construct the heuristic function.

6. The cross-border e-commerce logistics big data optimization system according to claim 5 is characterized by: The loading algorithm unit (3) further comprises a sequence optimization module (34). When the path planning algorithm is used to group the goods, the sequence optimization module (34) uses the matching degree between the shape of the goods and the unoccupied space in the heuristic function to optimize the placement order within the group of goods at the same destination, as follows: For goods with similar delivery time and the same destination, first determine the time proximity threshold based on the fluctuation range of historical delivery time data. If goods A and goods B are within this threshold, calculate the matching degree M of goods A and the current unoccupied space. A and the matching degree M between the cargo B and the current unoccupied space B , if M A >M B , then in the expected unloading sequence, cargo A is placed at the priority unloading position, and the distance between the priority unloading position and the expected unloading channel in the transport vehicle meets the preset shortest distance standard.

7. The cross-border e-commerce logistics big data optimization system according to claim 1 is characterized by: After receiving new cargo data and transportation plan change data, the real-time guidance unit (4) re-runs the loading algorithm, assigns priority calculation marks to newly added cargo or cargo with changed destinations, and determines the priority calculation order for temporarily added cargo based on its weight, shape and estimated joining time by calculating the degree of its influence on the subsequent unloading order. For cargo with changed destination information, the priority calculation order is determined according to the position of the new destination in the logistics network topology and the estimated delivery time.

Images