A multi-objective logistics path optimization method and system for intelligent logistics

Abstract

This invention relates to the field of path optimization technology, specifically to a multi-objective logistics path optimization method and system for intelligent logistics. First, based on the relatively fixed transportation paths near forward transit stations in logistics transportation, the probability of fixed nodes, initially representing node priority, is determined by combining the frequency of node occurrence. Then, the number of substitutable nodes, representing substitutability, is determined based on the probability of fixed nodes, spatial distance, and similarity in sequence distribution. Further, combining the distribution of the number of substitutable nodes and the probability of fixed nodes in the current and historical periods, a priority index for each transportation path node is comprehensively represented. Thus, the pheromone concentration of the local path segment corresponding to each transportation path node is adjusted in real time based on the priority index to obtain a more accurate adjusted pheromone concentration, resulting in better logistics path optimization based on the adjusted pheromone concentration.

Description

Technical Field

[0001] This invention relates to the field of path optimization technology, specifically to a multi-objective logistics path optimization method and system for intelligent logistics. Background Technology

[0002] When optimizing multi-objective logistics routes in smart logistics, existing technologies typically measure a fixed pheromone concentration for each local route segment in the ant colony algorithm, based on the historical transportation costs of that segment within the actual logistics network. However, a fixed pheromone concentration cannot adapt to environmental changes. For example, transportation costs fluctuate over time, reducing the reference value of historical transportation costs and thus decreasing the accuracy of the fixed pheromone concentration measured by existing technologies based on historical transportation costs.

[0003] Furthermore, the method of measuring pheromone concentration solely based on prioritization determined by historical transportation costs fails to consider the logistics attributes of the logistics network itself. For example, transportation routes between large logistics transit stations are usually relatively fixed, while flexible last-mile delivery typically has a higher rate of change. Prioritization based on historical transportation costs would disrupt the logistics attributes of the logistics network itself, leading to a decrease in the accuracy of the determined pheromone concentration.

[0004] In summary, the accuracy of existing technologies in measuring fixed-size pheromone concentrations based on historical transportation costs is low, resulting in poor performance in optimizing logistics routes based on pheromone concentrations. Summary of the Invention

[0005] To address the low accuracy of existing technologies that measure fixed-size pheromone concentrations based on historical transportation costs, this application aims to provide a multi-objective logistics path optimization method and system for intelligent logistics. The specific technical solution adopted is as follows:

[0006] The first aspect of this application provides a multi-objective logistics path optimization method for intelligent logistics, including:

[0007] Obtain the sequence of transportation path nodes for each transportation order in each transportation cycle of historical data;

[0008] In each transportation cycle, the probability of a fixed node is determined based on the distribution of each transportation path node in the sequence of transportation path nodes, the position of its index value, and its frequency of occurrence. In historical data, the number of replaceable nodes is determined based on the overall similarity of the probability of fixed nodes between each transportation path node and other transportation path nodes, the similarity of their sequence distribution, and their relative distance.

[0009] Based on the distribution of fixed node probabilities for each transportation route node across all transportation cycles and the number of alternative nodes, the corresponding priority influence weight is determined; based on the overall size of fixed node probabilities for each transportation route node and the priority influence weight, the corresponding priority index is determined.

[0010] The pheromone concentration of the path in the ant colony algorithm is adjusted using the priority index as a weight to determine the corrected pheromone concentration; the logistics path is then optimized based on the corrected pheromone concentration.

[0011] Furthermore, the process of obtaining the probability of the fixed node includes:

[0012] Each transportation path node is sequentially designated as the target node;

[0013] In each transportation cycle, the sequence of transportation path nodes that contain the target node is used as the reference sequence for the target node; the number of times the reference sequence of the target node appears in each transportation cycle is used as the corresponding transportation frequency.

[0014] The mean index value of the target node in all reference sequences is negatively correlated to determine the priority of the corresponding path.

[0015] The corresponding transport chain weight is determined based on the concentrated distribution of the target node in all transport path node sequences;

[0016] The negative correlation mapping value of the number of transportation orders in each transportation cycle, the transportation frequency, the path priority and the transportation chain weight are normalized to determine the probability of the target node being a fixed node in each transportation cycle.

[0017] Furthermore, the process of obtaining the transport chain weight includes:

[0018] The frequency of occurrence of the reference sequence in each type of transportation order is used as the reference frequency; the number of transportation path nodes in the transportation path node sequence corresponding to the type of transportation order with the highest reference frequency is used as the transportation chain weight of the target node; where the transportation path node sequences of the same type of transportation order are the same.

[0019] Furthermore, the process of obtaining the number of replaceable nodes includes:

[0020] The corresponding fixed node characteristic value is determined based on the average probability of fixed nodes for each transportation path node across all transportation cycles.

[0021] Other transportation path nodes besides the target node are used as corresponding comparison nodes; the corresponding comparison feature deviation value is determined based on the difference between the fixed node feature value of the target node and the fixed node feature value of each corresponding comparison node.

[0022] The degree of functional similarity is determined based on the local transportation sequence similarity between the target node and each comparison node in all transportation path node sequences.

[0023] The node distance weight is determined based on the negative correlation mapping value between the target node and each corresponding comparison node.

[0024] The negative correlation mapping value of the comparison feature deviation value, the product of the node distance weight and the functional similarity are normalized to determine the substitutability of each comparison node for the target node.

[0025] The number of replaceable nodes for the target node is determined based on the total number of comparison nodes whose substitutability exceeds a preset replacement threshold.

[0026] Furthermore, the process of obtaining the degree of functional similarity includes:

[0027] In each reference sequence of each transportation path node, the previous transportation path node of each transportation path node is combined with the corresponding next transportation path node to determine the corresponding path node tuple.

[0028] The set of all path node pairs corresponding to each transportation path node in all reference sequences is taken as the corresponding reference pair set.

[0029] The degree of functional similarity is determined by the number of elements in the intersection between the reference tuple set of the target node and the reference tuple set of each corresponding comparison node.

[0030] Furthermore, the process of obtaining the priority influence weight includes:

[0031] In each transportation cycle, the transportation path nodes whose probability of being a fixed node is greater than a preset probability threshold are designated as the fixed nodes of each transportation cycle.

[0032] The positive correlation mapping value of the probability of fixed nodes in the current transportation cycle of each transportation path node is used as the real-time influence weight.

[0033] The historical influence weight of each transportation path node is determined by multiplying the negative correlation mapping value of the number of alternative nodes with the total number of times each transportation path node belongs to a fixed period node in all transportation cycles.

[0034] The sum of the historical influence weights and the real-time influence weights is normalized to determine the corresponding priority influence weights.

[0035] Furthermore, the process of obtaining the priority index includes:

[0036] The corresponding priority index is determined by performing a positive correlation mapping between the fixed node feature value of each transportation path node and the priority influence weight.

[0037] Furthermore, the process of obtaining the corrected pheromone concentration includes:

[0038] Obtain all local path segments; the start and end points of each local path segment are transportation path nodes, and there are no transportation path nodes in each local path segment other than the start and end points;

[0039] The priority index of the transportation path node corresponding to the end point of each local path segment is used as a reference index.

[0040] The initial pheromone concentration of each local path segment is determined based on the ant colony algorithm; wherein, the initial pheromone concentration of all local path segments is the same; the corrected pheromone concentration of each local path segment is determined by multiplying the initial pheromone concentration of each local path segment with the corresponding reference index.

[0041] Furthermore, the process of obtaining the logistics route optimization based on the modified pheromone concentration includes:

[0042] The origin and destination of the current transportation order are input into the ant colony algorithm. Based on the corrected pheromone concentration of all local path segments, the optimized transportation path of the current transportation order is output.

[0043] Secondly, this application provides a multi-objective logistics path optimization system for intelligent logistics, the system comprising:

[0044] The data acquisition and preprocessing module is used to obtain the sequence of transportation path nodes for each transportation order in each transportation cycle of historical data;

[0045] The first determining module is used to determine the probability of a fixed node in each transportation cycle based on the distribution of each transportation path node in the sequence of transportation path nodes, the position of its index value, and its frequency of occurrence; and to determine the number of corresponding replaceable nodes in historical data based on the overall similarity of the probability of fixed nodes between each transportation path node and other transportation path nodes, the similarity of their sequence distribution, and their relative distance.

[0046] The second determining module is used to determine the corresponding priority influence weight based on the distribution of fixed node probability for each transportation path node in all transportation cycles and the number of alternative nodes; and to determine the corresponding priority index based on the overall size of fixed node probability for each transportation path node and the priority influence weight.

[0047] The logistics route optimization module is used to adjust the pheromone concentration of the path in the ant colony algorithm with the priority index as the weight, determine the corrected pheromone concentration, and optimize the logistics route according to the corrected pheromone concentration.

[0048] Thirdly, this application provides a computer device including a memory and a processor. The memory is used to store computer program code, and the processor is used to call and run the computer program code from the memory to perform the method as described in the first aspect of this application or any embodiment of the first aspect.

[0049] Fourthly, this application provides a computer program product comprising computer program code, which, when executed, performs the method as described in the first aspect of this application or any embodiment thereof.

[0050] Fifthly, this application provides a computer-readable storage medium that stores computer program code, which, when executed, performs the method as described in the first aspect of this application or any embodiment thereof.

[0051] This application has the following beneficial effects:

[0052] This application first determines the probability of fixed nodes that preliminarily represent node priority based on the relatively fixed transportation routes of transit stations near the front end in logistics transportation, combined with the frequency of node occurrence. Then, it determines the number of substitutable nodes that represent substitutability based on the probability of fixed nodes, spatial distance, and similarity of sequence distribution. Furthermore, it combines the distribution of the number of substitutable nodes and the probability of fixed nodes in the current period and historical periods to comprehensively represent the priority index of each transportation route node. Thus, the pheromone concentration of the local path segment corresponding to each transportation route node is adjusted in real time based on the priority index to obtain a more accurate adjusted pheromone concentration, making the logistics route optimization based on the adjusted pheromone concentration more effective. Attached Figure Description

[0053] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0054] Figure 1 A flowchart of a multi-objective logistics path optimization method for intelligent logistics provided in one embodiment of the present invention;

[0055] Figure 2 This is a structural diagram of a multi-objective logistics path optimization system for intelligent logistics provided in one embodiment of the present invention;

[0056] Figure 3 This is a schematic diagram of a computer device structure provided in one embodiment of the present invention. Detailed Implementation

[0057] To further illustrate the technical means and effects adopted by the present invention to achieve its intended purpose, the following, in conjunction with the accompanying drawings and preferred embodiments, details the specific implementation, structure, features, and effects of a multi-objective logistics path optimization method and system for intelligent logistics proposed according to the present invention. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment, and specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form. Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined with "first" or "second" may explicitly or implicitly include one or more of that feature.

[0058] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains.

[0059] The following description, in conjunction with the accompanying drawings, details the specific scheme of the multi-objective logistics path optimization method and system for intelligent logistics provided by this invention.

[0060] This application provides a multi-objective logistics path optimization method for intelligent logistics. Please refer to [link / reference]. Figure 1 The diagram illustrates a flowchart of a multi-objective logistics path optimization method for intelligent logistics according to an embodiment of the present invention. The method includes:

[0061] Step S101: Obtain the sequence of transportation path nodes for each transportation order in each transportation cycle of historical data.

[0062] First, in the logistics system analyzed in this application, the logistics system network nodes are obtained and analyzed as transportation path nodes; where, a logistics system network node refers to the junction connecting logistics lines in the logistics system network, which is a technical term well known to those skilled in the art, and will not be further described here.

[0063] In one specific implementation of this invention, the transportation path node sequence for each transportation order within each day of the three months prior to the current time is obtained; that is, the time range of the historical data in this embodiment is three months prior to the current time, and the transportation cycle length is set to one day; wherein, the transportation path node sequence is a sequence composed of all transportation path nodes traversed on the transportation path of each transportation order arranged in chronological order. It should be noted that the time range of the historical data and the length of the transportation cycle can be adjusted according to the specific implementation environment, and this embodiment is applicable to implementation environments with more than 3 nodes, which will not be further elaborated here.

[0064] Step S102: In each transportation cycle, determine the probability of a fixed node based on the distribution of each transportation path node in the sequence of transportation path nodes, the position of its index value, and its frequency of occurrence; in historical data, determine the number of replaceable nodes based on the overall similarity of the probability of fixed nodes between each transportation path node and other transportation path nodes, the similarity of their sequence distribution, and their relative distance.

[0065] In actual transportation, the logistics process always starts from a transportation origin, such as the shipper delivering the goods to a large transit node. At the large transit node, the goods are sorted and further processed before being transferred to medium-sized nodes, and finally, through flexible last-mile delivery. During transportation, the order and path of goods passing through large nodes are relatively consistent, resulting in significant change costs. However, in flexible last-mile delivery, the delivery address of the order's destination is usually more detailed, offering greater flexibility but also being significantly affected by real-time road conditions. Therefore, in each transportation cycle, the probability of a fixed node can be determined based on the distribution of each transportation path node in the sequence of transportation path nodes, its index value, and its frequency of occurrence. A higher probability of a fixed node indicates that the corresponding transportation path node is more likely to be a large fixed transportation node, and thus, the higher the priority of the corresponding transportation path node.

[0066] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the probability of fixed nodes includes:

[0067] Each transportation path node is sequentially designated as a target node. Within each transportation cycle, the sequence of transportation path nodes containing target nodes is designated as the reference sequence for the target nodes. The frequency of occurrence of the reference sequence for the target nodes within each transportation cycle is designated as the corresponding transportation frequency. For a target node, a higher frequency of occurrence within the same transportation cycle indicates a larger transportation volume within that cycle, thus increasing the likelihood of it being a fixed transportation node and consequently assigning it a higher priority.

[0068] By negatively correlating the mean index values ​​of the target node across all reference sequences, the path's leading position is determined. The earlier the target node appears in each reference sequence, the closer it is to the front end of the logistics transportation network. Since front-end transportation path nodes in a logistics transportation network are relatively fixed, a higher path leading position indicates a greater likelihood that the corresponding target node belongs to a fixed transportation node.

[0069] Based on the concentrated distribution of the target node in all transportation path node sequences, the corresponding transportation chain weight is determined. The process of obtaining the transportation chain weight includes: taking the frequency of occurrence of the reference sequence in each type of transportation order as the reference frequency; taking the number of transportation path nodes in the transportation path node sequence corresponding to the type of transportation order with the highest reference frequency as the transportation chain weight of the target node; wherein, the transportation path node sequences of the same type of transportation order are the same.

[0070] For a target node, the sequence of transport path nodes corresponding to the transport order type with the highest reference frequency represents the main logistics transport path characteristics of the target node. When the sequence of transport path nodes corresponding to the transport order type with the highest reference frequency of the target node is short, even if the overall index value of the target node is small, the short overall node sequence makes the target node's position in the node relatively less prominent. This reduces the reliability of the path prominence calculated solely based on the index value. Therefore, it is necessary to further weight the path prominence using the calculated transport chain weights.

[0071] Furthermore, considering the varying number of transport orders across different transport cycles, to reduce the impact of the total number of transport orders in a transport cycle on the calculation process, the number of transport orders is further analyzed during the calculation of the fixed node probability. By further combining correlations, the negative correlation mapping value of the number of transport orders in each transport cycle, the transport frequency, the path's leading position, and the transport chain weight are normalized to determine the fixed node probability of the target node in each transport cycle. Normalization limits the fixed node probability to between 0 and 1, which not only conforms to the value range represented by probability but also makes the numerical value of the fixed node probability more intuitive, facilitating the selection of fixed nodes in subsequent cycles.

[0072] In one specific implementation of this invention, the process of obtaining the probability of a fixed node is expressed by the following formula: Among them, H k,h M represents the probability of node k being a fixed node in the h-th transportation cycle of the transportation path; hLet F be the total number of transport path nodes in the h-th transport cycle. Since transport cycles with a larger total number of transport path nodes usually correspond to a higher transport frequency than transport cycles with a smaller total number of transport path nodes, to reduce the impact of limitations caused by different scenarios on the robustness of the calculation process, the negative correlation mapping value of the total number of transport path nodes is used in the calculation of the probability of fixed nodes; k,h S represents the transport frequency of node k in the h-th transport cycle. k,h It is the number of transportation path nodes in the transportation path node sequence corresponding to the transportation order type with the highest reference frequency in the h-th transportation cycle, which is also the transportation chain weight of transportation path node k. is the mean of the index values ​​of transport path node k in all reference sequences during the h-th transport cycle; exp() is an exponential function with the natural constant as the base, and other negative correlation mapping methods can be used, such as 1-Norm(); This represents the path advancement of node k in the h-th transportation cycle; Norm() is a linear normalization function, and the normalization method can be adjusted as needed.

[0073] In other implementations of this invention, methods other than multiplication can be used to determine the probability of fixed nodes. For example, each parameter can be normalized, and the mean of all parameters can be calculated to determine the probability of fixed nodes. Further details will not be provided here. It should be noted that, unless otherwise specified, the normalization method used in this invention is linear normalization, which will not be further elaborated upon hereafter.

[0074] The probability of a fixed node characterizes the likelihood that each transportation path node will be a high-priority fixed selection node in each transportation cycle. Therefore, the higher the probability of a fixed node, the higher the selection priority of that transportation path node. Furthermore, it is necessary to consider whether a transportation path node can be replaced. For each path node, the stronger its irreplaceability, the more likely it is to be a fixed node, and the higher its corresponding selection priority should be. In a logistics system network, the higher the similarity between two transportation path nodes, the stronger their substitutability should be. Therefore, this embodiment of the invention analyzes the number of substitutable nodes for each transportation path node in three dimensions: fixed node probability, sequence distribution, and spatial dimension.

[0075] Preferably, in some possible implementations of the embodiments of the present invention, the alternative process for obtaining the number of nodes includes:

[0076] Based on the average probability of fixed nodes for each transportation path node across all transportation cycles, the corresponding fixed node characteristic value is determined. Other transportation path nodes besides the target node are used as corresponding comparison nodes. The comparison feature deviation value is determined based on the difference between the fixed node characteristic value of the target node and the fixed node characteristic value of each comparison node. The fixed node characteristic value characterizes the probability of fixed nodes across all transportation cycles. The smaller the deviation of the fixed node characteristic value between the target node and each comparison node, the more similar the priorities are between the target node and the corresponding comparison node. Therefore, for the target node, in terms of priority, the comparison node is more likely to replace the target node. Thus, the smaller the comparison feature deviation value, the higher the substitutability of the corresponding comparison node for the target node.

[0077] The functional similarity is determined based on the local transportation sequence similarity between the target node and each comparison node across all transportation path node sequences. The process of obtaining the functional similarity includes: in each reference sequence of each transportation path node, combining the previous transportation path node with the corresponding next transportation path node to determine the corresponding path node tuple; taking the set of all path node tuples corresponding to each transportation path node across all reference sequences as the corresponding reference tuple set; and determining the functional similarity based on the number of elements in the intersection between the reference tuple set of the target node and the reference tuple set of each comparison node.

[0078] Considering that for any two transport path nodes, if their previous and next transport path nodes are the same across most transport path sequences or transport path node sequences, it indicates that the two transport path nodes can accept the same superior and subordinate nodes. The more similar their functions are in the logistics transport network, the more they can substitute for each other. Therefore, for the target node and each comparison node, the more elements there are in the intersection between the reference set of the target node and the reference set of each comparison node, i.e., the greater the degree of functional similarity, the more similar the functions of the comparison node and the target node are, and the higher the substitutability of the corresponding comparison node for the target node should be.

[0079] The node distance weights are determined based on the negative correlation mapping values ​​between the target node and each corresponding comparison node. Spatially, the smaller the node distance weight (i.e., the smaller the distance between the target node and each comparison node), the lower the spatial distance cost when the target node is replaced by its corresponding comparison node, and therefore the higher the substitutability of the comparison node for the target node. It should be noted that the distances between the target node and each comparison node use Euclidean distance, which can be adjusted according to the specific implementation environment, such as the shortest path distance between the target node and each comparison node; this will not be elaborated further here.

[0080] Finally, based on the correlation between the comparison feature deviation value, node distance weight, and functional similarity and the substitutability, the negative correlation mapping value of the comparison feature deviation value and the product of node distance weight and functional similarity are normalized to determine the substitutability of each comparison node to the target node; so that the greater the corresponding substitutability, the more likely the comparison node is to be a substitutable node of the target node.

[0081] In one specific implementation of this invention, the process of obtaining the degree of substitutability is expressed by the formula: W k,r Let $r$ be the substitutability of the $r$-th comparison node for the $k$-th node in the transportation path. It is the mean of the fixed node probabilities of transportation path node k in all transportation cycles, that is, the fixed node characteristic value corresponding to transportation path node k. It is the mean of the fixed node probability of the r-th comparison node corresponding to the transportation path node k in all transportation cycles, that is, the fixed node feature value of the r-th comparison node corresponding to the transportation path node k. X represents the comparison feature deviation value between transportation path node k and its corresponding r-th comparison node; k,r D represents the number of elements in the intersection of the reference tuple set of transportation path node k and the reference tuple set of the corresponding r-th comparison node, which is also the corresponding functional similarity. k,r Let k be the distance between node k in the transportation path and the corresponding r-th comparison node; Norm() is a linear normalization function, used here to normalize the distance and reduce the influence of dimensions on the calculation process; (1-Norm(D k,r )) represents the node distance weight between node k in the transportation path and the corresponding r-th comparison node.

[0082] Finally, by setting a threshold, the number of replaceable nodes for the target node is determined based on the total number of comparison nodes whose substitutability exceeds a preset replacement threshold. Considering that the substitutability is a normalized value, the preset replacement threshold needs to be limited to between 0 and 1. In a specific implementation of this invention, the preset replacement threshold is set to 0.7, which can be adjusted according to the specific implementation environment, and will not be elaborated further here. The more replaceable nodes there are, the more choices the target node has when replacing nodes, and the stronger the substitutability of the target node. Therefore, the probability of a node being a fixed choice should be smaller, that is, the priority of the node should be lower.

[0083] Step S103: Determine the corresponding priority influence weight based on the distribution of fixed node probabilities and the number of alternative nodes for each transportation path node throughout all transportation cycles; determine the corresponding priority index based on the overall size of fixed node probabilities and priority influence weight for each transportation path node.

[0084] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the priority influence weight includes:

[0085] In each transportation cycle, transportation path nodes with a probability greater than a preset probability threshold are designated as the fixed nodes for that cycle. In one specific implementation of this invention, the preset probability threshold is set to 0.7, which can be adjusted according to the specific implementation environment.

[0086] The positive correlation mapping value of the probability of each transportation route node being a fixed node in the current transportation cycle is used as the real-time impact weight. The timeliness weight corresponding to the current transportation cycle is greater than the timeliness weight of other transportation cycles. Therefore, if the probability of a transportation route node being a fixed node in the transportation cycle is higher, then the corresponding transportation route node should have a higher priority at the current moment.

[0087] The historical influence weight of each transportation route node is determined by multiplying the negative correlation mapping value of the number of replaceable nodes with the total number of times each transportation route node belongs to a fixed node in all transportation cycles. The greater the total number of times a transportation route node belongs to a fixed node in all transportation cycles, the more it conforms to the characteristics of a fixed node in historical data, and its corresponding priority should be higher. Conversely, the fewer replaceable nodes there are, the less replaceable the corresponding transportation route node is, and it also conforms to the characteristics of a fixed node, and its corresponding priority should be higher. Therefore, the greater the historical influence weight, the higher the priority of the corresponding transportation route node should be.

[0088] Furthermore, by combining the correlation between historical impact weights and real-time impact weights and priorities, the sum of historical impact weights and real-time impact weights is normalized to determine the corresponding priority impact weights, so that the larger the priority impact weight, the higher the priority of the corresponding transportation path node.

[0089] In one specific implementation of this invention, the process of obtaining the priority influence weight is expressed by the formula: Among them, U k The priority of node k in the transportation path is used to influence the weight; H ′ k Let α be the probability of a fixed node in the current transportation cycle of the transportation path node k; α is a positive correlation adjustment parameter used to adjust the weight of the probability of a fixed node in the current transportation cycle of the transportation path node k. In this embodiment, it is set to 5, but can be adjusted according to the specific implementation environment; α×H ′ k Y represents the real-time impact weight of node k in the transportation path; k The number of alternative nodes for node k in the transportation path; The negative correlation mapping value for the number of replaceable nodes is used, and a real number 1 is added to the denominator to avoid the problem of the denominator being 0; Z k This represents the total number of times that transportation path node k belongs to a fixed-period node in all transportation cycles of historical data. The historical influence weight of node k in the transportation path is given; Norm() is the linear normalization function.

[0090] For each transportation path node, the greater its probability of being a fixed node in all transportation cycles, the more it conforms to the characteristics of a fixed node in historical data, and the higher its corresponding priority should be. Therefore, the priority index is determined by comprehensively considering the overall probability of each transportation path node being a fixed node and the weight of priority influence. The higher the priority index, the higher the priority should be in the current path selection.

[0091] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the priority index includes:

[0092] The priority index is determined by performing a positive correlation mapping between the fixed node feature value and the priority influence weight of each transportation path node. In a specific implementation of this invention, the process of obtaining the priority index is expressed by the formula: B k U is the priority index of node k in the transportation path; k The priority of node k in the transportation path is affected by the weight; For fixed node feature values ​​of node k in the transportation path; Norm() is a linear normalization function; by... By performing a positive correlation mapping, the value range of the priority index is limited to 1 to 2, making the corrected pheromone concentration obtained when the initial pheromone concentration is weighted more accurate and intuitive.

[0093] Step S104: Adjust the pheromone concentration of the path in the ant colony algorithm with priority index as weight to determine the corrected pheromone concentration; optimize the logistics path according to the corrected pheromone concentration.

[0094] Since a higher priority index for a transportation path node corresponds to a higher priority for that node in the current path selection, the ant colony algorithm essentially measures the pheromone concentration of local path segments between nodes. However, the path selection process also involves node selection. Therefore, the pheromone concentration of each local path segment can be adjusted based on the node's priority, resulting in better logistics path optimization based on the adjusted pheromone concentration obtained from the ant colony algorithm.

[0095] Preferably, in some possible implementations of the embodiments of the present invention, the process of obtaining the corrected pheromone concentration includes:

[0096] Obtain all local path segments; the start and end points of a local path segment are both transportation path nodes, and there are no other transportation path nodes in the local path segment besides the start and end points; that is, after dividing the entire transportation path with transportation path nodes as the interval points, all local path segments can be obtained.

[0097] The priority index of the transportation path node corresponding to the end point of each local path segment is used as a reference index. First, in the logistics system network, the branching point of a node usually corresponds to the starting point of multiple local path segments. Therefore, selecting the end point of each local path segment is equivalent to selecting the entire path of the local path segment at the same time. Thus, the priority index of the transportation path node corresponding to the end point of each local path segment is used as a reference index.

[0098] Further, the initial pheromone concentration of each local path segment is determined based on the ant colony algorithm; wherein, the initial pheromone concentration of all local path segments is the same; in a specific implementation of this invention, the initial pheromone concentration of all local path segments is set to 1, which can be adjusted according to the specific implementation environment. Then, the product between the initial pheromone concentration of each local path segment and the corresponding reference index is used to determine the corrected pheromone concentration of each local path segment, so that when optimizing the path based on the corrected pheromone concentration, the priority of the transportation path node with the higher priority index is increased, resulting in better path optimization. Therefore, at the current moment, the transportation start and end points of the current transportation order are input into the ant colony algorithm, and the optimized transportation path of the current transportation order is output based on the corrected pheromone concentration of all local path segments, making the optimized transportation path more accurate. It should be noted that the ant colony algorithm is a well-known technique in the art, and pheromone concentration is a technical term in the ant colony algorithm, which will not be further limited or elaborated here.

[0099] In summary, a multi-objective logistics path optimization method for intelligent logistics first determines the probability of fixed nodes that initially represent node priority based on the relatively fixed transportation paths near transit stations in logistics transportation and the frequency of node occurrence. Then, it determines the number of substitutable nodes representing substitutability based on the probability of fixed nodes, spatial distance, and similarity in sequence distribution. Furthermore, it comprehensively represents the priority index of each transportation path node by combining the distribution of the number of substitutable nodes and the probability of fixed nodes in the current and historical periods. Thus, it adjusts the pheromone concentration of the local path segment corresponding to each transportation path node in real time based on the priority index to obtain a more accurate adjusted pheromone concentration, resulting in better logistics path optimization based on the adjusted pheromone concentration.

[0100] This application also provides a multi-objective logistics path optimization system for intelligent logistics. Please refer to [link / reference]. Figure 2 The diagram illustrates a structure of a multi-objective logistics path optimization system for intelligent logistics according to an embodiment of the present invention. The system includes: a data acquisition and preprocessing module 201, a first determination module 202, a second determination module 203, and a logistics path optimization module 204.

[0101] The data acquisition and preprocessing module 201 is used to acquire the sequence of transportation path nodes for each transportation order in each transportation cycle of historical data;

[0102] The first determining module 202 is used to determine the probability of a fixed node in each transportation cycle based on the distribution of each transportation path node in the sequence of transportation path nodes, the position of its index value, and its frequency of occurrence; and to determine the number of replaceable nodes in historical data based on the overall similarity of the probability of fixed nodes between each transportation path node and other transportation path nodes, the similarity of their sequence distribution, and their relative distance.

[0103] The second determining module 203 is used to determine the corresponding priority influence weight based on the distribution of the probability of fixed nodes and the number of alternative nodes for each transportation path node in all transportation cycles; and to determine the corresponding priority index based on the overall size of the probability of fixed nodes and the priority influence weight for each transportation path node.

[0104] The logistics route optimization module 204 is used to adjust the pheromone concentration of the path in the ant colony algorithm with priority index as weight, determine the corrected pheromone concentration, and optimize the logistics route based on the corrected pheromone concentration.

[0105] It should be noted that the system provided in the above embodiments is only an example of the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the computer device can be divided into different functional modules to complete all or part of the functions described above. In addition, the intelligent logistics multi-objective logistics path optimization system and the intelligent logistics multi-objective logistics path optimization method embodiment provided in the above embodiments belong to the same concept, and their specific implementation process can be found in the method embodiment, which will not be repeated here.

[0106] This application also provides a computer device; please refer to [link / reference]. Figure 3 The diagram illustrates a computer device structure according to an embodiment of the present invention. The computer device includes a memory 301, a processor 302, and a computer program 303 stored in the memory 301 and running on the processor 302. When the processor 302 executes the computer program 303, the computer device can execute any of the aforementioned intelligent logistics multi-objective logistics path optimization methods.

[0107] This application also provides a computer program product that, when run on a computer device, enables the computer device to execute any of the aforementioned intelligent logistics multi-objective logistics path optimization methods.

[0108] This application also provides a computer-readable storage medium storing computer program code. When the computer program code is run on a computer device, the computer device can execute any of the aforementioned intelligent logistics multi-objective logistics path optimization methods.

[0109] In the embodiments provided in this application, it should be understood that the computer device, computer program product and computer-readable storage medium provided are all used to perform the corresponding methods provided above, and therefore the beneficial effects they can achieve can be referred to the beneficial effects of the methods provided above, which will not be repeated here.

[0110] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired result. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0111] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

Claims

1. A multi-objective logistics path optimization method for intelligent logistics, characterized in that, The method includes: Obtain the sequence of transportation path nodes for each transportation order in each transportation cycle of historical data; In each transportation cycle, based on the distribution of each transportation path node in the sequence of transportation path nodes, its index value, and its frequency of occurrence, the probability of a corresponding fixed node is determined, including: Each transportation path node is sequentially designated as the target node; In each transportation cycle, the sequence of transportation path nodes that contain the target node is used as the reference sequence for the target node; the number of times the reference sequence of the target node appears in each transportation cycle is used as the corresponding transportation frequency. The mean index value of the target node in all reference sequences is negatively correlated to determine the priority of the corresponding path. The corresponding transport chain weight is determined based on the concentrated distribution of the target node in all transport path node sequences; The negative correlation mapping value of the number of transportation orders in each transportation cycle, the transportation frequency, the product of the path's leading position and the transportation chain weight are normalized to determine the fixed node probability of the target node in each transportation cycle; in historical data, the number of corresponding alternative nodes is determined based on the overall similarity of the fixed node probability of each transportation path node with other transportation path nodes, the similarity of the sequence distribution, and the relative distance, including: determining the corresponding fixed node feature value based on the mean of the fixed node probability of each transportation path node in all transportation cycles; Other transportation path nodes besides the target node are used as corresponding comparison nodes; the corresponding comparison feature deviation value is determined based on the difference between the fixed node feature value of the target node and the fixed node feature value of each corresponding comparison node. Based on the local transportation sequence similarity between the target node and each comparison node across all transportation path node sequences, the corresponding functional similarity is determined, including: In each reference sequence of each transportation path node, the previous transportation path node of each transportation path node is combined with the corresponding next transportation path node to determine the corresponding path node tuple. The set of all path node pairs corresponding to each transportation path node in all reference sequences is taken as the corresponding reference pair set. The degree of functional similarity is determined by the number of elements in the intersection between the reference binary set of the target node and the reference binary set of each corresponding comparison node. The node distance weight is determined based on the negative correlation mapping value between the target node and each corresponding comparison node. The negative correlation mapping value of the comparison feature deviation value, the product of the node distance weight and the functional similarity are normalized to determine the substitutability of each comparison node for the target node. The number of replaceable nodes for the target node is determined based on the total number of comparison nodes whose substitutability exceeds a preset replacement threshold. Based on the distribution of fixed node probabilities for each transportation route node across all transportation cycles and the number of alternative nodes, the corresponding priority influence weight is determined; based on the overall size of fixed node probabilities for each transportation route node and the priority influence weight, the corresponding priority index is determined. The pheromone concentration of the path in the ant colony algorithm is adjusted using the priority index as a weight to determine the corrected pheromone concentration; the logistics path is then optimized based on the corrected pheromone concentration.

2. The multi-objective logistics path optimization method for intelligent logistics according to claim 1, characterized in that, The process of obtaining the transport chain weights includes: The frequency of occurrence of the reference sequence in each type of transportation order is used as the reference frequency; the number of transportation path nodes in the transportation path node sequence corresponding to the type of transportation order with the highest reference frequency is used as the transportation chain weight of the target node; where the transportation path node sequences of the same type of transportation order are the same.

3. The multi-objective logistics path optimization method for intelligent logistics according to claim 1, characterized in that, The process of obtaining the priority influence weight includes: In each transportation cycle, the transportation path nodes whose probability of being a fixed node is greater than a preset probability threshold are designated as the fixed nodes of each transportation cycle. The positive correlation mapping value of the probability of fixed nodes in the current transportation cycle of each transportation path node is used as the real-time influence weight. The historical influence weight of each transportation path node is determined by multiplying the negative correlation mapping value of the number of alternative nodes with the total number of times each transportation path node belongs to a fixed period node in all transportation cycles. The sum of the historical influence weights and the real-time influence weights is normalized to determine the corresponding priority influence weights.

4. The multi-objective logistics path optimization method for intelligent logistics according to claim 1, characterized in that, The process of obtaining the priority index includes: The corresponding priority index is determined by performing a positive correlation mapping between the fixed node feature value of each transportation path node and the priority influence weight.

5. The multi-objective logistics path optimization method for intelligent logistics according to claim 1, characterized in that, The process of obtaining the corrected pheromone concentration includes: Obtain all local path segments; the start and end points of each local path segment are transportation path nodes, and there are no transportation path nodes in each local path segment other than the start and end points; The priority index of the transportation path node corresponding to the end point of each local path segment is used as a reference index. The initial pheromone concentration of each local path segment is determined based on the ant colony algorithm; wherein, the initial pheromone concentration of all local path segments is the same; the corrected pheromone concentration of each local path segment is determined by multiplying the initial pheromone concentration of each local path segment with the corresponding reference index.

6. The multi-objective logistics path optimization method for intelligent logistics according to claim 5, characterized in that, The process of obtaining the logistics route optimization based on the modified pheromone concentration includes: The origin and destination of the current transportation order are input into the ant colony algorithm. Based on the corrected pheromone concentration of all local path segments, the optimized transportation path of the current transportation order is output.

7. A multi-objective logistics path optimization system for intelligent logistics, characterized in that, The system includes: The data acquisition and preprocessing module is used to obtain the sequence of transportation path nodes for each transportation order in each transportation cycle of historical data; The first determining module is used to determine the probability of a fixed node in each transportation cycle based on the distribution of each transportation path node in the sequence of transportation path nodes, the position of its index value, and its frequency of occurrence; and to determine the number of corresponding replaceable nodes in historical data based on the overall similarity of the probability of fixed nodes between each transportation path node and other transportation path nodes, the similarity of their sequence distribution, and their relative distance. The first determining module is specifically used to sequentially identify each transportation path node as the target node. In each transportation cycle, the sequence of transportation path nodes that contain the target node is used as the reference sequence for the target node; the number of times the reference sequence of the target node appears in each transportation cycle is used as the corresponding transportation frequency. The mean index value of the target node in all reference sequences is negatively correlated to determine the priority of the corresponding path. The corresponding transport chain weight is determined based on the concentrated distribution of the target node in all transport path node sequences; The negative correlation mapping value of the number of transportation orders in each transportation cycle, the transportation frequency, the path priority and the transportation chain weight are normalized to determine the probability of the target node being a fixed node in each transportation cycle. The corresponding fixed node characteristic value is determined based on the average probability of fixed nodes for each transportation path node across all transportation cycles. Other transportation path nodes besides the target node are used as corresponding comparison nodes; the corresponding comparison feature deviation value is determined based on the difference between the fixed node feature value of the target node and the fixed node feature value of each corresponding comparison node. Based on the local transportation sequence similarity between the target node and each comparison node across all transportation path node sequences, the corresponding functional similarity is determined, including: In each reference sequence of each transportation path node, the previous transportation path node of each transportation path node is combined with the corresponding next transportation path node to determine the corresponding path node tuple. The set of all path node pairs corresponding to each transportation path node in all reference sequences is taken as the corresponding reference pair set. The degree of functional similarity is determined by the number of elements in the intersection between the reference binary set of the target node and the reference binary set of each corresponding comparison node. The node distance weight is determined based on the negative correlation mapping value between the target node and each corresponding comparison node. The negative correlation mapping value of the comparison feature deviation value, the product of the node distance weight and the functional similarity are normalized to determine the substitutability of each comparison node for the target node. The number of replaceable nodes for the target node is determined based on the total number of comparison nodes whose substitutability exceeds a preset replacement threshold; the second determining module is used to determine the corresponding priority influence weight based on the distribution of fixed node probability for each transportation path node in all transportation cycles and the number of replaceable nodes; and to determine the corresponding priority index based on the overall size of fixed node probability for each transportation path node and the priority influence weight. The logistics route optimization module is used to adjust the pheromone concentration of the path in the ant colony algorithm with the priority index as the weight, determine the corrected pheromone concentration, and optimize the logistics route according to the corrected pheromone concentration.

Images