Intelligent vehicle and goods matching method based on heuristic rule

Abstract

The invention provides an intelligent vehicle-cargo matching method based on a heuristic rule, and the method comprises the following steps: S1, obtaining historical vehicle cargo carrying data, vehicle-cargo matching data, and real-time to-be-carried cargo and to-be-distributed transport capacity data, carrying out the preprocessing of the obtained data, and constructing a data set; s2, mining a frequent item set for the data set by adopting an FP-growth algorithm, and generating a cargo type set suitable for being carried by each transport capacity; and S3, obtaining information of empty vehicle drivers who are reserved through mobile phones to pass the sea, generating a transport capacity pool, constructing a utility function of multi-dimensional comprehensive waybill recommendation according to a cargo type set carried by each transport capacity, and matching cargo type set information with the empty vehicle driver information by adopting a Pareto frontier search algorithm based on improved ant colony optimization, so as to obtain a final cargo type set. Generating a waybill recommendation candidate set; and S4, screening a candidate waybill list conforming to driver transportation in the transport capacity pool according to the vehicle and cargo matching rule, and pushing the candidate waybill list to the driver.

Description

Technical Field

[0001] The present invention relates to the technical field of logistics transportation, and particularly to an intelligent vehicle-cargo matching method based on heuristic rules. Background Art

[0002] In recent years, with the rapid development of e-commerce and cross-border trade, the scale of freight platforms has been continuously expanding, and the demand for vehicle-cargo matching has been increasing day by day. In the field of intelligent logistics and network freight technology, the optimization of vehicle-cargo matching systems is a key link to improve transportation efficiency. However, the existing waybill matching technologies have significant defects in aspects such as dynamic rule adjustment, multi-objective optimization, and the matching efficiency of newly added transport capacity: the static rule base is difficult to adapt to the real-time changes of cargo combinations and environmental parameters, the lack of multi-objective collaborative optimization ability, low cold start efficiency, and insufficient algorithm real-time performance. The existence of these problems seriously restricts the transport capacity matching efficiency in the cross-sea transportation scenario, resulting in the coexistence of a decline in ship turnover rate and cargo backlog.

[0003] Therefore, the present invention specifically proposes a solution combining a dynamic rule engine and multi-objective Pareto optimization. It realizes hourly rule iteration through the FP-growth algorithm updated by a sliding time window, and uses an improved ant colony algorithm to collaboratively optimize multiple objectives such as transportation cost, loading rate, and carbon emissions, so as to improve the transport capacity matching efficiency and optimize cross-sea transportation scheduling. Summary of the Invention

[0004] The purpose of the present invention is to provide an intelligent vehicle-cargo matching method based on heuristic rules to solve the problems mentioned in the above background art.

[0005] The present invention is realized through the following technical solutions:

[0006] An intelligent vehicle-cargo matching method based on heuristic rules, the method comprising the following steps:

[0007] Step S1: Obtain historical vehicle cargo data, vehicle-cargo matching data, as well as real-time to-be-transported cargo and to-be-allocated transport capacity data, and preprocess the obtained data to construct a data set;

[0008] Step S2: Use the FP-growth algorithm on the data set to mine frequent item sets and generate a set of cargo types suitable for each transport capacity to carry;

[0009] Step S3: Obtain the information of empty truck drivers who make appointments to cross the sea through mobile phones and generate a transport capacity pool. According to the set of cargo types suitable for each transport capacity, construct a utility function for multi-dimensional comprehensive waybill recommendation, and use the Pareto front search algorithm based on improved ant colony optimization to match the set information of cargo types with the information of empty truck drivers to generate a candidate set of waybill recommendations;

[0010] Step S4: Filter the candidate waybill list that meets the transportation requirements of the drivers in the capacity pool according to the vehicle-cargo matching rules, and push it to the drivers.

[0011] As a preferred solution of the present invention, the step S1 specifically includes: the preprocessing includes deleting abnormal data and duplicate data, filling in missing data, removing noise, and data standardization.

[0012] As a preferred solution of the present invention, the step S2 specifically includes:

[0013] Step S21: Construct a frequent pattern tree to compress the data set, and recursively mine the conditional pattern bases to gradually generate all frequent item sets;

[0014] Step S22: Set a minimum support threshold for frequent item sets with different numbers of items to filter low-frequency association combinations. As the number of items in the frequent item sets increases, appropriately reduce the threshold to mine more high-order association rules and generate a set of cargo types suitable for each vehicle's capacity to carry.

[0015] As a preferred solution of the present invention, in the step S22, to increase the mining ability for low-frequency association combinations, an analysis of cargo associations based on a co-occurrence matrix is introduced, specifically as follows:

[0016] Step S221: Construct a co-occurrence matrix, and respectively count the individual occurrence times of each cargo type and the co-occurrence times of their pairwise combinations;

[0017] Step S222: Calculate the cosine similarity between cargos based on the statistical data, and use the similarity value as the edge weight of the weighted graph. The nodes correspond to different cargo types to generate a cargo association graph;

[0018] Step S223: Further mine the potential association relationships between cargos through this cargo association graph to optimize the vehicle-cargo matching rules.

[0019] As a preferred solution of the present invention, the step S3 specifically includes:

[0020] Step S31: Based on the physical constraints and environmental constraints of the transportation carrier, generate an initial waybill candidate set that meets the transportation capacity of the current capacity pool;

[0021] Step S32: Construct a multi-dimensional comprehensive utility function based on spatial proximity, full load rate, and cargo type matching degree to quantify the matching degree of capacity-waybill;

[0022] Step S33: Introduce a two-way cardinality constraint mechanism in the improved ant colony optimization-based Pareto front search algorithm to control the recommendation behavior, search the initial waybill candidate set for the transportation capacity of the current capacity pool, ensure the balanced allocation of resources between waybills and drivers, and generate a waybill recommendation candidate set.

[0023] As a preferred solution of the present invention, the step 32 specifically includes:

[0024] Step 321: Calculating the spatial proximity includes obtaining the destination filled in by the driver and the waybill destination, searching the two separately using the A* shortest path search algorithm, and calculating the en route distance d ab , as follows:

[0025]

[0026] In the formula, I a represents the shortest path to the driver’s destination, I b Indicates the shortest path of the waybill;

[0027] Step 322: Calculate the full load factor including the ratio W of the weight of cargo j to the maximum load of vehicle i ij , as follows:

[0028]

[0029] In the formula, w j represents the weight of cargo j, W i represents the maximum load of vehicle i;

[0030] Step 323: Calculate the waiting time of vehicle i again and normalize it:

[0031]

[0032] Where, t i is the normalized waiting time, T i represents the waiting time before normalization;

[0033] Step 324: Calculate the matching degree between the waybill and the driver using the output of the vehicle-cargo matching rule module. ij If the goods to be allocated have been transported by the driver, the matching degree o ij is 1; if the goods to be assigned have not been carried by the driver, the maximum co-occurrence value of the goods to be assigned and the type of goods it has carried is searched and used as the matching fitness o oj ;

[0034] Step 325: Construct a utility function based on the spatial dimension, load dimension, time dimension, and category dimension calculated in steps 301 to 304, where the weights of different indicators are represented respectively:

[0035] S=α*d ab +β*W ij +ε*t ij +δ*o ij

[0036] Wherein, S represents the utility function, and α, β, ε, and δ respectively represent the weights of different dimensional indicators.

[0037] As a preferred embodiment of the present invention, step S33 specifically includes:

[0038] Step S331: Define a binary matrix, X = x fk , x fk = 1, indicating that driver f is recommended order c k , otherwise x fk = 0.

[0039] Step S332: Set the constraint conditions. Each driver f recommends n orders: ∑ k x fk = n, and each order appears at most n times;

[0040] Step S333: Set the objective function to globally maximize the utility function: maximize ∑ j,k S·x fk .

[0041] As a preferred embodiment of the present invention, step S33 further specifically includes: introducing a pheromone decay update mechanism in the improved ant colony optimization Pareto front search algorithm, specifically as follows:

[0042]

[0043] Wherein, π PQ represents the pheromone concentration on the path (P, Q) in the improved ant colony optimization algorithm, and ρ represents the control rate of pheromone evaporation.

[0044] As a preferred embodiment of the present invention, step S4 specifically includes: pushing the candidate order list that meets the transportation of the drivers in the capacity pool according to the vehicle - cargo matching rules to the driver APP client.

[0045] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0046] The intelligent vehicle - cargo matching method based on heuristic rules provided by the present invention comprehensively considers the carrying conditions of the vehicle itself, accurately matches the suitable orders, recommends the candidate order sequence with the highest adaptability to the driver, and continuously optimizes the recommendation result through the dynamic adjustment mechanism, significantly improving the driver's usage experience and satisfaction, enhancing the driver's loyalty to the freight platform, and at the same time improving the logistics efficiency. Description of the Drawings

[0047] To more clearly illustrate the technical solutions in the embodiments of the present invention, the following will briefly introduce the accompanying drawings required for the description of the embodiments. Obviously, the accompanying drawings in the following description are only the preferred embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0048] Figure 1 The flow of the intelligent vehicle-cargo matching method based on heuristic rules provided by the present invention Figure 1 。

[0049] Figure 2 The flow of the intelligent vehicle-cargo matching method based on heuristic rules provided by the present invention Figure 2 。

[0050] Figure 3 The flow of the intelligent vehicle-cargo matching method based on heuristic rules provided by the present invention Figure 3 。 Detailed implementation manners

[0051] In order to make the objectives, technical solutions and advantages of the present invention more obvious, the following will describe in detail exemplary embodiments according to the present invention with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments of the present invention. It should be understood that the present invention is not limited by the exemplary embodiments described herein. Based on the embodiments of the present invention described in the present invention, all other embodiments obtained by those skilled in the art without creative efforts shall fall within the protection scope of the present invention.

[0052] In the following description, numerous specific details are given to provide a more thorough understanding of the present invention. However, it is obvious to those skilled in the art that the present invention can be implemented without one or more of these details. In other examples, to avoid confusion with the present invention, some well-known technical features in the art are not described.

[0053] It should be understood that the present invention can be implemented in different forms and should not be construed as limited to the embodiments presented herein. On the contrary, providing these embodiments will make the disclosure thorough and complete, and will fully convey the scope of the present invention to those skilled in the art.

[0054] The purpose of the terms used herein is only to describe specific embodiments and is not a limitation of the present invention. As used herein, the singular forms "a", "an" and "the" are also intended to include the plural forms unless the context clearly dictates otherwise. It should also be understood that the terms "comprising" and / or "including", when used in this specification, identify the presence of the stated features, integers, steps, operations, elements and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups. As used herein, the term "and / or" includes any and all combinations of the related listed items.

[0055] To thoroughly understand the present invention, detailed structures will be presented in the following description to illustrate the technical solutions proposed by the present invention. The optional embodiments of the present invention are described in detail below. However, in addition to these detailed descriptions, the present invention may also have other embodiments.

[0056] See Figure 1 and Figure 2 , an intelligent vehicle-cargo matching method based on heuristic rules, the method comprising the following steps:

[0057] Step S1: Obtain historical vehicle cargo data, vehicle-cargo matching data, and real-time goods to be transported and transport capacity data to be allocated, and preprocess the obtained data to construct a data set;

[0058] Step S2: Use the FP-growth algorithm to mine frequent item sets from the data set to generate a set of cargo types suitable for each transport capacity;

[0059] Step S3: Obtain the information of empty vehicle drivers who make appointments to cross the sea by mobile phone and generate a transport capacity pool. According to the set of cargo types suitable for each transport capacity, construct a utility function for multi-dimensional comprehensive waybill recommendation, and use the Pareto front search algorithm based on improved ant colony optimization to match the set information of cargo types with the information of empty vehicle drivers to generate a candidate set of waybill recommendations;

[0060] Step S4: Screen the candidate waybill list that meets the transportation requirements of the drivers in the transport capacity pool according to the vehicle-cargo matching rules and push it to the drivers.

[0061] Exemplarily, based on the frequent itemset mining technology, the FP-growth algorithm updated by using a sliding time window is used to extract the set of goods that can be transported in the same vehicle from the historical cargo records of trucks, and combined with the load and volume limits of the vehicle, the vehicle-cargo matching rule constraints are dynamically constructed. In addition, in order to improve the waybill matching rate, the present invention also proposes a Pareto front search algorithm based on improved ant colony optimization (ACO), comprehensively considering factors such as the current position of the vehicle, the position of the goods, the type of historical transported goods, and the load limit, generating a set of waybill candidates suitable for each truck, and pushing this candidate set to each driver through the driver side of the vehicle-cargo matching APP, so as to improve the efficiency of capacity matching and optimize the cross-sea transportation scheduling.

[0062] As a preferred solution of the present invention, the step S1 specifically includes: the preprocessing includes deleting abnormal data and duplicate data, filling in missing data, removing noise, and data standardization, which is convenient for subsequent data analysis.

[0063] As a preferred solution of the present invention, the step S2 specifically includes:

[0064] Step S21: Construct a frequent pattern tree to compress the data set, and gradually generate all frequent itemsets by recursively mining conditional pattern bases;

[0065] Step S22: Set a minimum support threshold for frequent itemsets with different numbers of items to filter out low-frequency association combinations. As the number of items in the frequent itemset increases, the threshold is appropriately reduced to mine more high-order association rules, and generate a set of cargo types suitable for each transportation capacity.

[0066] As a preferred solution of the present invention, in the step S22, in order to increase the mining ability for low-frequency association combinations, the analysis of cargo association based on the co-occurrence matrix is introduced, specifically as follows:

[0067] Step S221: Construct a co-occurrence matrix, and respectively count the individual occurrence times of each cargo type and the co-occurrence times of their pairwise combinations;

[0068] Specifically, an N*N co-occurrence matrix is constructed, where N is the number of cargo types, the value of the matrix element is the co-occurrence value, and then the Jaccard coefficient normalization is performed on all the co-occurrence values (co-occurrence vectors) corresponding to each type of cargo.

[0069] Step S222: Calculate the cosine similarity between the goods based on the statistical data, and use the similarity value as the edge weight of the weighted graph, and the nodes correspond to different cargo types, generating a cargo association graph;

[0070] Step S223: Further mine the potential association relationships between the goods through this cargo association graph, and optimize the vehicle-cargo matching rules.

[0071] Specifically, this step provides a basis for optimizing vehicle-cargo matching rules. At the same time, it introduces a sliding time window mechanism and periodically updates the rule base to adapt to vehicle modifications and seasonal changes in cargo types, ensuring dynamic optimization and continuous applicability of matching rules.

[0072] As a preferred solution of the present invention, step S3 specifically includes:

[0073] Step S31: Based on the physical constraints (load threshold, volume specification) and environmental constraints (the transport carrier has no history of transporting the goods), an initial waybill candidate set that meets the transport capacity of the current transport pool is generated;

[0074] Step S32: construct a multi-dimensional comprehensive utility function based on spatial proximity, full load rate and cargo type matching to quantify the capacity-waybill matching;

[0075] Step S33: A bidirectional cardinality constraint mechanism is introduced into the Pareto frontier search algorithm of improved ant colony optimization to control the recommendation behavior, and the initial candidate set of waybills for the transportation capacity of the current capacity pool is searched to ensure the balanced allocation of resources for waybills and drivers, and generate a recommended candidate set of waybills.

[0076] As a preferred solution of the present invention, the step 32 specifically includes:

[0077] Step 321: Calculating the spatial proximity includes obtaining the destination filled in by the driver and the waybill destination, searching the two separately using the A* shortest path search algorithm, and calculating the en route distance d ab , as follows:

[0078]

[0079] In the formula, I a represents the shortest path to the driver’s destination, I b Indicates the shortest path of the waybill;

[0080] Step 322: Calculate the full load factor including the ratio W of the weight of cargo j to the maximum load of vehicle i ij , as follows:

[0081]

[0082] In the formula, w j represents the weight of cargo j, W i represents the maximum load of vehicle i;

[0083] Step 323: Calculate the waiting time of vehicle i again and normalize it:

[0084]

[0085] Where, t i is the normalized waiting time, and T i represents the waiting time before normalization;

[0086] Step 324: Calculate the matching fitness o of the waybill and the driver using the output of the vehicle-cargo matching rule module ij , if the goods to be allocated have been carried by the driver, the matching fitness o ij is 1; if the goods to be allocated have not been carried by the driver, search for the maximum co-occurrence value between the goods to be allocated and the types of goods it has carried, and use it as the matching fitness o ij ;

[0087] Step 325: Construct a utility function based on the spatial dimension, load dimension, time dimension, and category dimension calculated in Steps 301 to 304, where represent the weights of different indicators respectively:

[0088] S = α * d ab + β * W ij + ε * t ij + δ * o ij

[0089] Where, S represents the utility function, and α, β, ε, and δ represent the weights of different dimension indicators respectively.

[0090] As a preferred solution of the present invention, the specific steps of S33 include:

[0091] Step S331: Define a binary matrix, X = x fk , x fk = 1, indicating that driver f is recommended for waybill c k , otherwise x fk = 0.

[0092] Step S332: Set the constraint conditions, each driver f recommends n waybills k: ∑ k x fk = n, and each waybill c k appears at most n times;

[0093] Step S333: Set the objective function to globally maximize the utility function: maximize ∑ j,k S · x fk .

[0094] Specifically, to avoid over-competition for waybill resources, for example, if some waybills are recommended to too many drivers, it will lead to fierce competition and may cause problems such as some drivers being unable to successfully grab the waybill. And to alleviate the problem of driver decision overload. If a driver receives too many recommended waybills at one time, it may increase their decision-making pressure and affect the selection efficiency. A two-way cardinality constraint mechanism is proposed to control the system's recommendation behavior, which not only controls the number of recommended drivers on the waybill side but also limits the recommendation competition degree on the waybill side, ensuring an even distribution of resources between waybills and drivers. Specifically, the upper limit n of the maximum number of drivers pushed for a single waybill (to avoid over-competition for waybill resources) and the maximum number of waybills pushed for a single driver (to alleviate driver decision overload) are set, thus transforming the problem into a constrained combinatorial optimization model. This method improves the traditional ACO algorithm to efficiently solve this NP-Hard problem.

[0095] Specifically, the exposure rate of waybills is controlled through a two-way cardinality constraint mechanism; to address the problem of path dependence during the search for the optimal solution caused by the fixed pheromone in the original ant colony algorithm, a dynamic pheromone update method is proposed. According to the waybill exposure rate, the pheromone for searching for this result is adaptively reduced, and then a candidate set of waybills that conforms to the transport capacity rules is generated, avoiding falling into local optimal solutions. The algorithm flow is as Figure 3 shown.

[0096] However, the traditional ant colony algorithm has the problem that static pheromone updates cannot meet the waybill exposure (the above-mentioned constraint conditions to prevent waybills from being viewed by multiple people) conditions. To solve the above problems, a dynamic exposure constraint processing and a pheromone dynamic update method are respectively planned to improve the ACO algorithm to adapt to the current scenario. Among them, the dynamic exposure constraint processing designs a candidate set masking mechanism. During each iteration, the optional waybill set for the driver is dynamically constructed, and the waybills that have reached the exposure limit are masked to ensure that they no longer appear in the recommendation list. And each waybill c k maintains an exposure count value E(c k ), which is used to record the number of times this waybill has been recommended. When E(c k ) reaches the set exposure limit E_max, this waybill will be masked to prevent further exposure. At the same time, the traditional ant colony algorithm uses a fixed pheromone update strategy, which is prone to causing some waybills to maintain a high pheromone concentration for a long time in this scenario, affecting the diversity of recommendations. Therefore, we introduce a pheromone decay mechanism, so that the recommendation probability of waybills is dynamically adjusted as the number of exposure times increases: when the exposure times E(c k ) of waybill c k approaches the upper limit (such as 4 times), the pheromone evaporation rate is accelerated, so that the pheromone concentration of this waybill rapidly decreases, reducing its selection probability in subsequent recommendations. Among them, the dynamic exposure constraint processing designs a candidate set masking mechanism to prevent recommending duplicate waybills to the driver.

[0097] As a preferred embodiment of the present invention, the step S33 further specifically includes: introducing a pheromone decay update mechanism into the improved ant colony optimization Pareto front search algorithm, specifically as follows:

[0098]

[0099] In the formula, π PQ represents the pheromone concentration on the path (P, Q) in the improved ant colony optimization algorithm, and ρ represents the control pheromone evaporation rate. The value range of ρ is [0, 1]. In the present invention, ρ = 0.2 and n = 4.

[0100] As a preferred embodiment of the present invention, the step S4 specifically includes: screening a candidate waybill list that meets the transportation requirements of the drivers in the capacity pool according to the vehicle-cargo matching rule, and pushing it to the driver APP client to recommend suitable orders to the driver, thereby improving the logistics efficiency.

[0101] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. An intelligent vehicle-cargo matching method based on heuristic rules, characterized in that, The method includes the following steps: Step S1: Obtain historical vehicle cargo data, vehicle-cargo matching data, as well as real-time to-be-carried cargo and to-be-allocated transport capacity data, and preprocess the obtained data to construct a data set; Step S2: Use the FP-growth algorithm to mine frequent item sets from the data set to generate a set of cargo types suitable for each transport capacity to carry; Step S3: Obtain the information of empty truck drivers who make appointments to cross the sea through mobile phones and generate a transport capacity pool. According to the set of cargo types suitable for each transport capacity to carry, construct a utility function for multi-dimensional comprehensive waybill recommendation, and use the Pareto front search algorithm based on improved ant colony optimization to match the information of the set of cargo types with the information of empty truck drivers to generate a candidate set of waybill recommendations; Step S4: Screen the candidate waybill list that meets the drivers in the transport capacity pool according to the vehicle-cargo matching rules and push it to the drivers.

2. The intelligent vehicle-cargo matching method based on heuristic rules according to claim 1, wherein The specific content of step S1 includes: The preprocessing includes deleting abnormal data and duplicate data, filling in missing data, removing noise, and data standardization.

3. The intelligent vehicle-cargo matching method based on heuristic rules according to claim 2, wherein The specific content of step S2 includes: Step S21: Construct a frequent pattern tree to compress the data set, and recursively mine conditional pattern bases to gradually generate all frequent item sets; Step S22: Set a minimum support threshold for frequent item sets with different numbers of items to filter low-frequency association combinations. As the number of items in the frequent item sets increases, appropriately reduce the threshold to mine more high-order association rules and generate a set of cargo types suitable for each transport capacity to carry.

4. The intelligent vehicle-cargo matching method based on heuristic rules according to claim 3, characterized in that In step S22, to increase the mining ability for low-frequency association combinations, the co-occurrence matrix is introduced to analyze cargo associations, specifically as follows: Step S221: Construct a co-occurrence matrix, and respectively count the individual occurrence times of each cargo type and the co-occurrence times of their pairwise combinations; Step S222: Calculate the cosine similarity between cargos based on the statistical data, and use the similarity value as the edge weight of the weighted graph, where the nodes correspond to different cargo types to generate a cargo association graph; Step S223: Further mine the potential association relationships between cargos through this cargo association graph to optimize the vehicle-cargo matching rules.

5. The intelligent vehicle-cargo matching method based on heuristic rules according to claim 4, characterized in that, The specific content of step S3 includes: Step S31: Based on the physical constraints and environmental constraints of the transport carrier, generate an initial candidate set of waybills that meet the transport capacity of the current transport capacity pool; Step S32: Construct a multi-dimensional comprehensive utility function with spatial proximity, load factor, and cargo type matching degree to quantify the transport capacity-waybill matching degree; Step S33: Introduce a two-way cardinality constraint mechanism in the Pareto front search algorithm based on improved ant colony optimization to control the recommendation behavior, search the initial candidate set of waybills for the transport capacity of the current transport capacity pool, and ensure the balanced allocation of resources between waybills and drivers to generate a candidate set of waybill recommendations.

6. The intelligent vehicle-cargo matching method based on heuristic rules according to claim 5, wherein, The specific content of step 32 includes: Step 321: Calculating the spatial proximity includes obtaining the destination filled in by the driver and the waybill destination, searching the two separately using the A* shortest path search algorithm, and calculating the en route distance d ab , as follows: Wherein, I a represents the shortest path to the driver's destination, and I b represents the shortest path of the waybill; Step 322: Calculate the full load rate, which is the ratio W of the weight of cargo j to the maximum load of vehicle i, as follows: ij , specifically as follows: where w j represents the weight of cargo j, and W i represents the maximum load capacity of vehicle i; Step 323: Then calculate the waiting time of vehicle i and normalize it: where t i is the normalized waiting time, and T i represents the waiting time before normalization; Step 324: Calculate the matching fitness o between the waybill and the driver using the output of the vehicle-cargo matching rule module ij , if the goods to be allocated have been carried by the driver, the matching fitness o ij is 1; if the goods to be allocated have not been carried by the driver, search for the maximum co-occurrence value between the goods to be allocated and the types of goods it has carried, and use it as the matching fitness o ij ; Step 325: Construct a utility function according to the spatial dimension, load dimension, time dimension, and category dimension calculated in steps 301 to 304, where respectively represent the weights of different indicators: S = α * d ab + β * W ij + ε * t ij + δ * o ij In the formula, S represents the utility function, and α, β, ε, δ respectively represent the weights of different dimension indicators.

7. The intelligent vehicle-cargo matching method based on heuristic rules according to claim 6, wherein The specific content of step S33 includes: Step S331: Define a binary matrix, X = x fk , x fk = 1, indicating that driver f is recommended order c k , otherwise x fk = 0; Step S332: Set the constraint conditions, where each driver f recommends n waybills: ∑ k x fk = n, and each waybill appears at most n times; Step S333: Set the objective function to globally maximize the utility function: maximize ∑ j,k S·x fk .

8. The intelligent vehicle-cargo matching method based on heuristic rules according to claim 7, characterized in that Step S33 further specifically includes: introducing a pheromone decay and update mechanism into the improved ant colony optimization-based Pareto front search algorithm, which is specifically as follows: Where, π PQ represents the pheromone concentration on the path (P, Q) in the improved ant colony optimization algorithm, and · represents the evaporation rate of the pheromone control.

9. The intelligent vehicle-cargo matching method based on heuristic rules according to claim 8, wherein, Step S4 specifically includes: screening a candidate list of shipping orders that meet the transportation requirements of drivers in the capacity pool according to the vehicle-cargo matching rules, and pushing it to the driver APP client.

Images