A method and system for dynamic matching and optimization of logistics orders and transportation capacity resources
By using a multi-dimensional fit evaluation model and a hierarchical optimization strategy, the problem of low efficiency in matching logistics orders with transportation resources was solved, achieving efficient and real-time dynamic matching of orders and transportation resources, improving matching quality and transportation capacity utilization, and reducing computational complexity and logistics costs.
Patent Information
- Application Number
- CN202611494473.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-23
- Publication Date
- 2026-11-17
AI Technical Summary
Existing technologies struggle to achieve multi-dimensional dynamic optimization in matching logistics orders with transportation resources, resulting in low matching efficiency, inability to cope with frequent dynamic changes, and high computational complexity, failing to meet real-time requirements.
By constructing a multi-dimensional suitability evaluation model that includes distance, direction, capacity, time, route coordination, and cost, and combining a hierarchical optimization strategy and an incremental optimization adjustment mechanism, data is collected in real time for preprocessing and feature construction to generate an initial matching scheme, and local adjustments are made when dynamic changes occur.
It achieves precise matching between orders and transportation capacity, improves the average compatibility by 35%, reduces computational complexity, increases response speed by 10 times, achieves a capacity utilization rate of over 80%, and reduces the empty-running rate to below 15%, meeting the matching needs of order sizes from hundreds to tens of thousands.
Smart Images

Figure CN123117955A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of logistics scheduling technology, and in particular to a method and system for dynamic matching and optimization of logistics orders and transportation capacity resources. Background Technology
[0002] With the rapid development of e-commerce and modern logistics, the volume of logistics orders has exploded, posing a significant challenge for logistics companies to efficiently match massive orders with limited transportation resources. Traditional logistics companies primarily rely on dispatchers' experience to match and allocate orders with transportation capacity. Dispatchers manually assign orders based on information such as the destination, type of goods, and time requirements, combined with factors such as the location, load capacity, and routes of available vehicles.
[0003] Among the existing related patent technologies, some solutions only focus on matching in a single dimension. For example, they only consider distance or only consider load, lacking multi-dimensional comprehensive optimization; some solutions, although considering multiple factors, adopt static optimization strategies, which cannot cope with the frequent dynamic changes in logistics scenarios; and some solutions, although possessing certain dynamic adjustment capabilities, lack efficient incremental optimization mechanisms, resulting in poor performance in practical applications.
[0004] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention
[0005] This invention provides a method for dynamically matching and optimizing logistics orders and transportation capacity resources, comprising: real-time collection of raw data, including order data, transportation capacity resource data, and environmental data; preprocessing the raw data; constructing order feature vectors and order classifications for the preprocessed order data; constructing transportation capacity resource feature vectors and availability status assessments for the preprocessed transportation capacity resource data; calculating the matching fit degree between orders and transportation capacity resources to construct a fit degree matrix; generating an initial matching scheme based on the fit degree matrix; executing the matching scheme and updating the matching scheme based on dynamic change events during the execution process.
[0006] Furthermore, the raw data undergoes preprocessing, including the following steps: Perform data cleaning on the raw data; The raw data is standardized using the following formula:
[0007] in, These are the original eigenvalues; This represents the minimum value of the original eigenvalues. The maximum value of the original eigenvalues; The geographic information of the raw data is processed to convert the shipping address and delivery address into unified geographic coordinates; The time information of the raw data is standardized to convert it into UTC timestamp format.
[0008] Furthermore, an order feature vector is constructed from the preprocessed order data, including the following steps: For each order Construct order feature vectors The formula used is:
[0009] in, Geographic feature components; For cargo characteristic components; For time-sensitive characteristic components; Priority components; For special requirements; For customer characteristic components.
[0010] Geographic feature components The formula used is:
[0011] in, The longitude of the shipping address; The latitude and longitude of the shipping address; The longitude of the delivery address; The latitude and longitude of the delivery address; The transport direction angle is given by the formula:
[0012] The straight-line distance between the shipping location and the receiving location is calculated using the following formula:
[0013] in, The radius of the Earth; Cargo characteristic components The formula used is:
[0014] in, Standardized weight; Standardized volume; For cargo density; Code the type of goods; Time-dependent characteristic components The formula used is:
[0015] in, Remaining available time; This is an urgency index; Timeliness level; Urgency Index The formula used is:
[0016] in, To estimate the delivery time.
[0017] Further, order tiering includes the following steps: Based on the order feature vector, a comprehensive urgency score is calculated for each order using the following formula:
[0018] in, , , , These are the weighting coefficients; Scoring will be provided for special requirements; This is the original priority value for the order; Customer level; Based on the overall urgency score Orders are categorized into three levels: urgent orders, regular orders, and flexible orders.
[0019] Furthermore, a capacity resource feature vector is constructed from the preprocessed capacity resource data, including the following steps: For each transportation resource Construct a feature vector of transportation capacity resources The formula used is:
[0020] in, Location features; Capacity characteristics; State characteristics; As a characteristic of ability; Characteristics of the driver; This refers to route characteristics.
[0021] Location features The formula used is:
[0022] in, The longitude of the vehicle's current location; The latitude of the vehicle's current location; The vehicle's current location coding; The administrative region code for the vehicle; Capacity characteristics The formula used is:
[0023] in, The remaining available load capacity is calculated using the following formula:
[0024] in, This refers to the vehicle's maximum load capacity. This represents the current loaded weight. The available remaining volume is calculated using the following formula:
[0025] in, This refers to the maximum volume of the vehicle. This represents the currently loaded volume; The formula for weight utilization rate is as follows:
[0026] The formula for calculating physical exertion efficiency is as follows:
[0027] Driver characteristics The formula used is:
[0028] in, The normalized value of the driver's overall score; This is a normalized value for the remaining available driving time; The fatigue index is expressed by the following formula:
[0029] in, This represents the driver's current continuous driving time. The maximum continuous driving time stipulated by regulations; Route characteristics The formula used is:
[0030] in, This represents the number of orders currently allocated to this capacity. This is the estimated timestamp for completing all current delivery tasks; This is the main direction angle of the current route; The geographical area covered by the route.
[0031] Further, an availability status assessment is conducted, including the following steps: Calculate the overall availability score for each transportation capacity resource. The formula used is:
[0032] in, , , , These are the weighting coefficients; Assess availability; Evaluate the driver's overall performance; Rate the remaining capacity; Rate the route flexibility; Remaining capacity rating The formula used is:
[0033] Route flexibility rating The formula used is: ; in, This represents the maximum number of orders that the capacity can handle.
[0034] Furthermore, the fitness matrix is constructed, including the following steps: Perform constraint checks; Perform multi-dimensional adaptation calculation, which includes distance adaptation, direction adaptation, capacity adaptation, time adaptation, route coordination, and cost adaptation. Calculate the overall fit score; A fitness matrix is constructed based on the overall fitness score.
[0035] Furthermore, generating an initial matching scheme includes the following steps: For each order From the fitness matrix of the th Select the top of the fit ranking in the row Transportation capacity resources as a candidate set The formula used is:
[0036] in, For orders With transport capacity The overall fit score; The value of K is determined based on the scale of transportation capacity, using the following formula:
[0037] in, This refers to the amount of available transport capacity. The upper limit of the candidate set; Based on the order priority from highest to lowest, the optimal capacity is selected within each priority level based on the urgency score from highest to lowest to generate an initial plan. The initial scheme is optimized through a neighborhood search operation.
[0038] This invention also provides a dynamic matching and optimization system for logistics orders and transportation capacity resources, applying the method described above. It includes: a data preprocessing module, a feature construction and hierarchical evaluation module, a fitness matrix construction module, and a matching scheme generation module. The data preprocessing module is used to collect raw data in real time and preprocess the raw data, which includes order data and transportation capacity resource data. The feature construction and hierarchical evaluation module is used to construct order feature vectors and transportation capacity resource feature vectors based on the preprocessed order data and transportation capacity resource data, and to classify orders based on the order feature vectors and to perform a comprehensive availability score based on the transportation capacity resource feature vectors. The fitness matrix construction module is used to calculate multi-dimensional fitness scores for order-capacity pairs that meet the constraints, and to calculate a comprehensive fitness score through weighted summation, and to construct a fitness matrix based on the comprehensive fitness score. The matching scheme generation module is used to generate an initial matching scheme based on the fitness matrix through candidate set screening, hierarchical greedy matching, and local search optimization.
[0039] Furthermore, it also includes a dynamic incremental optimization module and a scheme output module; the dynamic incremental optimization module is used to detect dynamic change events when executing the matching scheme, to assess the scope of the event's impact, and to perform incremental optimization adjustments to update the matching scheme; the scheme output module is used to output the updated matching scheme.
[0040] Compared with the prior art, the present invention has the following advantages: This invention achieves precise matching of orders and transportation capacity by constructing a suitability evaluation model that includes six dimensions: distance, direction, capacity, time, route coordination, and cost. Compared with the traditional matching method based on a single distance, the average suitability is improved by about 35%.
[0041] Meanwhile, by employing a hierarchical optimization strategy (candidate set selection → hierarchical greedy matching → local search optimization), the computational complexity is reduced from... Reduce to (in With a scale of 500 orders × 80 transport capacity, the total calculation time is controlled within 4 seconds, meeting the real-time requirements.
[0042] Secondly, through the incremental optimization and adjustment mechanism, the matching scheme can be partially adjusted within seconds when dynamic events such as new order insertion, capacity failure, and traffic anomalies occur, without the need for global re-solution, and the response speed is more than 10 times faster than global re-optimization.
[0043] On the other hand, by using an adaptive weight adjustment strategy, the optimization focus of each dimension is automatically adjusted according to the order level, achieving a balanced optimization of multiple objectives such as timeliness, cost, and utilization.
[0044] This invention also optimizes route coordination and capacity adaptability, achieving an average capacity utilization rate of over 80% and reducing the empty-running rate to below 15%, effectively lowering logistics costs.
[0045] Meanwhile, the hierarchical optimization strategy and regional parallel computing mechanism enable this invention to adapt to matching needs ranging from hundreds to tens of thousands of order sizes. Attached Figure Description
[0046] Figure 1 Overall process diagram. Detailed Implementation
[0047] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of the invention.
[0048] Example 1: This invention provides a method for dynamically matching and optimizing logistics orders and transportation capacity resources, referring to... Figure 1 , Figure 1 This is a flowchart illustrating the process of this invention. It includes: real-time collection of raw data, which includes order data, transportation resource data, and environmental data. Specifically, order data includes: Order ID and shipping address information, including longitude. ,latitude Administrative division codes Delivery address information, including longitude. ,latitude Administrative division codes Cargo information. Includes cargo type. ,weight ,volume , number of pieces Delivery time requirements, including expected pickup time. Required delivery time Timeliness level Order priority. Special requirements, including whether cold chain is required, whether the item is fragile, and whether a sign is required. Customer information, including customer level and historical complaint rate.
[0049] Transportation capacity data includes VehicleID and the vehicle's current location, including longitude. ,latitude Vehicle Type. Vehicle Capacity Information, including Maximum Weight (MaxWeight) and Maximum Volume (MaxVolume). Currently Used Capacity, including Loaded Weight (UsedWeight) and Occupied Volume (UsedVolume). Vehicle Status. Driver Information, including DriverID, Driving Hours, and DriverScore. Planned Route Information, including Assigned Orders and Estimated Finish Time (EstFinishTime). Vehicle Special Capabilities, including whether it has Cold Chain capability (HasColdChain) and GPS real-time tracking capability (HasGPS).
[0050] Environmental data includes real-time traffic data, specifically the Congestion Index and EstTravelTime. Weather data includes Weather Type and Weather Level. Regional delivery density data includes the OrderDensity of orders awaiting delivery within the region.
[0051] Preprocessing the raw data includes the following steps: The raw data is cleaned to remove duplicates, correct outliers, and fill in missing values. For orders with missing address information, geocoding is used to complete the information; for vehicle location data with abnormal GPS signals, Kalman filtering is used for smoothing correction.
[0052] Standardization is performed on the raw data to normalize the feature data of different dimensions, mapping each feature value to the interval [0,1]. The formula is as follows:
[0053] in, These are the original eigenvalues; This represents the minimum value of the original eigenvalues. The maximum value of the original eigenvalues; The geographic information of the raw data is processed to convert the shipping address and receiving address into unified geographic coordinates, and a spatial index is built based on GeoHash encoding to support subsequent fast spatial queries. The GeoHash encoding precision is set to 7 bits, corresponding to a grid of approximately 150 meters × 150 meters.
[0054] The time information of the raw data is standardized to convert it into UTC timestamp format, and the remaining available time for each time window is calculated.
[0055] The process of constructing order feature vectors and order classifications from preprocessed order data includes the following steps: For each order Construct order feature vectors The formula used is:
[0056] in, Geographic feature components; For cargo characteristic components; For time-sensitive characteristic components; For priority components, the order priority value is directly retrieved; For special requirements; For customer characteristic components.
[0057] Geographic feature components The formula used is:
[0058] in, The longitude of the shipping address; The latitude and longitude of the shipping address; The longitude of the delivery address; The latitude and longitude of the delivery address; The transport direction angle is given by the formula:
[0059] The straight-line distance between the shipping location and the receiving location is calculated using the following formula:
[0060] in, The radius of the Earth; Cargo characteristic components The formula used is:
[0061] in, Standardized weight; Standardized volume; For cargo density; Code the type of goods.
[0062] Time-dependent characteristic components The formula used is:
[0063] in, The remaining available time (the time difference between the current time and the required delivery time); This is an urgency index; Timeliness level; Urgency Index The formula used is:
[0064] in, To estimate the delivery time.
[0065] When the urgency index If the value is greater than 1, it means that under the current conditions, delivery may not be possible on time and needs to be prioritized.
[0066] Priority components This is to directly retrieve the order priority value.
[0067] Special requirements for quantity To encode each special requirement into a binary vector, for example: .
[0068] Customer characteristic components ,include .in, The normalized historical customer complaint rate; Rate customers based on satisfaction or good track record.
[0069] Based on the order feature vector, a comprehensive urgency score is calculated for each order using the following formula:
[0070] in, , , , The weighting coefficient is the preferred value; the default value is [value to be filled in]. In practice, adjustments can be made based on business needs. Orders with special requirements will receive a higher rating. This is the original priority value for the order; Customer level; Based on the overall urgency score Orders are categorized into three levels, including: Urgent orders (Level A) ,default .
[0071] Standard order (Level B) ,default .
[0072] Flexible orders (Level C) .
[0073] The purpose of tiering is to prioritize the matching quality of higher-tier orders in subsequent matching optimization.
[0074] The preprocessed capacity resource data is used to construct a capacity resource feature vector and an availability status assessment, which specifically includes the following steps: For each transportation resource Construct a feature vector of transportation capacity resources The formula used is:
[0075] in, Location features; Capacity characteristics; State characteristics; As a characteristic of ability; Characteristics of the driver; This refers to route characteristics.
[0076] Location features The formula used is:
[0077] in, The longitude of the vehicle's current location; The latitude of the vehicle's current location; The vehicle's current location coding; The administrative region code for the vehicle; Capacity characteristics The formula used is:
[0078] in, The remaining available load capacity is calculated using the following formula:
[0079] in, This refers to the vehicle's maximum load capacity. This represents the current loaded weight. The available remaining volume is calculated using the following formula:
[0080] in, This refers to the maximum volume of the vehicle. This represents the currently loaded volume; The formula for weight utilization rate is as follows:
[0081] The formula for calculating physical exertion efficiency is as follows:
[0082] State characteristics Specifically, the vehicle status is encoded as a numerical value: Idle = 1.0, Loading = 0.8, En route = 0.6, Unloading = 0.4, Unavailable = 0.0.
[0083] Ability characteristics , to be encoded as a binary vector .in, To indicate whether the facility has cold chain transportation capabilities (1 = yes, 0 = no). Indicate whether real-time GPS positioning is provided (1=yes, 0=no).
[0084] Driver characteristics The formula used is:
[0085] in, The normalized value of the driver's overall score; This is a normalized value for the remaining available driving time; The fatigue index is expressed by the following formula:
[0086] in, This represents the driver's current continuous driving time. The maximum continuous driving time stipulated by regulations; Route characteristics The formula used is:
[0087] in, This represents the number of orders currently allocated to this capacity. This is the estimated timestamp for completing all current delivery tasks; This is the main direction angle of the current route; The geographical area covered by the route.
[0088] Calculate the overall availability score for each transportation capacity resource. The formula used is:
[0089] in, , , , The weighting coefficient is the preferred value; the default value is [value to be filled in]. .
[0090] Rate the remaining capacity. Specifically:
[0091] Score the availability of the status (i.e.) (Encoded value).
[0092] A comprehensive rating is given to the driver. Specifically, .
[0093] in, This is a normalized driver's historical comprehensive score.
[0094] Rate the route flexibility; Route flexibility rating The formula used is: ; in, This represents the maximum number of orders that the capacity can handle. This represents the number of orders currently allocated to this capacity.
[0095] when (default When this occurs, the transportation resource is marked as unavailable and will not participate in subsequent matching calculations.
[0096] Calculate the matching degree between orders and transportation capacity resources to construct a matching degree matrix. This includes the following steps: Before constructing the fit matrix, constraint checks are required. Specifically, for orders... and transport capacity The following hard constraints need to be met: Capacity constraints:
[0097] Capability constraints: Specific requirements of an order must be covered by available transportation capacity. For example, if an order requires cold chain transportation, then the transportation capacity must have cold chain capabilities.
[0098] Time window constraints: The estimated time for the delivery vehicle to arrive at the pickup point is no later than the latest pickup time of the order, and delivery can be completed before the delivery deadline.
[0099]
[0100] in, For transport capacity Drive from current location to order The estimated arrival time required at the pickup point. For orders The latest pickup time. For orders Estimated delivery time from the pickup point to the delivery point. For orders The required delivery deadline.
[0101] Driver working hours constraints:
[0102] in, To place orders After being allocated to this capacity, additional driving time is expected to be required.
[0103] For orders and shipping capacity that do not meet any of the above constraints, their fit is directly set to 0.
[0104] Multidimensional adaptability calculation is performed, which includes distance adaptability, direction adaptability, capacity adaptability, time adaptability, route coordination, and cost adaptability.
[0105] Specifically, distance adaptability for:
[0106] in, This represents the distance from the current location of the delivery vehicle to the order pickup point. This is the distance attenuation coefficient. Preferably, it is set to 0.05 by default.
[0107] Orientation adaptability for:
[0108] in, This indicator measures the consistency between the order's transportation direction and the current route direction of the transport capacity. The value range is [0,1], and it is 1 when the directions are completely consistent.
[0109] Capacity compatibility
[0110]
[0111] in, For orders The original weight of the goods. The target load rate is set to 0.8 by default.
[0112] This indicator encourages allocating orders to capacity that is close to the target load factor after loading, thus avoiding resource waste or over-concentration.
[0113] Time Adaptability for:
[0114] in, To place orders Allocate to transport capacity The estimated completion time. The larger this value, the more time leeway there is.
[0115] Route Coordination for:
[0116] in, For orders With transport capacity Assigned orders Geographical similarity between them:
[0117] in, For transport capacity The first in the allocated order set One order; This is the geographical similarity attenuation coefficient; For orders Delivery points and orders The distance between the receiving points; For orders Delivery points and orders The distance between pickup points.
[0118] This metric measures the route synergy between new orders and existing delivery tasks within the delivery capacity; higher synergy means higher efficiency in consolidated delivery.
[0119] Cost fit for:
[0120] in, To place orders Allocate to transport capacity The marginal cost (including additional fuel cost, time cost, detour cost, etc.); This represents the highest marginal cost among all candidate capacities.
[0121] The overall fit score is calculated by weighted summation. ;
[0122] Among them, the weight vector satisfy .
[0123] The weights are set in the following way: For Level A (urgent) orders: Increase (Time Adaptability Weight), Default .
[0124] For B-level (normal) orders: Balance all dimensions, default. .
[0125] For Category C (flexible) orders: Increase and (Route coordination and cost suitability weights), default .
[0126] Based on the overall fit score Constructing the fitness matrix The element in the i-th row and j-th column is... .
[0127]
[0128] in, The number of orders to be matched. This represents the amount of available transport capacity.
[0129] Based on the fitness matrix, an initial matching scheme is generated. This includes the following steps: Perform candidate set screening (coarse screening): For each order From the fitness matrix of the th Select the top of the fit ranking in the row Transportation capacity resources as a candidate set The formula used is:
[0130] in, For orders With transport capacity The overall fit score; The value of K is determined based on the scale of transportation capacity, using the following formula:
[0131] in, This refers to the amount of available transport capacity. This is the upper limit of the candidate set, with a default value of 20.
[0132] This step will match the search space from Reduced to This significantly reduces the computational complexity of subsequent optimizations.
[0133] Perform a tiered greedy matching algorithm: Based on order priority from highest to lowest, within each priority level, the optimal capacity is selected based on urgency rating from highest to lowest: For each order within each order level in {A, B, C} according to Sort in descending order.
[0134] From the candidate set Select the capacity with the highest suitability and still available capacity. The formula used is:
[0135] Will Assigned to ,renew The state, and recalculate. Availability score .
[0136] Thus, a feasible initial solution is generated. .
[0137] For the initial scheme Local search optimization is performed through neighborhood search operations. The specific process is as follows: Single-point relocation: Moving an order from its currently allocated capacity to another capacity:
[0138] Two-way swap: Swap the allocation of one order each on two different shipping lines.
[0139] Chain Relocate: Moving an order from one capacity to another. Migrate to At the same time One order from the previous one has been migrated to This forms a chain of adjustments:
[0140] The objective function of the local search is the global matching quality score. :
[0141] in, This represents the amount of transport capacity currently used in the plan; For transport capacity Utilization rate. This is the utilization rate reward coefficient. This represents the penalty coefficient for unmatched orders.
[0142] In summary, the first item is the sum of the compatibility of all matched pairs, the second item is the capacity utilization bonus, and the third item is the penalty for unmatched orders.
[0143] Local search employs a tabu search strategy, maintaining a tabu table. Record recently executed operations to avoid looping searches. The search termination condition is: Reaching the maximum number of iterations The default value is 500 times.
[0144] continuous No improvement was found in the second iteration. The default is 50 iterations.
[0145] Calculation time exceeded the time limit The default time is 5 seconds.
[0146] Matching scheme representation stage: Final matching scheme Represented as a set of order-capacity allocation pairs:
[0147] Simultaneously record the set of orders that failed to match. and unused capacity .
[0148] To address various dynamic changes that occur during the matching scheme execution process, an incremental optimization and adjustment mechanism is used to quickly update the matching scheme, avoiding a global re-solution. The specific process is as follows: Define the following dynamic event types:
[0149] When a dynamic event occurs, first assess the scope of its impact to determine the local areas that need adjustment: For new order insertion events : Calculate the radius around the pickup point for new orders All available transportation capacity within (default 30 km) .
[0150] The scope of influence is and its allocated orders For capacity failure events : Acquiring faulty transport capacity The set of all assigned but not yet completed orders .
[0151] The scope of influence is And the available shipping capacity near the pickup / delivery points for these orders.
[0152] For traffic abnormalities : Identify all available transport capacity passing through the affected road sections. .
[0153] Assess the time impact of traffic disruptions on each affected capacity. .
[0154] like If the time window constraints of some orders are violated, these orders will be added to the set that needs to be readjusted.
[0155] Based on the event type and its scope of impact, perform corresponding incremental optimization adjustments: A) Incremental optimization for inserting new orders: Input: New Order Current matching scheme P. Output: Updated matching scheme P'.
[0156] Step 1: Building eigenvectors Step 2: Calculation and The adaptability of each transport capacity Step 3: Sort candidate transport capacity in descending order of suitability. Step 4: For each candidate capacity : Check if the hard constraints are satisfied; if so, the calculation will proceed. insert Optimal location in the delivery sequence Optimal insertion position
[0157] CostIncrease represents the total additional delivery cost after insertion.
[0158] Calculate the change in solution quality ΔQ after insertion; if ΔQ>0 (the solution quality is improved or acceptable), perform insertion and update the solution.
[0159] Step 5: If all candidate transport capacity is not feasible, Add to the unmatched queue and trigger a wider range of local search optimizations.
[0160] Calculation of the optimal insertion position in the delivery sequence: Regarding transport capacity Current delivery sequence new orders Insertion position The subsequent cost increment is:
[0161] in For the first in the sequence The location of that location, For the first The location of each position needs to be verified. Simultaneously, it needs to be confirmed that the time window constraints for all subsequent orders are still satisfied after the insertion.
[0162] B) Incremental optimization of capacity failures Input: Faulty Capacity The current matching scheme P; Output: The updated matching scheme P'; Step 1: Obtain The collection of all incomplete orders ; Step 2: Marked as unavailable; Step 3: Sort by urgency in descending order Order sorting in the system; Step 4: For each affected order : Search for the best alternative capacity among nearby available capacity. If a feasible alternative is found, perform a reallocation. If no feasible alternative is found, trigger an alarm and attempt to expand the search scope.
[0163] Step 5: Perform local search optimization on the affected areas to further improve the solution.
[0164] C) Incremental optimization of traffic anomalies Input: Traffic anomaly information (affected road sections, estimated delay time), current matching scheme P; Output: The updated matching scheme P'; Step 1: Identify all affected transport capacity ; Step 2: For each affected capacity : Recalculate the estimated arrival time for each order in the delivery sequence and identify the set of orders where time window constraints are violated. .
[0165] Step 3: [Regarding...] Orders in progress: Try adjusting the delivery order to meet the time window constraints. If adjusting the order does not solve the problem, try migrating some orders to other delivery capacity. If the problem still does not solve the problem, replan the detour route and update the estimated time.
[0166] Step 4: Perform local search optimization: After each incremental adjustment, a consistency check is performed: Verify that all hard constraints are still met; verify that no orders have been duplicated; verify that no orders have been lost during the adjustment process; update the status information of all affected capacity.
[0167] For the final matching scheme Calculate the following evaluation indicators: Overall match rate:
[0168] in, The number of successfully matched orders; This represents the total number of orders to be matched in the system.
[0169] Average fit:
[0170] Average capacity utilization rate:
[0171] in, The collection of transport capacity being used; Timeliness fulfillment rate:
[0172] in, A collection of orders that are delivered on time; Total delivery cost:
[0173] in, For transport capacity Fuel costs; For transport capacity Time cost; For transport capacity Toll fees.
[0174] Empty run rate:
[0175] in, For transport capacity Empty mileage; For transport capacity Total mileage.
[0176] The output matching scheme contains the following information: Each delivery capacity's delivery task list (including order information, delivery sequence, and estimated arrival time); recommended routes for each delivery capacity; a list of unmatched orders and explanations of the reasons; an overall plan quality assessment report; and dynamic monitoring trigger condition settings.
[0177] In summary, this invention achieves precise matching of orders and transportation capacity by constructing a suitability evaluation model that includes six dimensions: distance, direction, capacity, time, route coordination, and cost. Compared with traditional matching methods based on a single distance, the average suitability is improved by about 35%.
[0178] Meanwhile, by employing a hierarchical optimization strategy (candidate set selection → hierarchical greedy matching → local search optimization), the computational complexity is reduced from... Reduce to (in With a scale of 500 orders × 80 transport capacity, the total calculation time is controlled within 4 seconds, meeting the real-time requirements.
[0179] Secondly, through the incremental optimization and adjustment mechanism, the matching scheme can be partially adjusted within seconds when dynamic events such as new order insertion, capacity failure, and traffic anomalies occur, without the need for global re-solution, and the response speed is more than 10 times faster than global re-optimization.
[0180] On the other hand, by using an adaptive weight adjustment strategy, the optimization focus of each dimension is automatically adjusted according to the order level, achieving a balanced optimization of multiple objectives such as timeliness, cost, and utilization.
[0181] This invention also optimizes route coordination and capacity adaptability, achieving an average capacity utilization rate of over 80% and reducing the empty-running rate to below 15%, effectively lowering logistics costs.
[0182] Meanwhile, the hierarchical optimization strategy and regional parallel computing mechanism enable this invention to adapt to matching needs ranging from hundreds to tens of thousands of order sizes.
[0183] To facilitate understanding of the above process, the following calculation examples are provided: Daily batch order matching scenario: Scenario Description: A logistics company processes the logistics orders accumulated from the previous day in the morning, and needs to match 500 orders to be delivered with 80 available vehicles.
[0184] Implementation process: The system retrieves 500 order data entries in batches from the order management system, real-time status data of 80 vehicles from the vehicle management system, and also obtains daily traffic forecast data and weather data.
[0185] Feature vectors were constructed and categorized for 500 orders: Level A urgent orders: 45; Level B regular orders: 320; Level C flexible orders: 135.
[0186] Feature modeling and availability assessment were performed on 80 vehicles. Of these, 75 passed the availability threshold check, while 5 were marked as unavailable because the drivers had reached their working hours or the vehicles were under maintenance.
[0187] Construct a 500×75 fit matrix, and reduce the candidate capacity for each order to the top 9 through candidate set filtering. The search space was reduced from 37,500 to 4,500.
[0188] Perform layered optimization: The first-level candidate set screening took approximately 0.3 seconds; The second-level hierarchical greedy matching takes about 0.5 seconds to generate an initial solution (matching rate 96.2%, average fit 0.72). The third-layer local search optimization iteration took 287 iterations and approximately 2.8 seconds. The optimized solution had a matching rate of 97.8% and an average fit of 0.81. Final solution: 489 orders were successfully matched with 73 vehicles, and 11 orders were placed into the manual processing queue due to special requirements. The average capacity utilization rate was 82.5%, the estimated delivery time fulfillment rate was 98.6%, and the empty-run rate was 12.3%.
[0189] Real-time dynamic matching scenarios Scenario Description: During delivery, the system continuously receives new orders and processes various dynamic events. Within a certain time period, the following events occur: 14:00 15 new urgent orders added.
[0190] At 14:15, vehicle V_023 experienced a malfunction, with 6 uncompleted orders on it.
[0191] Traffic control measures were implemented on the main roads in the city at 14:30, and the impact is expected to last for 2 hours.
[0192] Implementation process: 4:00 New order insertion processing: The system constructs feature vectors for 15 new orders, 3 of which are Class A emergency orders.
[0193] Search for available delivery capacity within a 30-kilometer radius of the pickup point for each new order.
[0194] Through incremental optimization algorithms, 12 orders were matched within 0.8 seconds, while 3 orders, after expanding the search scope, were matched within 2.1 seconds.
[0195] The matching process involved four delivery sequence rearrangements and two cross-capacity order migrations.
[0196] 14:15 Handling capacity issues: The system immediately retrieved 6 incomplete orders from V_023.
[0197] After sorting by urgency, alternative transportation capacity is sought in turn.
[0198] Four orders were assigned to three nearby vehicles, while two orders were upgraded to Grade A and assigned to the nearest available vehicle due to tight time windows.
[0199] The entire fault handling process took 1.5 seconds.
[0200] 14:30 Traffic Disruption Handling: The system identified that the routes of 8 transport vehicles passed through the controlled section.
[0201] The estimated arrival times for a total of 35 orders across these 8 vehicles were recalculated.
[0202] The time window constraint for 5 orders has been found to be violated.
[0203] Through route replanning and two cross-capacity migrations, the timeliness of four orders was guaranteed, and the customer was notified of the delay for one order and received confirmation.
[0204] The processing time was 2.3 seconds.
[0205] Peak-hour large-scale matching scenarios: Scenario description: During an e-commerce promotion, the system needs to match 5,000 orders with 300 delivery vehicles.
[0206] Implementation process: A parallel computing strategy is adopted to divide the orders into 8 sub-regions, and each sub-region independently performs candidate set screening and initial matching.
[0207] After sub-region matching is completed, cross-region optimization adjustments are made to orders at the sub-region boundaries.
[0208] The entire matching process took approximately 15 seconds, with a matching rate of 96.5% and a capacity utilization rate of 85.2%.
[0209] Example 2: A dynamic matching and optimization system for logistics orders and transportation capacity resources applies the method described in Example 1. It includes: a data preprocessing module, a feature construction and hierarchical evaluation module, a fitness matrix construction module, a matching scheme generation module, a dynamic incremental optimization module, and a scheme output module.
[0210] The data preprocessing module is used to collect raw data in real time and preprocess the raw data, which includes order data and transportation resource data. The feature construction and hierarchical evaluation module is used to construct order feature vectors and transportation resource feature vectors based on preprocessed order data and transportation resource data, and to classify orders based on order feature vectors and to score comprehensive availability based on transportation resource feature vectors. The fit matrix construction module is used to calculate multidimensional fit scores for order-capacity pairs that meet the constraints, calculate the comprehensive fit score through weighted summation, and construct the fit matrix based on the comprehensive fit score; The matching scheme generation module is used to generate an initial matching scheme based on the fitness matrix through candidate set filtering, hierarchical greedy matching, and local search optimization.
[0211] The dynamic incremental optimization module is used to detect dynamic changes in events when executing a matching scheme, to assess the scope of the event's impact, and to perform incremental optimization adjustments to update the matching scheme. The solution output module is used to output the updated matching solution.
[0212] It is understood that the system provided in the embodiments of the present invention corresponds to the method provided in the embodiments of the present invention, and the explanation, examples and beneficial effects of the relevant content can be referred to the corresponding parts of the above method.
[0213] Furthermore, it should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to include non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element.
[0214] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.
[0215] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as read-only memory (ROM) / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.
[0216] The above are merely preferred embodiments of the present invention and do not limit the scope of the patent. Any equivalent structural or procedural transformations made based on the description and drawings of the present invention, or direct or indirect applications in other related technical fields, are similarly included within the scope of patent protection of the present invention.
[0217] It is understood that the system provided in the embodiments of the present invention corresponds to the method provided in the embodiments of the present invention, and the explanation, examples and beneficial effects of the relevant content can be referred to the corresponding parts of the above methods.
Claims
1. A method for dynamically matching and optimizing logistics orders and transportation capacity resources, characterized in that, include: Real-time collection of raw data, including order data, transportation capacity data, and environmental data; The raw data is preprocessed; The preprocessed order data is used to construct order feature vectors and order classifications; Construct a capacity resource feature vector and an availability status assessment from the preprocessed capacity resource data; Calculate the matching degree between orders and transportation resources to construct a matching degree matrix; Based on the fitness matrix, an initial matching scheme is generated; The matching scheme is executed, and the matching scheme is updated based on dynamic change events during the execution process.
2. The method for dynamic matching and optimization of logistics orders and transportation capacity resources as described in claim 1, characterized in that, The preprocessing of the raw data includes the following steps: The raw data is cleaned. The original data is standardized using the following formula: in, These are the original eigenvalues; This represents the minimum value of the original eigenvalues. The maximum value of the original eigenvalues; The geographic information of the raw data is processed to convert the shipping address and receiving address into unified geographic coordinates; The time information of the original data is standardized to convert it into UTC timestamp format.
3. The method for dynamic matching and optimization of logistics orders and transportation capacity resources as described in claim 1, characterized in that, Constructing order feature vectors from the preprocessed order data includes the following steps: For each order Construct the order feature vector The formula used is: in, Geographic feature components; For cargo characteristic components; For time-sensitive characteristic components; Priority components; For special requirements; For customer characteristic components; The geographical feature components The formula used is: in, The longitude of the shipping address; The latitude and longitude of the shipping address; The longitude of the delivery address; The latitude and longitude of the delivery address; The transport direction angle is given by the formula: The straight-line distance between the shipping location and the receiving location is calculated using the following formula: in, The radius of the Earth; The cargo characteristic components The formula used is: in, Standardized weight; Standardized volume; For cargo density; Code the type of goods; The time-sensitivity feature component The formula used is: in, Remaining available time; This is an urgency index; Timeliness level; The urgency index The formula used is: in, To estimate the delivery time.
4. The method for dynamic matching and optimization of logistics orders and transportation capacity resources as described in claim 1, characterized in that, The order tiering includes the following steps: Based on the order feature vector, a comprehensive urgency score is calculated for each order using the following formula: in, , , , These are the weighting coefficients; Scoring will be provided for special requirements; This is the original priority value for the order; Customer level; Based on the comprehensive urgency score Orders are categorized into three levels: urgent orders, regular orders, and flexible orders.
5. The method for dynamic matching and optimization of logistics orders and transportation capacity resources as described in claim 1, characterized in that, The process of constructing a capacity resource feature vector from the preprocessed capacity resource data includes the following steps: For each transportation resource Construct a feature vector of transportation capacity resources The formula used is: in, Location features; Capacity characteristics; State characteristics; As a characteristic of ability; Characteristics of the driver; Route characteristics; The location features The formula used is: in, The longitude of the vehicle's current location; The latitude of the vehicle's current location; The vehicle's current location coding; The administrative region code for the vehicle; The capacity feature The formula used is: in, The remaining available load capacity is calculated using the following formula: in, This refers to the vehicle's maximum load capacity. This represents the current loaded weight. The available remaining volume is calculated using the following formula: in, This refers to the maximum volume of the vehicle. This represents the currently loaded volume; The formula for weight utilization rate is as follows: The formula for calculating physical exertion efficiency is as follows: The driver's characteristics The formula used is: in, The normalized value of the driver's overall score; This is a normalized value for the remaining available driving time; The fatigue index is expressed by the following formula: in, This represents the driver's current continuous driving time. The maximum continuous driving time stipulated by regulations; The route features The formula used is: in, This represents the number of orders currently allocated to this capacity. This is the estimated timestamp for completing all current delivery tasks; This is the main direction angle of the current route; The geographical area covered by the route.
6. The method for dynamic matching and optimization of logistics orders and transportation capacity resources as described in claim 1, characterized in that, The assessment of the available status includes the following steps: Calculate the overall availability score for each transportation capacity resource. The formula used is: in, , , , These are the weighting coefficients; Assess availability; Evaluate the driver's overall performance; Rate the remaining capacity; Rate the route flexibility; The remaining capacity score The formula used is: The route flexibility score The formula used is: ; in, This represents the maximum number of orders that the capacity can handle.
7. The method for dynamic matching and optimization of logistics orders and transportation capacity resources as described in claim 1, characterized in that, The construction of the fitness matrix includes the following steps: Perform constraint checks; Perform multi-dimensional adaptation calculation, which includes distance adaptation, direction adaptation, capacity adaptation, time adaptation, route coordination, and cost adaptation. Calculate the overall fit score; The fitness matrix is constructed based on the comprehensive fitness score.
8. The method for dynamic matching and optimization of logistics orders and transportation capacity resources as described in claim 1, characterized in that, The process of generating the initial matching scheme includes the following steps: For each order From the first of the fitness matrix Select the top of the fit ranking in the row Transportation capacity resources as a candidate set The formula used is: in, For orders With transport capacity The overall fit score; The value of K is determined based on the transportation capacity, using the following formula: in, This refers to the amount of available transport capacity. The upper limit of the candidate set; Based on the order level from highest to lowest, the optimal capacity is selected within each level based on the urgency score from highest to lowest to generate an initial plan; The initial scheme is optimized through a neighborhood search operation.
9. A dynamic matching and optimization system for logistics orders and transportation capacity resources, using the method described in any one of claims 1 to 8, comprising: Data preprocessing module, feature construction and hierarchical evaluation module, fitness matrix construction module, matching scheme generation module; The data preprocessing module is used to collect raw data in real time and preprocess the raw data, which includes order data and transportation resource data. The feature construction and hierarchical evaluation module is used to construct order feature vectors and transportation resource feature vectors based on preprocessed order data and transportation resource data, and to perform order hierarchical evaluation based on the order feature vectors and to perform comprehensive availability scoring based on the transportation resource feature vectors. The fit matrix construction module is used to calculate a multidimensional fit score for order-capacity pairs that meet the constraints, calculate a comprehensive fit score by weighted summation, and construct a fit matrix based on the comprehensive fit score. The matching scheme generation module is used to generate an initial matching scheme based on the fitness matrix through candidate set filtering, hierarchical greedy matching, and local search optimization.
10. The logistics order and transportation capacity resource dynamic matching optimization system as described in claim 9 further includes a dynamic incremental optimization module and a solution output module; The dynamic incremental optimization module is used to detect dynamic change events when executing the matching scheme, to assess the scope of the event's impact, and to perform incremental optimization adjustments to update the matching scheme. The scheme output module is used to output the updated matching scheme.
