Self-adaptive logistics task allocation method for multi-warehouse cooperation

Abstract

The invention discloses a self-adaptive logistics task allocation method for multi-warehouse cooperation, and particularly relates to the field of task allocation of logistics transportation, and the method comprises the steps: building a behavior offset map of a logistics warehouse in a logistics alliance, solving a path inertia function and a response rigid vector, and generating a response model baseline of the logistics warehouse under non-profit guidance; generating a path inertia superposition risk value according to fusion of the path repetition matching value and the response model baseline, and executing dynamic scheduling weight adjustment by taking the path inertia superposition risk value as a sorting intervention factor; and constructing a rejection chain structure and a rejection behavior frequency table based on the rejection behavior of each task, and forming a responsibility sensitive sorting function for sorting punishment. By constructing a multi-dimensional scheduling weight dynamic adjustment mechanism fusing warehouse natural response behaviors, path preference trends, rejection behavior records, strategy simulation degrees and performance expressions, the problems of scheduling imbalance and long-term delay of partial order transportation caused by tendentiousness order receiving of alliance warehouses are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of task allocation in logistics transportation. More specifically, the present invention relates to an adaptive logistics task allocation method for multi-warehouse collaboration. Background Art

[0002] In a logistics alliance composed of multiple brands, warehouses of different enterprises will jointly participate in the unified scheduling of orders. Although it seems to be resource sharing, problems are likely to occur in actual operation. We found in actual situations that each warehouse is more willing to grab orders with high profits and simple transportation, while being indifferent to tasks with low profits and complex transportation. The final result is that the overall distribution efficiency decreases, the platform scheduling becomes increasingly unbalanced, and even there are situations where orders are not taken over for a long time. Therefore, in an alliance composed of multiple enterprise warehouses, how to enable each warehouse to fairly share tasks is an urgent problem to be solved. Summary of the Invention

[0003] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides an adaptive logistics task allocation method for multi-warehouse collaboration. By constructing a multi-dimensional scheduling weight dynamic adjustment mechanism that integrates the natural response behavior, path preference trend, rejection behavior record, strategy imitation degree, and performance fulfillment of the warehouse, to solve the problems of unbalanced scheduling and long-term lag of some orders caused by the tendency of alliance warehouses to accept orders in the above-mentioned background art.

[0004] To achieve the above object, the present invention provides the following technical solution: An adaptive logistics task allocation method for multi-warehouse collaboration, including the following steps: S1: By constructing a behavior deviation map of logistics warehouses in the logistics alliance, solving the path inertia function and the response rigidity vector, and generating a response model baseline of the logistics warehouse under non-profit orientation. S2: According to the path repeatability matching value and the response model baseline, generate a path inertia superposition risk value, and use it as a sorting intervention factor to perform dynamic scheduling weight adjustment. S3: Based on the rejection behavior of each task, construct a rejection chain structure and a rejection behavior frequency table to form a responsibility-sensitive sorting function for sorting punishment. S4: Use the similarity of strategy feature vectors to construct a strategy imitation trend map between warehouses, calibrate the spread risk boundary through the diffusion coefficient, and perform sorting interruption intervention on the starting warehouse. S5: By integrating the performance fulfillment status, construct a responsibility performance vector, and update the initial sorting weight and the alliance reputation record based on the incentive adaptation weight to achieve two-way binding of responsibility fulfillment incentives.

[0005] In a preferred embodiment, S1 further includes: extracting historical order receiving record data of each warehouse from the logistics task log of the logistics alliance. The historical order receiving record data includes task path, task response delay, and inventory turnover record, and the historical order receiving record data serves as the basic dataset of warehouse behavior; Perform structured classification processing on the basic dataset of warehouse behavior, and sequentially solve the volatility of the path direction distribution, the spatial gradient characteristics of the inventory flux density field, and the change trend of the response frequency, and define them as the triple of warehouse behavior indicators; input the triple of warehouse behavior indicators into the graph construction process, and combine with the structural path of the task execution graph to establish a warehouse behavior deviation map, which is used to reflect the natural response differences of the warehouse to different types of tasks; Based on the node connection strength of the warehouse behavior deviation map, deduce the warehouse path inertia function, and combine with the response period distribution to form a response rigidity vector, which serves as the response model baseline of the warehouse under non-profit constraints.

[0006] In a preferred embodiment, S2 further includes: extracting the path trajectory set of each warehouse from the historical order receiving record data, and extracting the path attribute set of the current scheduling task from the currently to-be-allocated task set. Subsequently, perform path vectorization processing on the path trajectory set of the warehouse and the path set of the current scheduling task respectively, construct a standard vector structure representing the historical path and the current path direction, and calculate the path repetition matching value between each warehouse and the current scheduling task set through the path direction coincidence degree; statistically analyze the central tendency of path selection based on the distribution of the matching values of each warehouse, and finally generate a path aggregation density index representing the degree of risk of repeated path locking; Perform vector fusion on the path aggregation density index and the path inertia function in the response model baseline of the warehouse, and output the path inertia superposition risk value, which is used to measure the trend of repeated path locking; Inject the path inertia superposition risk value into the scheduling sorting model as a dynamic intervention factor, and perform weight mapping in combination with the current task value vector to generate an initial matching sorting table of tasks to warehouses.

[0007] In a preferred embodiment, S3 further includes: setting a unique task identifier for each to-be-allocated task, and accordingly establishing a corresponding rejection chain structure. Each node in the rejection chain structure records the warehouse number, response status, and response time corresponding to the task when it is rejected, constituting a warehouse response behavior trajectory for traceability; During the task allocation process, if scheduling fails or response times out, append the rejection behavior corresponding to the task to the rejection chain structure of the task in chronological order, and expand it in round order to form a rejection behavior linked list; Aggregate the rejection chain structures of all tasks according to the task identifiers, count the cumulative frequency of rejection behaviors with the warehouse number as the index, construct a frequency table of rejection behaviors for each warehouse, and use it to characterize the historical density characteristics of rejection responses; Determine the rejection impact factor for each warehouse based on the frequency table of rejection behaviors, and involve the rejection impact factor in the calculation of the scheduling sorting weight. By constructing a responsibility-sensitive sorting function, dynamically lower its sorting priority in the next round of task allocation to achieve punitive sorting adjustment.

[0008] In a preferred embodiment, S4 further includes: extracting the task information received by each warehouse within a similar scheduling period from the task allocation history, sequentially extracting three types of task attributes based on the profit interval, path direction attribute, and trigger response time period corresponding to each task, and encoding the three types of task attributes respectively to form a set of warehouse scheduling strategy feature vectors as the policy behavior expression of each warehouse in the current period; Use the set of warehouse scheduling strategy vectors as input to perform similarity comparison. Based on the similarity distribution between vectors, construct a strategy imitation trend map between warehouses, and mark the strategy trajectory paths with similarity scores exceeding the preset similarity score threshold in the strategy imitation trend map as high-risk imitation paths; Calculate the strategy diffusion coefficient for each high-risk imitation path. The strategy diffusion coefficient is determined based on the number of warehouses on the imitation path, the synchronization degree and duration of strategy updates, and mark the paths with strategy diffusion coefficients exceeding the intervention threshold as the spread risk boundary; Perform a strategy intervention action on the starting warehouse node in the spread risk boundary, set the corresponding scheduling punishment coefficient according to its imitation propagation level, and use the scheduling punishment coefficient in the process of adjusting the scheduling sorting weight to dynamically inhibit the trend of behavior pattern replication.

[0009] In a preferred embodiment, S5 further includes: after each scheduling period ends, extract three task completion attributes: response success rate, low-profit task completion ratio, and path selection diversity, based on the scheduling response results of the tasks completed by each warehouse within the corresponding scheduling period, and integrate the three task completion attributes by warehouse dimension to generate a responsibility performance vector; Use the responsibility performance vector as input data, perform corresponding mapping with the preset set of incentive factors, and calculate the incentive adaptation weight that each warehouse is allowed to obtain in the next scheduling period according to the weight settings of the incentive factors in the set of incentive factors on different performance dimensions; Inject the incentive adaptation weight as the initial sorting weight into the task scheduling sorting process of the next period, and write the incentive adaptation weight corresponding to each warehouse into the periodic credit record table to update the credit status information of the warehouse within the alliance.

[0010] The technical effects and advantages of the present invention: The present invention reflects the natural task response behavior of a warehouse under non-profit orientation by constructing a response model baseline that includes a path inertia function and a response rigidity vector, thereby suppressing the unbalanced tendency of preferentially pre-empting high-profit tasks in the joint transportation scheduling of multi-enterprise warehouses and ensuring fair distribution of low-profit tasks; The present invention dynamically identifies the degree of preference for repeated paths in the scheduling sorting by extracting the coincidence trend between the warehouse path trajectory and the current task path and generating a path inertia superposition risk value in combination with the response model, so as to achieve early intervention in the behavior of locking high-frequency paths and avoid excessive resource concentration; The present invention constructs a task rejection chain structure, counts the behavior trajectories and frequencies of task rejections, and deduces the rejection impact factor in combination with the task level and response delay, so as to realize the quantitative expression of the degree of warehouse responsibility fulfillment and perform a disciplinary adjustment of responsibility for negative response behaviors during the sorting process; By constructing a policy feature vector and analyzing its similarity distribution among warehouses, generating a policy imitation trend map, and constructing a spread risk boundary in combination with the strength of the policy diffusion path, the identification and intervention of the convergence trend of the inter-warehouse scheduling behavior are realized, and the diversity of responses is improved; By integrating the fulfillment results of the warehouse within a cycle, constructing a fulfillment vector based on the response success rate, the proportion of low-profit tasks, and path diversity, and introducing an incentive adaptation weight to update the sorting weight, the integration closed-loop of positive feedback on fulfillment and a dynamic incentive mechanism in task scheduling is realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 is a flowchart of the steps of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0012] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0013] Refer to the attached Figure 1 description. A method for adaptive logistics task allocation for multi-warehouse collaboration according to an embodiment of the present invention includes the following steps: S1: By constructing a behavior deviation map of logistics warehouses in a logistics alliance, solving out a path inertia function and a response rigidity vector, and generating a response model baseline of the logistics warehouse under non-profit orientation; S2: Fusing the path repetition matching value with the response model baseline to generate a path inertia superposition risk value, and using it as a sorting intervention factor to perform dynamic scheduling weight adjustment; S3: Based on the rejection behavior of each task, construct a rejection chain structure and a rejection behavior frequency table to form a responsibility-sensitive sorting function for sorting and punishment; S4: Use the similarity of policy feature vectors to construct a policy imitation trend map between warehouses, calibrate the spread risk boundary through the diffusion coefficient, and perform sorting interruption intervention on the starting warehouse; S5: Construct a responsibility performance vector by integrating the performance status, and update the initial sorting weight and the alliance reputation record based on the incentive adaptation weight to achieve two-way binding of responsibility performance incentives.

[0014] S1 also includes: Extract the historical order receiving record data of each warehouse from the logistics task log of the logistics alliance. The historical order receiving record data includes the task path, task response delay, and inventory turnover record, and the historical order receiving record data is used as the basic dataset of warehouse behavior; Perform structured classification processing on the basic dataset of warehouse behavior, and successively solve the volatility of the path direction distribution, the spatial gradient characteristics of the inventory flux density field, and the change trend of the response frequency, and define them as the triple of warehouse behavior indicators; Input the triple of warehouse behavior indicators into the graph construction process, and combine the structural path of the task execution graph to establish a warehouse behavior deviation map, which is used to reflect the natural response differences of warehouses to different types of tasks; Based on the node connection strength of the warehouse behavior deviation map, derive the warehouse path inertia function, and combine the response period distribution to form a response rigidity vector as the response model baseline of the warehouse under non-profit constraints.

[0015] It should be further noted that in the formula structure involved in this solution, the dimensionless term can be used as a proportional or structural adjustment factor. When combined with a quantity with a unit, it only plays a role in numerical scaling and does not introduce a new physical dimension. Therefore, it will not change or confuse the overall unit system of the expression; The combination of such "dimensionless term and quantity term with unit" can be understood as the composite structure expression form commonly used in mathematical and physical modeling, which conforms to the principle of dimensional consistency and has a clear physical interpretation basis; Secondly, in the formula structure of this solution, if there are multiple variable terms with different physical units, including but not limited to time-type, mass-type, or energy-type variables, their joint appearance is to express the collaborative modeling relationship of multiple physical mechanisms. Each variable forms a unified structure through function mapping, ratio combination, or normalization adjustment, with clear units and clear meanings. The overall expression conforms to the principle of dimensional consistency and the common norms of engineering modeling; In the solution based on "S1 also includes", it includes: extracting historical order acceptance record data from the logistics task log, successively solving the volatility of the path direction distribution, the spatial gradient characteristics of the inventory flux density field, and the response frequency change trend, and constructing a triple of warehouse behavior indicators; deriving the path inertia function and the response rigidity vector based on the behavior deviation map, and constructing a response model baseline under non-profit orientation: ; where is the response model baseline of the warehouse , [unit: ; is the path direction fluctuation function of the warehouse at time , [unit: number of path segments, segment]; is the behavior modeling observation period, [unit: hour]; is the inventory flux density function, [unit: ; is the inventory flux direction vector field, [unit: direction vector, dimensionless]; is the inventory operation area, [unit: ; is the response times function per unit time, unit: [times / hour]; is the path scale normalization constant, and the path scale normalization constant is defined as the maximum path segment number normalization factor, unit: [segment]; For the first item of the response model baseline is the path inertia term, [unit: , which is achieved by integrating the square of the second-order change rate of the path; The second item is the inventory flux intensity term, unit: [pieces / hour]; in order to maintain the same dimension as the path term, the normalization scale is used to cancel the inventory density dimension; The exponential part of the third item refers to the response frequency fluctuation term, dimensionless; The overall unit closed result in the response model baseline is: ; The formula construction of the response model baseline is used to extract the response tendency under non-profit conditions from the warehouse behavior. First, the second derivative of the path function reflects the stability of path selection, and the reciprocal of its integral represents the inertia intensity. The inventory gradient multiplied by the direction vector represents the local inventory flow intensity, and the overall activity is reflected after integration; in order to avoid unit mismatch, the inventory flux term introduces to normalize the path scale; the response frequency fluctuation term decays exponentially, making the response more unstable and the overall baseline lower; the finally output It represents the product effect of response ability, path inertia, and inventory activity, which is the baseline value used for risk suppression and capacity guidance in scheduling and sorting.

[0016] S2 also includes: extracting the set of path trajectories of each warehouse from the historical order receiving record data, and extracting the set of path attributes of the current scheduling task from the current task set to be assigned. Subsequently, vectorize the path of the warehouse's path trajectory set and the current scheduling task's path set respectively, construct a standard vector structure representing the historical path and the current path direction, and calculate the path repeatability matching value between each warehouse and the current scheduling task set through the path direction coincidence degree; statistically analyze the central tendency of path selection based on the distribution of the matching values of each warehouse, and finally generate a path aggregation density index representing the degree of path repeat lock risk; Fuse the path aggregation density index with the path inertia function in the response model baseline of the warehouse, and output the path inertia superposition risk value, which is used to measure the trend of repeated path locking; Inject the path inertia superposition risk value into the scheduling and sorting model as a dynamic intervention factor, and jointly perform weight mapping with the current task value vector to generate an initial matching sorting table from tasks to warehouses.

[0017] Based on the solution of "S2 also includes", it includes: extracting the historical path direction vector of each warehouse from the historical path direction field, calculating the direction coincidence integral and aggregation gradient between it and the current scheduling task path direction, and combining with the response model baseline in S1 to generate a path inertia superposition risk value as a sorting intervention factor: ; where is the path inertia superposition risk value, unit: ; is the historical path direction vector field, unit: [dimensionless direction vector]; is the current task path direction vector, unit: [dimensionless direction vector]; is the path definition space domain, unit: ; is the direction vector gradient tensor field, unit: [1 / m]; is generated by the formula in S1, unit: ; For in the formula, it should be noted that the first term is the normalized path direction coincidence integral term, with the dimension of [dimensionless]; the second term is the square root of the path gradient modulus integral, with the dimension of , which is dimensionless; the third term is , with the unit of ; the overall unit is ; Among them In the formula, the core is to comprehensively evaluate the coincidence trend of warehouse path selection and path distribution density, expressed as a risk intervention factor. The first item captures the path reuse trend by taking the dot product integral of two direction fields and normalizing. The second item calculates the square of the change rate of the path direction field and takes the root mean square to reflect the gradient intensity of path aggregation. The combination of the two expresses the path inertia intensity. Finally, it is multiplied by the response model baseline , to achieve the fusion judgment of response ability and path tendency, as Applied to the adjustment of scheduling sorting weights to achieve early intervention in path competition behavior.

[0018] S3 also includes: setting a unique task identifier for each task to be assigned, and accordingly establishing a corresponding rejection chain structure. Each node in the rejection chain structure records the warehouse number, response status, and response time corresponding to the task when it is rejected, forming a warehouse response behavior trajectory for traceability; During the task assignment process, if scheduling fails or response times out, append the rejection behavior corresponding to the task to the rejection chain structure of the task in chronological order, and expand it in round order to form a rejection behavior linked list; Aggregate the rejection chain structures of all tasks according to the task identifier, count the cumulative frequency of rejection behaviors with the warehouse number as the index, and construct a rejection behavior frequency table in units of warehouses to depict the historical density characteristics of rejection responses; Determine the rejection impact factor of each warehouse based on the rejection behavior frequency table, and participate in the calculation of the scheduling sorting weight of the rejection impact factor. Dynamically lower its sorting priority in the next round of task assignment by constructing a responsibility-sensitive sorting function to achieve punitive sorting adjustment.

[0019] Based on the solution of "S3 also includes", it includes: recording the rejection response behavior of the warehouse to the task during the scheduling process based on the rejection chain structure, and constructing a dynamically normalized rejection penalty integral by combining task scheduling urgency, task level, and scheduling density to generate the rejection impact factor of the warehouse: ; Among them is the rejection impact factor of warehouse , [unit: dimensionless]; is the duration of the scheduling evaluation period, [unit: hour]; is at time the number of rejected tasks of warehouse , [unit: times]; is the scheduling urgency factor of task , [unit: dimensionless, defined as the normalized value of the urgency level score]; is the task The cumulative waiting time before being responded to, [unit: hour]; is the task level number of the task, [unit: level number, dimensionless (integer value)]; is the reference item of the system task level mean, [unit: level number, dimensionless] For the components, the unit of is hour / hour = dimensionless; is dimensionless; has been normalized to dimensionless; The total amount of items is dimensionless, the unit of is hour, and the unit of the integration result is [hour]; Multiply the whole by ( ) and the time dimension is completely cancelled; The final result is dimensionless; The formula of is used to evaluate the influence intensity of the warehouse's rejection behavior of the scheduling task in the current cycle. By combining each rejection with the task urgency, response delay time and task level complexity to construct a penalty term, and accumulating it over time, and finally dividing by the cycle length to complete the normalization, the output is the rejection influence factor that can be directly called in the scheduling ranking .

[0020] S4 also includes: extracting the task information received by each warehouse in the similar scheduling cycle from the task assignment history, extracting three types of task attributes in sequence based on the profit interval, path direction attribute and trigger response time period corresponding to each task, and encoding the three types of task attributes respectively to form a set of warehouse scheduling strategy feature vectors as the expression of the strategy behavior of each warehouse in the current cycle; Taking the set of warehouse scheduling strategy vectors as the input to perform similarity comparison, constructing a strategy imitation trend map of warehouses based on the similarity distribution between vectors, and marking the strategy trajectory paths with similarity scores exceeding the preset similarity score threshold as high-risk imitation paths in the strategy imitation trend map; Calculating the strategy diffusion coefficient for each high-risk imitation path. The strategy diffusion coefficient is measured based on the number of warehouses on the imitation path, the synchronization degree and duration of strategy updates, and marking the paths with strategy diffusion coefficients exceeding the intervention threshold as the spread risk boundary; Performing a strategy intervention action on the starting warehouse node in the spread risk boundary, setting the corresponding scheduling disciplinary coefficient according to its imitation propagation level, and using the scheduling disciplinary coefficient in the process of adjusting the scheduling ranking weight to dynamically inhibit the trend of behavior pattern replication.

[0021] In the solution based on "S4 also includes", it includes: extracting the scheduling policy feature vectors of each warehouse within the scheduling period, calculating the policy diffusion intensity comprehensively through behavioral similarity, policy update speed, and structural behavior path distance, and outputting the policy diffusion coefficient: ; where is the policy diffusion coefficient of warehouse , [unit: dimensionless]; , is the scheduling policy feature vector of the warehouse (encoded by three types of task attributes: profit interval, path direction attribute, and trigger response time period), [unit: dimensionless vector]; represents the set of neighboring warehouses with highly similar scheduling behaviors to warehouse , is the number index of set ; is the policy evaluation time window, [unit: hour]; is the policy vector change rate, [unit: 1 / hour]; is the perturbation constant, [unit: dimensionless], and the perturbation constant is used to prevent the denominator from being zero; is the length of the behavioral structure path between warehouses, [unit: segment]; is the policy synchronization degree factor, [unit: dimensionless], and the policy synchronization degree factor represents the coincidence ratio of the scheduling policy update times of two warehouses; in the formula is the upper limit of integration, is the lower limit of integration, and the upper limit of integration represents the length of the time sliding window for evaluating the diffusion of scheduling policy behaviors, with the unit of hour, and is uniformly set as a constant for all warehouses; In the units of the above formula: The square of the vector difference is dimensionless; In the denominator is [1 / hour] × [hour] = dimensionless; is the path segment, and taking the reciprocal is ; is dimensionless; Therefore The unit result in the formula is: dimensionless × dimensionless × [1 / segment] × dimensionless = [1 / segment]; if further normalized (i.e., divided by the maximum path length), a dimensionless result is obtained; The formula objective of Used to determine whether to implement scheduling sorting interruption intervention for this warehouse.

[0022] S5 also includes: After each scheduling cycle ends, according to the scheduling response results of the tasks completed by each warehouse during the corresponding scheduling cycle, extract three task completion attributes: response success rate, low-profit task completion ratio, and path selection diversity, and integrate the three task completion attributes by warehouse dimension to generate a responsibility fulfillment vector; Take the responsibility fulfillment vector as input data, perform corresponding mapping with the preset set of incentive factors, and calculate the incentive adaptation weight that each warehouse is allowed to obtain in the next scheduling cycle according to the weight settings of the incentive factors in the set of incentive factors on different fulfillment dimensions; Inject the incentive adaptation weight into the task scheduling and sorting process of the next cycle as the initial sorting weight, and write the incentive adaptation weight corresponding to each warehouse into the cycle credit record table for updating the credit status information of the warehouse within the alliance.

[0023] Based on the solution of "S5 also includes", it includes: Combining the three indicators of the response success rate, low-profit task completion ratio, and path selection diversity of the warehouse's response to scheduling tasks, constructing a fulfillment vector through time integration, and normalizing to form a dimensionless incentive adaptation weight: ; Where is the incentive adaptation weight of the warehouse , [unit: dimensionless]; is the length of the scheduling evaluation cycle, [unit: hour]; is the task response success frequency per unit time, [unit: times / hour]; is the reference maximum response frequency, [unit: times / hour]; is the low-profit task completion ratio, [unit: dimensionless]; is the path diversity factor, [unit: dimensionless], and the path diversity factor includes path coverage; Taking the square root remains dimensionless; Composition description: The unit of the integrand is: times / hour × dimensionless × dimensionless = [times / hour]; after integration, it is: times; the denominator The unit is hour × times / hour = times; the final unit is: times / times = dimensionless; is the formula used to measure the true fulfillment quality of the warehouse during the current scheduling cycle; the three indicators constitute the core performance characteristics, and after convolution and time integration, a complete cycle score is formed; then it is normalized with the maximum success frequency to output a dimensionless incentive weight , which is used to affect the initial sorting weight of the next cycle and is also used for updating the warehouse credit factor synchronously.

[0024] Overall, it should be noted that in a logistics alliance composed of multiple brands, although the warehouse resources achieve apparent synergy, in the actual task scheduling practice, since each warehouse belongs to a different enterprise, its scheduling decision is restricted by its respective profit orientation, resulting in a structural imbalance in resource allocation. When the alliance platform faces the distribution of a large number of parallel tasks, there is often a situation where some warehouses actively prefer to receive orders with high profits and low complexity, while forming a "negative selection" or even "cold treatment" for orders with low profits or sub-optimal paths. The long-term existence of this phenomenon will make the task distribution show a deteriorating trend: high-value orders are crowded in competition, low-value orders are postponed for a long time, the overall scheduling system loses fairness and task acceptance stability, and the scheduling ranking lacks a feedback mechanism and cannot actively identify and intervene in behavioral imbalances. Therefore, in order to achieve fair performance and responsible order acceptance behavior of the warehouses within the alliance, this solution constructs a five-level progressive adaptive task allocation method consisting of response model construction, path risk suppression, behavior punishment, imitation trend intervention, and performance incentive feedback, so as to achieve the dual goals of improving the fairness of alliance scheduling and the overall response ability. The goal of S1 is to construct a response model baseline as the starting point of scheduling behavior. The S1 solution first collects the historical order acceptance behavior logs of each warehouse in the logistics alliance, extracts its task path characteristics, response delay records, and inventory turnover status, and constructs a basic warehouse behavior dataset based on this. After structured classification of this dataset, three core indicators, namely the volatility of the path direction distribution, the spatial gradient characteristics of the inventory flux density field, and the response frequency change trend, are respectively extracted, and then a triple of warehouse behavior indicators is constructed. By mapping the triple to the task execution graph, a warehouse behavior deviation map is generated to describe the natural response distribution of the warehouse to different types of tasks without profit orientation. On this basis, the path inertia function is extracted according to the second derivative of the path direction change, and the response rigidity vector is constructed in combination with the response frequency change. After the two are fused, the response model baseline of the warehouse is formed as a reference benchmark for the platform to evaluate the scheduling stability and behavioral naturalness of the warehouse. The goal of S2 is to fuse the response model based on path coincidence and output a risk intervention factor. The historical path direction field of the warehouse and the path direction of the current scheduling task are vectorized, and their direction coincidence integral and path aggregation density index are calculated to characterize whether there is a risk of "path inertia repeated locking" in the current task. Subsequently, this path repeatability index is fused with the path inertia function in the response model baseline to output the path inertia superposition risk value. This risk value is used as a dynamic intervention factor in the sorting and scheduling link to identify possible path monopoly behaviors of the warehouse in advance. It is jointly weighted with the task value vector to output the initial sorting weight to ensure that the warehouse does not continuously preempt the scheduling priority due to path matching advantages. The goal of S3 is to construct a rejection impact factor based on responsibility traceback to achieve sorted punishment; During the scheduling execution process, a rejection chain structure is constructed for each task, recording the time, status, and warehouse number when each task is rejected by a certain warehouse; these rejection events are appended by round, and finally a rejection behavior linked list is formed; after the scheduling cycle ends, the system counts the cumulative number of rejections for each warehouse, and calculates its rejection impact factor in combination with the scheduling urgency, response delay time, and task level of the task; this rejection impact factor is injected into the responsibility-sensitive sorting function as a punishment item for scheduling sorting, used to lower the scheduling priority of warehouses that frequently reject orders or evade responsibility, and achieve the scheduling punishment of behaviorally deviated warehouses; The goal of S4 is to imitate the behavior of warehouse strategies globally and achieve trend interruption intervention; To prevent the spread of the behavior of imitating high-yield strategies among warehouses, the platform extracts the strategy vector (three-dimensional attribute coding vector) expressed by the fulfillment of warehouse tasks in each scheduling cycle and constructs a strategy similarity map; if it is detected that some warehouses form a highly similar strategy path in a short period of time, it is judged as a "strategy imitation trend", and the strategy diffusion coefficient is calculated to calibrate the potential spread risk boundary; the system performs punitive sorting intervention on the starting propagation warehouse node, and by adjusting its scheduling weight, suppresses the excessive spread of the isomorphic scheduling mode between warehouses and restores the differential response ability between warehouses; The goal of S5 is to integrate the fulfillment results after the cycle ends and form an incentive adaptation weight feedback; After each scheduling cycle ends, the platform will quantitatively evaluate the fulfillment behaviors of all warehouses, extract three indicators: response success rate, low-profit task completion ratio, and path selection diversity, and construct a fulfillment vector; match this vector with the preset incentive factor set to form the corresponding incentive adaptation weight; this weight is injected into the scheduling engine as the initial factor for scheduling sorting in the next cycle, and at the same time updates the reputation record of the warehouse within the alliance, realizing the positive binding of scheduling incentives and responsibility fulfillment, and stimulating the enthusiasm of each warehouse to actively participate in all task types of scheduling; This solution is applicable to distributed warehousing and distribution platforms jointly operated by multiple brands and multiple enterprises, especially applicable to shared platform-based logistics organizations, such as regional joint distribution alliances, e-commerce platform warehousing and distribution integration systems, or cross-regional stock preparation and transfer collaborative networks; in such organizational structures, although the participating warehouses share the scheduling platform, due to inconsistent interests, systemic task biases or fulfillment inclinations occur frequently; traditional polling, weighted, or static scoring scheduling solutions cannot dynamically identify the changing trends of warehouse behaviors, nor can they effectively punish order rejection behaviors or inhibit the spread of strategy imitation; the five-level progressive mechanism proposed by the present invention realizes the regulation goals of scheduling fairness, response diversity, and behavioral differentiation between warehouses from response baseline modeling to rejection behavior punishment, then to strategy diffusion control and fulfillment incentive feedback.

[0025] 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 protection scope of the present invention.

Claims

1. An adaptive logistics task allocation method for multi-warehouse collaboration, characterized in that, it includes the following steps: S1: By constructing a behavior deviation map of logistics warehouses in the logistics alliance, solving the path inertia function and response rigidity vector, and generating a response model baseline of the logistics warehouse under non-profit orientation; S2: According to the path repeatability matching value and the response model baseline, generate a path inertia superposition risk value, and use it as a sorting intervention factor to perform dynamic scheduling weight adjustment; S3: Based on the rejection behavior of each task, construct a rejection chain structure and a rejection behavior frequency table to form a responsibility-sensitive sorting function for sorting punishment; S4: Use the similarity of policy feature vectors to construct a policy imitation trend map between warehouses, calibrate the spread risk boundary through the diffusion coefficient, and perform sorting interruption intervention on the starting warehouse; S5: By integrating the performance status, construct a responsibility performance vector, and update the initial sorting weight and the alliance reputation record based on the incentive adaptation weight to achieve two-way binding of responsibility and incentive.

2. The adaptive logistics task allocation method for multi-warehouse collaboration according to claim 1, characterized in that: S1 further includes: extracting the historical order receiving record data of each warehouse from the logistics task log of the logistics alliance. The historical order receiving record data includes task paths, task response time delays, and inventory turnover records, and the historical order receiving record data serves as the warehouse behavior basic data set; Perform structured classification processing on the warehouse behavior basic data set, and successively solve the volatility of the path direction distribution, the spatial gradient characteristics of the inventory flux density field, and the response frequency change trend, and define them as the triple of warehouse behavior indicators; input the triple of warehouse behavior indicators into the graph construction process, and combine the structural paths of the task execution graph to establish a warehouse behavior deviation map, which is used to reflect the natural response differences of warehouses to different types of tasks; Based on the node connection strength of the warehouse behavior deviation map, deduce the warehouse path inertia function, and combine the response period distribution to form a response rigidity vector as the response model baseline of the warehouse under non-profit constraints.

3. The adaptive logistics task allocation method for multi-warehouse collaboration according to claim 2, characterized in that: S2 further includes: extracting the path trajectory set of each warehouse from the historical order receiving record data, and extracting the path attribute set of the current scheduling task from the current task set to be allocated. Subsequently, perform path vectorization processing on the path trajectory set of the warehouse and the path set of the current scheduling task respectively, construct a standard vector structure representing the historical path and the current path direction, and calculate the path repeatability matching value between each warehouse and the current scheduling task set through the path direction coincidence degree; statistically analyze the path selection central tendency according to the matching value distribution of each warehouse, and finally generate a path aggregation density index representing the degree of path repeat lock risk; Fuse the path aggregation density index with the path inertia function in the response model baseline of the warehouse, and output the path inertia superposition risk value, which is used to measure the repeated path lock trend. Inject the path inertia superposition risk value as a dynamic intervention factor into the scheduling and sorting model, and perform weight mapping in combination with the current task value vector to generate an initial matching sorting table of tasks to warehouses.

4. An adaptive logistics task allocation method for multi-warehouse collaboration according to claim 3, characterized in that: S3 further includes: setting a unique task identifier for each task to be allocated, and accordingly establishing a corresponding rejection chain structure. Each node in the rejection chain structure records the warehouse number, response status, and response time corresponding to the task when it is rejected, constituting a warehouse response behavior trajectory for traceability; During the task allocation process, if scheduling fails or the response times out, append the rejection behavior corresponding to the task to the rejection chain structure of the task in chronological order, and expand it in round order to form a rejection behavior linked list; Aggregate the rejection chain structures of all tasks according to the task identifier, count the cumulative frequency of rejection behaviors with the warehouse number as the index, and construct a rejection behavior frequency table for each warehouse to characterize the historical density characteristics of rejection responses; Determine the rejection impact factor for each warehouse based on the rejection behavior frequency table, and involve the rejection impact factor in the calculation of the scheduling and sorting weight. Dynamically lower its sorting priority in the next round of task allocation by constructing a responsibility-sensitive sorting function to achieve disciplinary sorting adjustment.

5. An adaptive logistics task allocation method for multi-warehouse collaboration according to claim 4, characterized in that: S4 further includes: extracting the task information received by each warehouse within a similar scheduling cycle from the task allocation history, successively extracting three types of task attributes based on the profit interval, path direction attribute, and trigger response time period corresponding to each task, and encoding the three types of task attributes respectively to form a set of warehouse scheduling strategy feature vectors as the policy behavior expression of each warehouse in the current cycle; Perform similarity comparison with the set of warehouse scheduling strategy vectors as the input, construct a strategy imitation trend map between warehouses based on the similarity distribution between vectors, and mark the strategy trajectory paths with similarity scores exceeding the preset similarity score threshold in the strategy imitation trend map as high-risk imitation paths; Calculate the strategy diffusion coefficient for each high-risk imitation path. The strategy diffusion coefficient is determined based on the number of warehouses on the imitation path, the synchronization degree and duration of strategy updates, and mark the paths with strategy diffusion coefficients exceeding the intervention threshold as the spread risk boundary; Perform a strategy intervention action on the starting warehouse node in the spread risk boundary, set a corresponding scheduling disciplinary coefficient according to its imitation propagation level, and use the scheduling disciplinary coefficient in the process of adjusting the scheduling and sorting weight to dynamically suppress the trend of behavior pattern replication.

6. An adaptive logistics task allocation method for multi-warehouse collaboration according to claim 5, characterized in that: S5 further includes: after each scheduling cycle ends, extract three task completion attributes: response success rate, low-profit task completion ratio, and path selection diversity, based on the scheduling response results of the tasks completed by each warehouse within the corresponding scheduling cycle, and integrate the three task completion attributes in the warehouse dimension to generate a responsibility fulfillment vector; Using the responsibility fulfillment vector as input data, perform corresponding mapping with a preset set of incentive factors, and calculate the incentive adaptation weight that each warehouse is allowed to obtain in the next scheduling cycle according to the weight settings of the incentive factors in the set of incentive factors on different fulfillment dimensions; Inject the incentive adaptation weight into the task scheduling and sorting process of the next cycle as the initial sorting weight, and write the incentive adaptation weight corresponding to each warehouse into the cycle reputation record table for updating the reputation status information of the warehouse within the alliance.

Images