CN117332995A

CN117332995A - Narrow-channel blocking avoidance-based picking order allocation planning method, device and medium

Info

Publication number: CN117332995A
Application number: CN202311628622.4A
Authority: CN
Inventors: 屈挺; 李珊珊; 潘扬华; 张俊; 黄国全
Original assignee: Super Communications Co ltd; Jinan University
Current assignee: Super Communications Co ltd; Jinan University
Priority date: 2023-12-01
Filing date: 2023-12-01
Publication date: 2024-01-02
Anticipated expiration: 2043-12-01
Also published as: CN117332995B

Abstract

The application relates to a narrow-channel blocking avoidance-based picking order allocation planning method, a narrow-channel blocking avoidance-based picking order allocation planning device and a narrow-channel blocking avoidance-based picking order allocation planning medium, wherein the method comprises the following steps: in order picking process based on the first picking planning information, detecting real-time running state information corresponding to the target warehouse, and acquiring first running state information corresponding to the target warehouse from the first picking planning information; determining whether the real-time running state information and the first running state information have deviation, and determining a target correction strategy according to initial channel congestion state information corresponding to a warehouse channel under the condition that the deviation exists; acquiring correction running state information from correction picking planning information, and determining first warehouse congestion state information corresponding to the correction running state information; and carrying out re-planning updating on the corrected picking planning information according to the target picking correction planning according to the first warehouse congestion state information, and carrying out allocation planning on the target picking order according to the target picking planning information generated by the re-planning updating.

Description

Narrow-channel blocking avoidance-based picking order allocation planning method, device and medium

Technical Field

The application relates to the technical field of computer intelligent application, in particular to a picking order allocation planning method, a device and a medium based on narrow-channel blocking avoidance.

Background

In the related art, in order to realize the high area utilization rate and the high picking efficiency of the warehouse at the same time, the problem of narrow channel congestion in the picking process needs to be solved; current e-commerce warehouse picking planning often presents three challenges: the method comprises the steps of difficulty in order batch, disorder of picking paths and difficulty in solving narrow-channel congestion, wherein the difficulty in order batch refers to difficulty in considering a plurality of factors in an order batch method based on manual experience, and the factors comprise dynamics of order arrival time, diversification of order deadline, difference of picking paths caused by different batch results and real-time state of narrow-channel congestion of a warehouse; the disorder of the picking path means that the non-planned picking path is easy to generate the problem of transportation waste of repeatedly traversing the goods channel, and the probability of occurrence of narrow-channel congestion is increased; narrow-channel congestion resolution is difficult to mean that after congestion problems occur, the processing is generally not timely, and a better solution is difficult to quickly obtain so as to guide related workers to work; meanwhile, the e-commerce warehouse picking planning method or technology is usually only aimed at optimizing one link of picking batch or path planning, and the problems of disordered picking planning, labor cost, increased order fulfillment cost and the like caused by narrow channel congestion are ignored.

At present, aiming at the problems that narrow channel congestion is ignored, and the sorting planning disorder, the labor cost and the order fulfillment cost are easily caused by the e-commerce warehouse sorting planning method in the related technology, no effective solution is proposed.

Disclosure of Invention

The embodiment of the application provides a picking order allocation planning method, a device and a medium based on narrow channel blocking avoidance, which at least solve the problems that in the related art, the E-commerce warehouse picking planning method ignores narrow channel blocking, so that the picking planning is messy, the labor cost and the order fulfillment cost are increased.

In a first aspect, an embodiment of the present application provides a method for distributing and planning a picking order based on narrow-channel blocking avoidance, including: in order picking process based on first picking planning information, detecting corresponding real-time running state information in a target warehouse, and acquiring first running state information corresponding to the target warehouse from the first picking planning information, wherein the first picking planning information is generated by periodically picking allocation processing of multi-scale fusion data corresponding to a target picking order by using a pre-planning model, the pre-planning model is constructed based on a preset optimizing strategy target and an improved nested genetic algorithm, the picking allocation processing comprises picking batch and path planning, and a corresponding warehouse channel is formed in the target warehouse by deploying narrow channel shelves; determining whether the real-time running state information and the first running state information have deviation, and determining a target correction strategy according to initial channel congestion state information corresponding to the warehouse channel under the condition that the deviation exists, wherein the target correction strategy comprises a target picking correction plan; acquiring correction operation state information from correction picking planning information, and determining first warehouse congestion state information corresponding to the correction operation state information, wherein the correction picking planning information is generated by carrying out target picking correction planning on the real-time operation state information and the initial channel congestion state information; and carrying out rescheduling updating on the corrected picking planning information according to the first warehouse congestion state information and carrying out distribution planning on target picking orders according to target picking planning information generated by rescheduling updating, wherein when order picking is carried out based on the target picking planning information, the target warehouse releases warehouse channel congestion.

In a second aspect, an embodiment of the present application provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the steps of the narrow channel unblocking-based pick order allocation planning method according to the first aspect are implemented when the computer program is executed by the processor.

In a third aspect, embodiments of the present application provide a storage medium having stored thereon a computer program which, when executed by a processor, implements a narrow channel unblocking-based pick order allocation planning method as described in the first aspect above.

Compared with the related art, the narrow-channel blocking-avoidance-based picking order allocation planning method, device and medium provided by the embodiment of the application detect real-time running state information corresponding to a target warehouse in order picking process based on first picking planning information, and acquire first running state information corresponding to the target warehouse from the first picking planning information, wherein the first picking planning information is generated by periodically picking allocation processing of multi-scale fusion data corresponding to the target picking order by using a pre-planning model, the pre-planning model is constructed based on a preset optimization strategy target and an improved nested genetic algorithm, the picking allocation processing comprises picking batch and path planning, and the target warehouse forms a corresponding warehouse channel by deploying narrow-channel shelves; determining whether the real-time running state information and the first running state information have deviation, and determining a target correction strategy according to initial channel congestion state information corresponding to the warehouse channel under the condition that the deviation exists, wherein the target correction strategy comprises a target picking correction plan; acquiring correction operation state information from correction picking planning information, and determining first warehouse congestion state information corresponding to the correction operation state information, wherein the correction picking planning information is generated by carrying out target picking correction planning on the real-time operation state information and the initial channel congestion state information; according to the first warehouse congestion state information, re-planning updating is carried out on the modified picking planning information according to the target picking correction planning, and distribution planning is carried out on target picking orders according to target picking planning information generated by the re-planning updating, wherein when order picking is carried out based on the target picking planning information, the target warehouse removes warehouse channel congestion, the problem that narrow channel congestion is ignored by an e-commerce warehouse picking planning method in the related art, the problem that picking planning is messy, labor cost and order fulfillment cost are easy to increase is solved, and the beneficial effects of improving the smoothness of warehouse picking operation and reducing the labor and picking operation cost are achieved by periodically pre-planning and dynamically carrying out real-time planning correction on the picking orders and executing a narrow channel blocking avoidance mechanism.

Drawings

The accompanying drawings, which are included to provide a further understanding of the application and are incorporated in and constitute a part of this application, illustrate embodiments of the application and together with the description serve to explain the application and do not constitute an undue limitation to the application. In the drawings:

FIG. 1 is a hardware block diagram of a terminal of a narrow channel unblocking-based pick order allocation planning method according to an embodiment of the present application; FIG. 2 is a flow chart of a pick order allocation planning method based on narrow channel unblocking in accordance with an embodiment of the present application; FIG. 3 is a schematic diagram of a digital twin system architecture for intelligent warehouse picking planning in accordance with an embodiment of the present application; FIG. 4 is a flow chart of a pick order allocation planning method based on narrow channel unblocking in accordance with a preferred embodiment of the present application; FIG. 5 is a framework diagram of an improved nested genetic algorithm according to a preferred embodiment of the present application; FIG. 6 is a schematic diagram of a selection operator operation process according to a preferred embodiment of the present application; FIG. 7 is a schematic diagram one of a crossover operator operation process according to a preferred embodiment of the present application; FIG. 8 is a schematic diagram two of a crossover operator operation process according to a preferred embodiment of the present application; FIG. 9 is a schematic diagram III of a crossover operator operation process according to a preferred embodiment of the present application; FIG. 10 is a flow chart for solving an optimal picking path according to a preferred embodiment of the present application; FIG. 11 is a warehouse aisle layout schematic and key feature diagram in accordance with a preferred embodiment of the present application; FIG. 12 is a schematic diagram of a best path sub-graph in accordance with a preferred embodiment of the present application; FIG. 13 is a schematic view of a path within a warehouse aisle located in a warehouse setting area; FIG. 14 is a flow chart of narrow channel analysis and path adjustment according to a preferred embodiment of the present application; FIG. 15 is a schematic diagram of a sort sub-path having a spatial conflict; FIG. 16 is a block diagram of a narrow channel unblocking-based pick order allocation planning device in accordance with an embodiment of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described and illustrated below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the present application. All other embodiments, which can be made by one of ordinary skill in the art without undue burden on the person of ordinary skill in the art based on the embodiments provided herein, are intended to be within the scope of the present application. Moreover, it should be appreciated that while such a development effort might be complex and lengthy, it would nevertheless be a routine undertaking of design, fabrication, or manufacture for those of ordinary skill having the benefit of this disclosure, and thus should not be construed as having the benefit of this disclosure.

The method embodiment provided in this embodiment may be executed in a terminal, a computer or a similar computing device. Taking the example of running on a terminal, fig. 1 is a hardware structural block diagram of the terminal of the picking order allocation planning method based on narrow channel blocking avoidance according to the embodiment of the present application. As shown in fig. 1, the terminal may include one or more processors 102 (only one is shown in fig. 1) (the processor 102 may include, but is not limited to, a microprocessor MCU or a processing device such as a programmable logic device FPGA) and a memory 104 for storing data, and optionally, a transmission device 106 for communication functions and an input-output device 108. It will be appreciated by those skilled in the art that the structure shown in fig. 1 is merely illustrative and not limiting on the structure of the terminal described above. For example, the terminal may also include more or fewer components than shown in fig. 1, or have a different configuration than shown in fig. 1.

The memory 104 may be used to store computer programs, such as software programs and modules of application software, such as computer programs corresponding to the narrow channel blocking-based picking order allocation planning method in the embodiment of the present invention, and the processor 102 executes the computer programs stored in the memory 104 to perform various functional applications and data processing, that is, implement the method described above. Memory 104 may include high-speed random access memory, and may also include non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include memory remotely located relative to the processor 102, which may be connected to the terminal via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.

The transmission device 106 is used to receive or transmit data via a network. The specific example of the network described above may include a wireless network provided by a communication provider of the terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, simply referred to as NIC) that can connect to other network devices through a base station to communicate with the internet. In one example, the transmission device 106 may be a Radio Frequency (RF) module, which is configured to communicate with the internet wirelessly.

The embodiment provides a picking order allocation planning method based on narrow channel blocking avoidance, which is operated at the terminal, and fig. 2 is a flowchart of a picking order allocation planning method based on narrow channel blocking avoidance according to an embodiment of the application, as shown in fig. 2, and the flowchart includes the following steps:

step S201, in order picking process based on first picking planning information, detecting corresponding real-time running state information in a target warehouse, and acquiring first running state information corresponding to the target warehouse from the first picking planning information, wherein the first picking planning information is generated by periodically picking allocation processing of multi-scale fusion data corresponding to the target picking order by utilizing a pre-planning model, the pre-planning model is constructed based on a preset optimization strategy target and an improved nested genetic algorithm, the picking allocation processing comprises picking batch and path planning, and a corresponding warehouse channel is formed in the target warehouse by deploying narrow channel shelves.

In this embodiment, the execution body of the picking order allocation planning method of the embodiment of the application is an intelligent warehouse digital twin system which is deployed in a terminal, a server and a cloud server and is suitable for dynamic planning of warehouse picking, the intelligent warehouse digital twin system acquires real-time information (such as identity information, position information and action signals of objects such as staff, forklifts and the like) of a picking operation process according to sensing equipment deployed in a target warehouse, stores the real-time information as real-time data in a database of a virtual layer, uploads the real-time information of the warehouse picking operation to the virtual layer through communication equipment and a communication protocol, and the virtual layer carries out an example simulation and linkage decision on the picking system through the real-time operation data uploaded to the physical layer and downloads an optimization result to the physical layer to guide worker operation.

In the embodiment, picking pre-planning is periodically performed on orders in an order pool, wherein the periodicity refers to a time period, such as picking pre-planning is performed once every 15 minutes, in a pre-planning stage, a digital twin system builds a warehouse picking global multi-link pre-planning model based on full-flow real-time multi-scale fusion data by adopting an improved nested genetic algorithm, and the pre-planning model performs picking batch, task allocation and path planning multi-link initial plan formulation under preset system optimization target guidance and preset optimization rule constraint; in this embodiment, the whole-flow real-time multi-scale fusion data includes all order information in the order pool, the task execution progress of the picking workers, the future walking track and the warehouse congestion condition; in this embodiment, pre-planning takes order delay cost reduction and congestion occurrence rate reduction as optimization targets, so as to achieve the effect of preventing channel congestion.

In this embodiment, after the pre-planning is completed, by executing the picking scheme obtained in the pre-planning, that is, the first picking planning information, the relevant picking object is arranged to start picking goods in the target warehouse according to the planning result, then, the picking operation state in the target warehouse is monitored in real time, and the actual picking operation state (corresponding to the real-time operation state information) is compared with the system pre-planning optimization state (corresponding to the first operation state information), so as to determine whether the system disturbance occurs in the target warehouse according to whether the deviation occurs between the actual operation state and the system pre-planning optimization state and whether the warehouse channel congestion occurs beyond the planning.

Step S202, determining whether deviation exists between the real-time running state information and the first running state information, and determining a target correction strategy according to the initial channel congestion state information corresponding to the warehouse channel under the condition that the deviation exists, wherein the target correction strategy comprises a target picking correction plan.

In this embodiment, when judging whether a system disturbance occurs according to an actual operation state and a system pre-planning optimization state, firstly judging whether a deviation occurs between real-time operation state information and first operation state information, when the deviation occurs, and if the congestion of a warehouse channel in a target warehouse is higher than a specified parameter value, the system disturbance of a corresponding level occurs in the system, and then performing correction planning by adopting a target correction strategy corresponding to the system disturbance of the occurred level so as to realize alleviation or solution of a narrow channel congestion condition; in this embodiment, the intelligent warehouse digital twin system acquires real-time operation state information during warehouse picking operation and initial channel congestion state information corresponding to a warehouse channel, evaluates congestion index values of narrow channel congestion points based on the acquired real-time operation data, and adopts different correction planning strategies according to the different index values.

In this embodiment, the corresponding target correction strategies include four types, which are respectively: internal adjustment, multiparty linkage, resource linkage and demand linkage, wherein the internal adjustment refers to the selection planning of only one selecting object, and comprises a selection path or task execution sequence or selection waiting time so as to relieve the narrow-channel congestion condition; multiparty linkage refers to the fact that the relieving of the narrow-channel congestion condition requires simultaneous coordination of picking paths of a plurality of picking objects, the corresponding target correction strategy aims at minimizing the number of the picking objects involved, namely, as much as possible, adjusting the picking plans of fewer picking objects and realizing the solving of the congestion problem, and resource linkage refers to the fact that the narrow-channel congestion problem in the system cannot be effectively relieved by simply adjusting the picking task plans being executed, only external resources can be called, or the number of staff walking in a warehouse and picking equipment is reduced; demand linkage refers to modifying the demand of an order that is performing a task, primarily to order deadlines, by which the pressure of warehouse picking operations is reduced by relaxation of the demand.

Step S203, obtaining corrected operation state information from the corrected picking planning information, and determining first warehouse congestion state information corresponding to the corrected operation state information, wherein the corrected picking planning information is generated by performing target picking correction planning on the real-time operation state information and the initial channel congestion state information.

In this embodiment, after determining the target correction strategy, a corresponding picking correction plan or a corresponding correction plan model is constructed by using a global planning idea and an interference management idea, and based on the result of the pre-planning, a correction plan (corresponding to correction picking plan information) with the minimum system disturbance (the system disturbance appears as a narrow channel congestion degree and a picking cost deviation degree) is generated according to the target picking correction plan corresponding to the target correction strategy, so as to solve the channel congestion problem; in this embodiment, the idea of interference management is to emphasize real-time adjustment of the original plan, and to build a corresponding interference management model with the goal of minimizing the deviation, and to timely provide an adjustment plan for handling the interference event.

And S204, carrying out reprofiling updating on the corrected picking planning information according to the first warehouse congestion state information and carrying out distribution planning on the target picking order according to the target picking planning information generated by the reprofiling updating, wherein the warehouse channel congestion is relieved by the target warehouse when order picking is carried out based on the target picking planning information.

In this embodiment, after correction planning is executed, the corresponding picking object is instructed to perform order picking operation in the target warehouse according to correction picking planning information generated by the correction planning, meanwhile, whether the warehouse channel congestion condition in the target warehouse is relieved after the correction planning is implemented is judged, if not, the correction planning is repeated according to the channel congestion condition (corresponding to the first warehouse congestion state information) and the real-time running state information (monitored by the sensing equipment) in the target warehouse after the previous correction planning is completed, and the channel congestion condition after the repeated correction planning is verified until the congestion is relieved; and after judging that the congestion is relieved, continuously monitoring the real-time running state information in the target warehouse, and performing periodical pre-planning.

Through the steps S201 to S204, in the order picking process based on the first picking planning information, real-time operation state information corresponding to the target warehouse is detected, and first operation state information corresponding to the target warehouse is obtained from the first picking planning information, wherein the first picking planning information is generated by periodically picking and distributing multi-scale fusion data corresponding to the target picking order by using a pre-planning model, the pre-planning model is constructed based on a preset optimizing strategy target and an improved nested genetic algorithm, the picking and distributing process comprises picking batch and path planning, and a corresponding warehouse channel is formed in the target warehouse by deploying narrow channel shelves; determining whether deviation exists between the real-time running state information and the first running state information, and determining a target correction strategy according to initial channel congestion state information corresponding to a warehouse channel under the condition that the deviation exists, wherein the target correction strategy comprises a target picking correction plan; acquiring correction operation state information from correction picking planning information, and determining first warehouse congestion state information corresponding to the correction operation state information, wherein the correction picking planning information is generated by carrying out target picking correction planning on real-time operation state information and initial channel congestion state information; according to the first warehouse congestion state information, re-planning and updating modified picking planning information according to target picking correction planning, and carrying out distribution planning on target picking orders according to target picking planning information generated by the re-planning and updating, wherein when order picking is carried out based on the target picking planning information, a target warehouse releases warehouse channel congestion, the problem that narrow channel congestion is ignored by an e-commerce warehouse picking planning method in the related art, picking planning disorder, labor cost and order fulfillment cost are easy to cause is solved, and the beneficial effects of improving warehouse picking operation smoothness and reducing labor and picking operation cost are achieved by periodically pre-planning and dynamically planning and correcting the picking orders in real time and executing a narrow channel blocking avoidance mechanism.

It should be noted that the pre-planning and correction rules performed in this embodiment are planning for picking a target pick order, and it is understood that picking a target pick order is also planning for assigning a pick order.

In some embodiments, the determining the target correction policy in step S202 according to the initial warehouse channel congestion status information corresponding to the warehouse channel is implemented by the following steps:

and step 21, detecting a first congestion rate corresponding to a congestion detection node positioned on the warehouse channel in the initial warehouse channel congestion state information, wherein the first congestion rate is used for representing the congestion degree of the corresponding warehouse channel.

In this embodiment, a corresponding warehouse aisle is formed in the target warehouse due to deployment of the narrow aisle racks, so that a picking object (a picker) picks goods placed on the narrow aisle racks, and the picking object picks orders along the warehouse aisle, so that aisle congestion can be generated.

Step 22, calculating first total congestion rates corresponding to all warehouse channels in the target warehouse based on the first congestion rates, and determining channel congestion levels in the target warehouse according to the first total congestion rates.

In this embodiment, the current congestion condition of the target warehouse and the current channel congestion level of the target warehouse are further determined by determining the congestion degree of each warehouse channel.

Step 23, acquiring a target picking correction plan from preset picking correction plans based on the channel congestion level, wherein the picking correction plan at least comprises one of the following steps: a single picking object picking plan adjustment, a linkage adjustment of the picking plans of a plurality of picking objects, the addition and deletion of the picking plans, and the correction of the first picking plan information.

In this embodiment, the preset picking correction plans include an internal adjustment correction plan, a multiparty linkage correction plan, a resource linkage correction plan, and a demand linkage correction plan, and each picking correction plan corresponds to a correction mode, that is, corresponds to a picking correction scheme, for example: the internal adjustment correction planning is to adjust the picking path, the task execution sequence or the picking waiting time of one picking object, and after determining the target picking correction planning according to the channel congestion level, the correction plan which enables the minimum narrow channel congestion degree and the minimum picking cost deviation in the target warehouse is formulated through the correction mode corresponding to the target picking correction planning.

Through the steps 21 to 23, the target picking correction plan is determined, and based on the result of the pre-planning, a correction plan is generated to minimize the congestion degree of the narrow channel in the target warehouse and the deviation of the picking cost, so as to solve the problem of channel congestion, improve the smoothness of warehouse picking operation, and reduce the labor and picking operation cost.

In some of these embodiments, the objective picking correction planning of the real-time running status information and the initial warehouse aisle congestion status information in step S203 includes the steps of:

step 31, detecting first pick sub-plan information from the real-time running state information, wherein the first pick sub-plan information comprises a first pick object and a first pick sub-plan.

In this embodiment, the first picking sub-schedule information is schedule information of a corresponding first picking object performing a picking operation, where the first picking sub-schedule information includes a unique identifier or number of goods that the first picking object needs to pick, a required number, a storage location in a target warehouse, and an order picking deadline.

And step 32, determining a target picking correction plan according to the congestion state information of the initial warehouse channel.

In this embodiment, according to the initial warehouse aisle congestion status information, a target picking correction plan is determined, and a picking correction plan for correcting the first picking sub-plan information corresponding to the first picking object is determined, for example: for the current first pick sub-schedule information, the pick path, the task execution sequence or the pick waiting time is adjusted.

And step 33, carrying out correction processing on the first picking sub-plan based on the target picking correction plan to generate a candidate picking sub-plan, wherein the correction processing comprises adjustment of the first picking sub-plan and/or adjustment of the first picking object, and the candidate picking sub-plan comprises a second picking object after adjustment of the first picking object and a second picking sub-plan after adjustment of the first picking object.

In this embodiment, since the correction schemes corresponding to the target picking correction plans are different, when the correction process is performed, not only adjustment of the picking path, the order of task execution, or the picking waiting time corresponding to the first picking sub-plan, but also adjustment of the picking object, for example: when the target picking rule is a multiparty linkage correction plan, the picking paths of a plurality of picking objects are linked, the picking path i of the picking object a is adjusted to the picking path i of the picking object b, and in some other embodiments, the picking of the picking objects and the corresponding picking paths in the target warehouse is reduced, or the task arrangement in the first picking plan information of the pre-plan is corrected, for example: and modifying the deadline of the order, so that the corresponding picking path, task allocation, task execution sequence and picking waiting time are correspondingly modified, thereby realizing the improvement of the congestion condition in the current target warehouse.

And step 34, determining a deviation value of the first cable casting cost corresponding to each candidate picking sub-plan and a preset cost threshold value, determining a second congestion rate of a warehouse channel corresponding to each candidate picking sub-plan, and selecting the candidate picking sub-plan with the minimum deviation value and the lowest second congestion rate to obtain modified picking planning information, wherein the modified picking planning information comprises the candidate picking sub-plans with the minimum deviation value and the lowest second congestion rate.

In this embodiment, after correction planning is performed on the first picking sub-plan, order picking operation is performed in the target warehouse according to a candidate picking sub-plan generated by the correction planning and corresponding to the picking object, and whether the congestion condition of the warehouse channel where the picking object is located is relieved after the execution of the candidate picking sub-plan executed by the corresponding picking object is judged, so as to determine the target picking sub-plan corresponding to the picking object, and after the determination of the target picking sub-plan of the picking object currently executing the picking operation is completed, correction picking planning information is obtained.

Through the steps 31 to 34, correction of the picking planning information corresponding to the single picking object is realized, so that global correction of global picking planning information is further realized, and a correction plan which minimizes the narrow-channel congestion degree and the picking cost deviation in the target warehouse is generated based on the pre-planning result, so as to solve the channel congestion problem.

In some embodiments, the step S203 of rescheduling and updating the corrected picking plan information according to the current warehouse congestion status information and the target picking correction plan includes the following steps:

and step 41, detecting a third congestion rate corresponding to a congestion detection node positioned on the warehouse channel in the first warehouse congestion state information, wherein the third congestion rate is used for representing the congestion degree of the warehouse channel when the picking operation is performed according to the corrected picking planning information.

In this embodiment, the sensing device (e.g., a camera) disposed at a position (congestion detection node) corresponding to the warehouse aisle detects the flowing and walking condition of the picking object in the warehouse aisle to determine that the congestion condition in the warehouse aisle corresponds to the time of picking operation according to the modified picking planning information, and in this embodiment, the third congestion rate refers to the current congestion degree of the warehouse aisle and does not indicate that there are multiple congestion rates.

And 42, calculating third total congestion rates corresponding to all warehouse channels in the target warehouse based on the third congestion rates, and judging whether the target warehouse relieves the warehouse channel congestion according to the third total congestion rates.

In this embodiment, the current congestion condition of the target warehouse is determined by determining the congestion degree of each warehouse aisle when the picking operation is performed according to the corrected picking planning information.

In step 43, determining the target pick plan information includes correcting the pick plan information if it is determined that the target warehouse is free of warehouse aisle congestion.

In this embodiment, if it is determined that the target warehouse is not congested, it indicates that the picking operation is guided according to the modified picking planning information, and the congestion problem in the target warehouse can be solved, where the modified picking planning information meets the requirement of the picking operation in the target warehouse.

And step 44, correcting and planning the corrected picking and planning information according to the target picking and correcting planning until the target picking and planning information is generated under the condition that the warehouse channel congestion is not relieved by the target warehouse.

In this embodiment, if it is determined that the target warehouse is still congested, it indicates that the picking operation is guided according to the modified picking planning information, and the congestion in the target warehouse is not relieved, and the modification needs to be continued according to the modified picking planning information until the picking planning information meeting the requirement of the picking operation in the target warehouse is generated.

Through the steps 41 to 44, the judgment of whether the generated modified picking planning information meets the picking operation requirement in the target warehouse is realized, so that a picking operation scheme meeting the picking operation requirement and solving the congestion problem in the target warehouse is obtained, the smoothness of the picking operation of the warehouse is ensured, and the labor and the picking operation cost are reduced.

In some embodiments, the improved nested genetic algorithm includes a genetic algorithm and a picking path optimization ratlif & Rosenthal algorithm, and the step S201 of performing periodic picking allocation processing on the multi-scale fusion data corresponding to the target picking order by using the pre-planning model, to generate first picking planning information includes the following steps:

step 51, coding and population initializing are performed on the picking object information and the order information in the multi-scale fusion data by using a genetic algorithm, and a plurality of first coding arrays corresponding to the picking batch tasks are generated, wherein the first coding arrays comprise a plurality of first codes, the gene values of the first codes are used for representing a third picking object corresponding to the picking object information, the position numbers of the first codes are used for representing a target picking order corresponding to the order information, and the first coding arrays are also related to path parameters corresponding to the picking batch tasks.

Step 52, based on the path parameters, solving a first optimal picking path corresponding to each first coding array by utilizing a Ratliff & Rosenthal algorithm, carrying out narrow-channel congestion analysis and path adjustment processing on the first optimal picking path to obtain a second optimal picking path and second congestion rate amplification corresponding to the second optimal picking path, and taking the second optimal picking path as the optimal picking path corresponding to the first coding array currently, wherein the second congestion rate amplification is used for representing congestion amplification of a congestion detection point in a target warehouse caused by the second optimal picking path.

In this embodiment, a ratlif & Rosenthal algorithm (also called as RR algorithm) is an accurate algorithm for solving a single-block narrow-channel rectangular warehouse picking path, and is proposed by two scholars of ratlif and Rosentha, because an author does not name the algorithm in a corresponding document, and the initial letters of the two authors are used for calling the algorithm as the RR algorithm, the core idea of the RR algorithm is to apply the idea of building an euler loop, and a walking route of each channel and a connection mode of the paths between the channels are sequentially built from left to right, so that the euler loop is finally formed, thereby obtaining a picking optimal path, fully utilizing the uniqueness of the warehouse layout and the walking route in the warehouse to seek an optimal solution, and reducing the time complexity of the algorithm while solving the optimal solution; in this embodiment, an RR algorithm is applied to solve an optimal picking path loop of a given picking batch, that is, input information is related information of the picking batch, including which orders the picking batch includes, how many storage locations need to be accessed in total, which storage locations need to be accessed in particular, a shelf and channel layout of a warehouse, a picking start point coordinate of the warehouse, and the like, and after calculation by the RR algorithm, an output result is a path loop from the picking start point coordinate through all storage locations needing to be picked and finally returns to the picking start point; it will be appreciated that the operation of solving the first best-order picking path using the RR algorithm is well within the skill of the art and does not constitute an unclear definition of the solution of the first best-order picking path.

Step 53, inspiring the first coding array to perform a preset genetic evolution operation based on the second congestion rate increase to generate a second coding array, wherein the genetic evolution operation includes fitness calculation, tournament selection, genetic crossover and genetic variation.

In this example, the corresponding genetic evolution operation is an existing genetic operation, and in some alternative embodiments, the selection operator employs a tournament selection strategy, the genetic crossover employs a two-point crossover strategy, and the genetic variation employs a reverse mutation strategy.

And step 54, sequentially carrying out the processes of solving the optimal picking path based on Ratliff & Rosenthal algorithm, carrying out narrow channel congestion analysis and path adjustment processing and genetic evolution operation on the solved corresponding optimal picking path, and determining a target coding array from the generated candidate coding arrays, wherein the first picking planning information comprises the target coding array, the gene value of the target coding array represents the corresponding target picking object, the position number of the target coding represents the corresponding target picking order, and the target coding array is related to the path parameters corresponding to the target picking batch task.

Through the steps 51 to 54, the pre-planning based on the multi-scale fusion data is realized, and the first picking planning information is generated by adopting the establishment of the picking batch, the task allocation and the path planning multi-link initial plan under the preset system optimization target guidance and the preset optimization rule constraint, so that the effects of reducing the order delay cost, reducing the congestion occurrence rate and preventing the channel congestion are realized.

It should be noted that, an improved nested genetic algorithm is used for solving a picking planning scheme in pre-planning, the algorithm is integrated with an accurate solving algorithm for solving path planning in a warehouse-RR algorithm on the basis of a traditional genetic algorithm calculation frame, narrow-channel congestion analysis is added in fitness calculation, a batch scheme and a path planning scheme are evaluated according to specific indexes, a multi-person path planning scheme is coordinated and adjusted, and finally a picking scheme with minimum cost and minimum narrow-channel congestion rate is obtained; in this embodiment, since the RR algorithm fully utilizes the uniqueness of the warehouse layout and the walking route in the warehouse and applies the idea of graph theory to find the optimal solution, the time complexity of the algorithm is low, the improved nested genetic algorithm obtained by combining the algorithm with the genetic algorithm can also obtain a better scheme on the basis of not greatly increasing the solving time.

In some embodiments, the narrow-channel congestion analysis and path adjustment processing in step 52 are performed on the first optimal picked path to obtain a second optimal picked path and a second congestion rate increase corresponding to the second optimal picked path, including the following steps:

step 521, in the first optimal picking path, detecting a first picking sub-path corresponding to each third picking object, and detecting a second picking sub-path with path conflict from the first picking sub-paths, wherein the path conflict comprises that at least two third picking objects cross and the picking paths cross when one warehouse channel picks orders.

In this embodiment, in a case where there is a time window conflict and a space conflict that are both present, it is determined that there is congestion in the corresponding first optimal picking path.

And 522, carrying out path adjustment on the second picking sub-path, carrying out path conflict recognition on the second picking sub-path subjected to the path adjustment, and determining a third picking sub-path with path conflict and a fourth picking sub-path without conflict in the second picking sub-path, wherein a warehouse channel corresponding to the third picking sub-path is congested.

In this embodiment, after detecting the second picking sub-path having the path conflict, path adjustment is performed on the second picking sub-path, for example: and adjusting the direction of a picking path of a certain picking object, and after path adjustment, removing the congestion of the second picking sub-path, taking the second picking sub-path after the congestion removal, namely the fourth picking sub-path as a target picking path, and taking the third picking sub-path as a picking sub-path corresponding to the current first coding array and a picking sub-path for measuring the analysis congestion condition corresponding to the first coding array, so that planning is realized by taking the corresponding congestion condition as a target in subsequent genetic operation, and finally generating a coding array for solving the congestion problem of warehouse channels in the target warehouse, namely obtaining the target coding array corresponding to the first picking planning information.

Step 523, determining a second congestion rate increase according to the ratio of the third culling sub-path to all the first culling sub-paths, and determining that the second optimal culling path includes the first culling sub-path, the third culling sub-path, and the fourth culling sub-path of all the first culling sub-paths except the second culling sub-path.

In some embodiments, the target coding array is determined from the generated candidate coding arrays by: and calculating the adaptability of the candidate coding arrays by using a genetic algorithm, and selecting the candidate coding array with the highest adaptability from the candidate coding arrays according to the adaptability to obtain a target coding array.

In some of these embodiments, prior to generating the first pick plan information, the following steps are also implemented:

step 61, obtaining the user requirement of the target user from the preset platform, wherein the user requirement comprises the order information purchased by the target user.

Step 62, based on the order information, generating corresponding picking requirement information in the target warehouse, and fusing the picking requirement information with full-flow real-time data corresponding to the target warehouse to generate multi-scale fused data, wherein the full-flow real-time data at least comprises one of the following: the method comprises the steps of picking task execution progress information, planned picking path information and warehouse channel congestion state information.

Fig. 3 is a schematic diagram of a smart warehouse picking planning digital twin system architecture according to an embodiment of the present application, fig. 4 is a flow chart of a picking order allocation planning method based on narrow channel avoidance according to a preferred embodiment of the present application, fig. 5 is a frame diagram of an improved nested genetic algorithm according to a preferred embodiment of the present application, fig. 6 is a schematic diagram of a selection operator operation process according to a preferred embodiment of the present application, fig. 7 is a schematic diagram of a cross operator operation process according to a preferred embodiment of the present application, fig. 8 is a schematic diagram of a cross operator operation process according to a preferred embodiment of the present application, fig. 9 is a schematic diagram of a cross operator operation process according to a preferred embodiment of the present application, fig. three, fig. 10 is a flow chart of a solution of an optimal picking path according to a preferred embodiment of the present application, fig. 11 is a schematic diagram of a warehouse channel layout and key feature diagram according to a preferred embodiment of the present application, fig. 12 is a schematic diagram of an optimal path in a preferred embodiment of the present application, fig. 13 is a schematic diagram of a path within a warehouse channel in a set area of the present application, fig. 14 is a schematic diagram of a narrow channel analysis and path adjustment according to a preferred embodiment of the present application, fig. 15 is a schematic diagram of a sub-path collision space, fig. 15 is a schematic diagram of the following a schematic diagram of a preferred channel picking order allocation planning method according to a preferred embodiment of the present application, and the following further description of the preferred embodiment of the present application is based on a system of the following a system for picking order allocation planning method according to a preferred embodiment of the following a planned order:

The present embodiments construct an intelligent warehouse digital twin system suitable for warehouse picking dynamic planning, with reference to fig. 3, which is divided into two layers, a physical layer and a virtual layer, wherein,

the physical layer is provided with corresponding sensing equipment for all physical objects involved in the picking operation system so as to acquire real-time information of the picking operation process, including but not limited to identity information, position information, action signals and the like of objects such as staff, forklifts and the like; the real-time operation data is transmitted and interacted with the virtual layer in real time through 5G, WIFI and is stored in a database of the virtual layer as real-time data, so that basic data support is provided for accurate control and linkage decision of each link in a dynamic operation environment.

The virtual layer can establish a mirror image model of the physical object according to basic data and real-time data of the physical object monitored in the database, the model can be continuously developed and evolved according to the real-time data, virtual-real symbiosis is realized, and historical data in the evolution process are further stored in the database for subsequent data analysis, so that a decision basis is provided for a linkage control center; the mirror image model and the data warehouse in the virtual layer support an upper layer application service module together and comprise a static resource management service, a dynamic operation information service, a real-time operation guidance service and a decision planning auxiliary service; the static resource management service comprises management and maintenance of various resources such as warehouse shelves, channel layout, goods storage, sensing equipment and the like; the dynamic operation information service is used for extracting key information from real-time operation data to construct and present a real-time operation model, and mainly comprises a warehouse real-time picking operation state, a narrow channel congestion state, a picking staff working state, a picking equipment using state, a goods position adjusting state and the like, and provides data support for real-time operation guidance and linkage decision planning; the real-time operation guiding service can guide staff in the warehouse to execute operation in real time in a voice, text and other communication mode, so that a decision planning scheme made by a linkage control center or a warehouse manager can be conveniently and well implemented in a floor mode, meanwhile, the memory burden of the staff is reduced, the error rate is reduced, and the service module mainly comprises a picking indoor navigation service, a staff/equipment positioning service, an operation action track tracking service and the like. The decision-making planning auxiliary service is a core sub-module, which can assist a warehouse manager to generate the current optimal picking planning based on static resource management and dynamic operation information, and realize the functions of real-time operation state visualization, narrow-channel congestion simulation and prediction, on-line simulation of picking planning, dynamic batch and path optimization, multi-person picking path coordination, historical planning data analysis and the like; the real-time linkage control center in the virtual layer is the most core part in the intelligent warehouse digital twin system and provides real-time decision and control for the intelligent warehouse digital twin system.

In the intelligent warehouse digital twin system, a physical layer uploads real-time information of warehouse picking operation to a virtual layer, and the virtual layer carries out instantiation simulation and linkage decision making on the picking system through the real-time operation data uploaded by the physical layer and downloads an optimization result to the physical layer to guide the operation of workers.

Referring to fig. 4, the present embodiment provides a dynamic real-time planning and narrow-channel blocking avoidance linkage mechanism for a picking order, which specifically includes the following steps:

step 1, receiving demand information of a customer placed on an electronic commerce platform, wherein the demand information comprises information such as product names, quantity, order delivery date, order delivery address, order delivery mode and the like.

Step 2, converting the customer demand information into warehouse picking demand information (also called picking orders), including the unique identification or number of the goods to be picked, the number required respectively, the storage location in the warehouse, the order picking deadline, etc. Pick orders are trapped in the warehouse pick order pool when pick batches are not planned.

In this embodiment, because of the uncertainty in customer order placement time, pick order arrival time is dynamic, i.e., orders arrive continuously and randomly within 24 hours, orders in the pool of orders accumulate continuously, and some of them are also continuously planned into new pick batches. The order has accumulated a certain amount before the start of the daily on-duty time.

And 3, periodically performing picking pre-planning on the orders in the order pool.

In this embodiment, "periodic" refers to a period of time, such as a picking preplanning every 15 minutes. At this stage, a warehouse picking global multi-link pre-planning model is constructed based on full-flow real-time multi-scale fusion data, picking batch, task allocation and path planning multi-link initial plan making are carried out through an improved nested genetic algorithm under the constraint of system optimization target guidance and optimization rules based on the pre-planning model, in the embodiment, the full-flow real-time multi-scale fusion data comprises all order information in an order pool, picking worker task execution progress, future walking tracks, warehouse congestion conditions and the like, and sensing equipment of an information physical layer is obtained; pre-planning aims at reducing order delay costs and reducing congestion occurrence.

And 4, executing the picking scheme obtained in the pre-planning, and arranging relevant pickers (picking objects) to start goods picking in the target warehouse according to the planning result.

And 5, monitoring the picking operation state in the target warehouse in real time, comparing the actual picking operation state with the pre-planned operation state, and judging that the system disturbance occurs if the deviation between the actual operation state and the pre-planned operation state occurs, the channel congestion outside the plan occurs, and the congestion index value is larger than the appointed parameter value.

And step 6, if the system disturbance is generated, entering a correction planning stage.

In this embodiment, the correction planning phase includes: acquiring dynamic data such as real-time information of warehouse picking operation, warehouse channel congestion state information and the like through a digital twin information framework; based on the collected real-time operation data, estimating congestion index values of the narrow-channel congestion points, and adopting different correction planning strategies according to different index values; a correction planning model capable of timely solving the channel congestion problem is constructed by utilizing a global planning idea and an interference management idea; based on the pre-planning result, a correction plan with the minimum disturbance of the system is formulated according to a correction strategy; the method comprises the steps of establishing a corresponding interference management model by taking the minimum deviation degree as a target and giving an adjustment plan for processing interference events in time, wherein the interference management idea is to emphasize the real-time adjustment of an original plan; in this embodiment, the system disturbance mainly appears on the deviation of the congestion degree of the narrow channel and the picking cost, and the correction strategy can be divided into four types according to different congestion degrees of the narrow channel, which are respectively internal adjustment, multiparty linkage, resource linkage and demand linkage, wherein the internal adjustment refers to the picking planning of only one picking employee, and the internal adjustment includes the picking route or the task execution sequence or the picking waiting time so as to relieve the congestion condition of the narrow channel; multiparty linkage refers to the fact that the alleviation of the narrow-channel congestion condition requires simultaneous coordination of the picking paths of multiple pickers, and the correction strategy aims at minimizing the number of involved employees, i.e. adjusting the picking plans of fewer employees as much as possible and realizing the solution of the congestion problem; resource adjustment refers to the fact that narrow-channel congestion problems in a system cannot be effectively relieved by simply adjusting a picking task plan being executed, only external resources can be called, or the number of workers and picking equipment walking in a warehouse is reduced; demand linkage refers to modifying the order demand of a task being performed, and mainly refers to order deadlines, by which the pressure of warehouse picking operations is reduced by relaxation of demand.

And 7, executing correction planning, and instructing staff to perform picking operation according to the correction planning result.

And 8, judging whether the congestion condition is relieved after the correction planning is performed in real time, if not, repeating the steps 6 to 7, and if so, continuously monitoring the real-time operation state of the picking system, and repeating the steps 5 to 7.

The following describes a dynamic order batch and path planning optimization algorithm considering narrow-channel blocking avoidance according to the present application, as follows:

the present embodiment provides a mathematical model of the pick pre-planning problem, the relevant parameters of which are defined as follows:

selecting an unallocated order set in the order pool for the time t;representing a pick order and,；representing the number of items that order i needs to pick;a position number representing the q-th item that order i needs to pick;representing ordersA set of location numbers for items to be picked,；representing ordersLoading deadlines (latest picking completion time);representing ordersThe total weight of the contained goods;representing a collection of all pickers within the warehouse;indicating that the person is to be picked up,；representing a pickerWhether it is in an idle state at time t,；representing a picker A point in time at which the current task is expected to be completed;representing the path sequence number, i.e. the kth path track;representing a pickerThe first of the remaining non-walked path trajectories at time tThe number of road segments is one,；representing a pickerThe first of the remaining non-walked path trajectories at time tThe channel of each road section;representing a pickerThe first of the remaining non-walked path trajectories at time tThe traveling direction of the individual road segments,；representing a pickerThe first of the remaining non-walked path trajectories at time tStarting point position numbers of the individual road segments;representing a pickerThe first of the remaining non-walked path trajectories at time tStart time of each road segment;representing a pickerThe first of the remaining non-walked path trajectories at time tEnd position numbers of the individual road sections;representing a pickerThe first of the remaining non-walked path trajectories at time tThe end time of each road segment;representing a pickerA set of remaining non-walked path trajectories at time t,；representing a set of congestion monitoring points;indicating the point of congestion monitoring,；indicating congestion monitoring pointsAt the position ofThe congestion status of the moment of time,；representing monitoring pointsAt the position ofThe predicted future period of congestion at the moment in time,；indicating that at time t, it is idleIs a collection of pickers of (a),；representing a collection of picking orders assigned by picker j at time t; Representing a pick order allocation and path planning decision scheme for picker j at time t,；representing a decision scheme at time t；Representing the maximum capacity of the picking container;indicating the average walking rate of the pickers.

The goal of picking pre-planning is to minimize the total cost of the planned picking cycle, including labor costs, aisle congestion management costs, order delay costs.

The object is:wherein, the method comprises the steps of, wherein,

。

constraint:

。

the improved nested genetic algorithm is a heuristic algorithm, integrates an RR algorithm for solving the accurate solution of path planning in a warehouse on the basis of a traditional genetic algorithm calculation frame, adds narrow-channel congestion analysis in fitness calculation, evaluates a batch scheme and a path planning scheme according to specific indexes, coordinates and adjusts a multi-person path planning scheme, and finally obtains a picking scheme with minimum cost and minimum narrow-channel congestion rate; in this embodiment, the genetic algorithm mainly solves the problem of batch of orders, and the RR algorithm further solves an optimal path loop of the batch on the batch result of the genetic algorithm, where the length of the optimal path loop is an important index for evaluating the batch result; meanwhile, as an optimal path loop can be converted into a plurality of paths with directions, the paths of a plurality of persons are coordinated through narrow-channel congestion analysis, the optimal path is obtained, the fitness of chromosomes representing the batch scheme is calculated by using the optimal path, so that the chromosomes meeting the requirements are further screened, and referring to fig. 5, the improved nested genetic algorithm specifically performs the following steps:

1. Encoding

The genetic algorithm is mainly used for obtaining a better order batch scheme, and determining the number of order batches according to the number of current idle staff, namely, the corresponding problem is defined as follows: based on N orders in the current order pool and M unoccupied pickers, respectively distributing the orders to each staff as picking tasks of the next picking period; because the number of orders and the number of staff are integers and have continuity, an integer coding mode is adopted, the chromosome length is N, the chromosome length is consistent with the number of orders in an order pool, the position numbers of genes respectively correspond to the order numbers in the order pool, for example, the first gene represents order information with the number of 1, and the nth gene represents order information with the number of N; the value range of each gene in the chromosome is 0~M, which indicates which employee the order should be allocated to, and the value of the gene corresponds to the employee number, for example, a value of 1 indicates an employee with a number of 1, and a value of M indicates an employee with a number of M; the first gene code is 3 in the following description, which means that orders numbered 1 in an order pool are distributed to staff numbered 3, and a set formed by orders corresponding to all the gene codes of 3 is an order batch required to be selected by staff numbered 3; in particular, a gene encoded as 0 indicates not to participate in this round of picking allocation;

。

2. Population initialization

The initial population is a starting point of the genetic algorithm for loop iteration optimization, and a good initial population can reduce the iteration times of the genetic algorithm and improve the efficiency and quality of solving the optimization problem; generally, the initial population can be obtained through random generation, the initial population quantity is generally in the range of 50-200, and too small or too large initial population quantity can increase the iteration times of the genetic algorithm or excessively long solving time, so that the operation efficiency is low. In the algorithm, the population number is set to be 50, the maximum iteration number is 100, and the population is generated in a random initialization mode.

3. Fitness function calculation

In the genetic algorithm, the fitness function is used for measuring the environmental fitness of individuals in the population, the individuals are subjected to superior and inferior elimination according to the fitness value, the higher the fitness is, the higher the probability of inheriting offspring is, the lower the fitness is, and the probability of being eliminated is high. Because the genetic algorithm is only based on the fitness function value when being selected, the quality of the fitness function selection can directly influence the speed of solving the genetic algorithm and the searching of the optimal solution, and when solving the actual problem, a mapping relation between an optimization target and the individual fitness is required to be established; since the goal of picking pre-planning is to minimize the total cost of the planned picking cycle, including labor costs, aisle congestion management costs, order delay costs; therefore, the fitness value f of the individual is also measured from the three aspects, and the following formula is specific:

；

The following explains in detail three parts of the fitness value calculation formula, wherein,

representing the cost of labor time, i.e. at time t, the system state isAt this time, take a picking planThe generated picking labor cost is，The specific calculation formula of (2) is as follows:

is a labor cost coefficient, can be set according to the actual situation of enterprises, and has a formulaAnd the time cost is converted into manual cost by calculating the time generated by each employee in the picking plan, when the employee is not assigned with a task at the time t, the generated cost is calculated to be 0, otherwise, the sum of the walking time, the picking operation time and the congestion waiting time is calculated. Wherein the travel time is a ratio of an optimal picking path length of the picking batch to an average worker travel rate;

representing the management cost or loss cost due to congestion, i.e. at time t, the system state isAt this time, take a picking planThe congestion management cost for the system is thatThe specific calculation formula is as follows:

wherein,for the congestion management cost coefficient, the value is taken according to the actual condition of an enterprise, and the formula is as followsFor calculating a congestion rate, whereinAn increased value representing the length of time that the batch scheme caused congestion at congestion monitoring point q, Indicating the magnitude of increase in the number or number of congestion caused by the scheme, coefficientAndthe method comprises the steps that deviation of an enterprise on congestion time length and the number of congestion points is measured, and relevant data of congestion state prediction are obtained through a narrow-channel congestion analysis operator;

representing the order delay cost, i.e. at time t, the system state isAt this time, take a picking planThe congestion management cost for the system is thatThe specific calculation formula is as follows:

wherein,the cost coefficient of order delay can be valued according to the actual situation of enterprisesTo estimate the final delay cost for order i under the picking strategy; since the order delay can only be judged after the order is completed, the order delay cost is the sameHysteresis in the problem; for the treatment of hysteresis problems, the solution adopted here is to employ a short-looking cost approximation strategy; in pre-picking planning, i.e. estimating the cost of the order delay it eventually causes according to a pre-picking planning scheme, and by introducing parametersThe estimation method is accurate, and parameters are set through a steep drop algorithmIs set to the optimum value of (2).

4. Selection operator

Referring to fig. 6, a tournament selection strategy is employed to take a number of individuals from the population at a time (sample-back) and then select the best one of them to enter the offspring population. Repeating the operation until the new population size reaches the original population size; because the tournament selection strategy has the characteristics of smaller complexity, easy parallelization, difficult sinking into local optimal points and no need of sorting all fitness values, in the embodiment, a ternary tournament strategy is adopted, namely 3 individuals are taken out of the overall samples at one time, then the optimal individuals are taken out of the individuals and put into a set reserved to the next generation population, and the specific operation flow is as follows: 1. firstly, randomly selecting three individuals from a population; 2. comparing the fitness values of the three individuals, and selecting the individual with the largest fitness value as a winner to be placed into a offspring population set; 3. and (3) repeating the steps 1 to 2 until the number of the offspring population sets reaches the set population size.

5. Crossover operator

Referring to fig. 7 to fig. 9, the crossover operator in this embodiment adopts a two-point crossover strategy, where two-point crossover refers to a specific operation procedure of randomly setting two crossover points in an individual chromosome and then performing partial gene exchange, where the two-point crossover is: two individuals paired with each other were randomly selected as shown in fig. 7; two cross points are randomly arranged in two individual code strings paired with each other as shown in fig. 8; the partial chromosomes of the two individuals between the two set crossover points are swapped, as shown in fig. 9.

6. Mutation operator

In the embodiment, the mutation probability is set to be 0.005, a strategy of turning mutation is adopted, one gene is randomly selected from chromosomes, and then the gene is randomly turned to other selectable gene values; the specific mutation logic is as follows: a random number is first generated for each individual selected. If the random number is not greater than the preset mutation rate, the offspring individual needs mutation operation; otherwise, the offspring individuals remain unchanged; if the conditions for performing the mutation operation are satisfied, a mutation point is randomly selected in the range of [0, N ], and this mutation point determines which gene will be mutated. Subsequently, randomly mutating the gene of the variation point, wherein the mutation range is [0, M ]; finally, the new progeny after mutation is preserved.

7. Termination of

Repeating the steps 3 to 6 until the preset iteration times or the convergence of the result are reached; in this embodiment, the iteration number is set to 100, and after the iteration process is finished, an individual with the highest fitness value is found from the population, and the individual is the optimal solution.

The following description of the RR algorithm of the embodiment of the present application solves for the optimal picking path is as follows:

referring to fig. 10, the method for solving the optimal picking path by using the RR algorithm according to the embodiment of the present application includes the following steps:

step 1, inputting relevant information of a picking batch, wherein the relevant information mainly comprises which orders the picking batch comprises, which storage positions are specifically required to be accessed, the goods shelf and channel layout of a warehouse, the picking departure point coordinates of the warehouse and the like; the applicable scene of the embodiment of the application is a parallel channel double-block plane warehouse, the warehouse shelf, namely the channel layout is shown in the left diagram of fig. 11, fig. 11 shows that a plane warehouse is divided into two blocks by a middle transverse channel region, the upper part is an X region, the lower part is a Y region, each square in one block represents a storage position, a plurality of square forms together form a shelf region, and a plurality of longitudinal channels which are parallel to each other are arranged; the darkened squares represent the pick points involved in this pick batch, with the drop being the pick departure point. The path required to be output by the algorithm passes through all black pick points and departure points, and the key characteristic information of the left graph in fig. 11 is extracted to obtain a right graph, wherein the transverse channels are respectively indicated by letters a, b and c.

Step 2, determining a part of path equivalent class storage list according to warehouse layout informationTwo partial loop child graphs that will satisfy the following two conditions belong to the same equivalence class: (1)Parity of degree (odd U, even E, 0) is the same in both, andthe parity of the degree of (2) is the same in both, andthe parity of the degree of (2) is the same in both; (2) With the exception of the node of degree 0, both partial ring-stream subgraphs have no connected component, or only one contains at leastDots orThe connected components of the points, or bothDots andthe connected components of the points, as in the parallel channel dual block plane warehouse described in step 1, are 25 in total, each equivalent class being represented by a tuple containing 4 or 5 elements, as follows:

wherein the first element of the tuple representsParity of dot degree, 0 representing degree of 0, u representing degree of even, e representing degree of odd, and the second element of tuple representsThe third element of the tuple represents the parity of the degree of (2)Parity of the degree of (2); the fourth element represents how many connected variables are contained in the partial loop-stream subgraph; the fifth element is only needed when the values of the current four elements are (e, e, e, 2), and is used for indicating which two points of three points a, b and c belong to the same connected component; if "a-bc" means that the two points b and c belong to the same connected component, and the point a belongs to the other connected component; after determining all equivalence types, initializing a storage list For storing all possible optimal path subgraphs.

Step 3, calculating the path scheme among the channels, and marking the path scheme asAnd update the list. The path scheme between channels refers to the path scheme of the transverse channel between the channel j-1 and the channel j, and for the parallel channel double-block plane warehouse, the path scheme between the channels comprises the following 14 (shown by referring to fig. 12), to be listedAll equivalent paths in the table are respectively connected with the following path schemes to obtain new sub-paths, the equivalent category of each sub-path is judged, and the optimal path in each equivalent category is taken and stored in the listFor the next calculation. In particular, when j=1,=None。

step 4, calculating the path scheme of the Y area in the channel j, which is recorded asAnd update the listThe path scheme in the channel is composed of 6 possibilities, as shown in FIG. 13, where scheme (6) is only likely to be adopted, the current list, when there are no pickpoints in the channel that need to go throughAll equivalent paths in the table are respectively connected with the following path schemes to obtain new sub-paths, and judge whether the equivalent category of each sub-path and the equivalent category are possible to be solved, after the non-feasible solution is thrown out, the optimal path in each equivalent category is taken and stored in the list For the next calculation.

Step 5, calculating the path scheme of the X area in the channel j, which is recorded asAnd update the listThe specific operation is similar to step 4.

And 6, adding 1 to the channel number j, wherein the channel number j represents the channel currently being calculated, and the channel numbers are respectively numbered 1 to N from left to right, and N is the number of the channels.

Step 7, judging whether the channel number j currently being calculated is in a calculation range, namely smaller than or equal to N, if yes, repeating the steps 3 to 6; otherwise, step 8 is entered.

Step 8, when traversing all channels from left to right, storing the optimal path in the listIn the method, a path record with the shortest recorded path length in the following 8 equivalence classes is taken, and the path record is an optimal solution of the path planning problem, which is specifically as follows:

。

step 9, according to the listAnd backtracking the complete information of the optimal path by the change history, and finally obtaining the Euler loop of the optimal path.

The following describes a narrow-channel congestion analysis operator according to an embodiment of the present application as follows:

the narrow-channel congestion analysis operator is mainly used for identifying the congestion waiting time and the increasing amplitude of the congestion rate caused by a given optimal path loop scheme and the current narrow-channel congestion state of a warehouse, so that an optimal path combination scheme is obtained, wherein the optimal picking loop scheme is obtained by the RR algorithm, only an optimal path loop is specified, no path direction is specified, so that if one optimal loop starts to walk from different directions, different path schemes can be generated, and different paths of a plurality of pickers can finally form different optimal path combination schemes. The congestion waiting time and the increasing amplitude of the congestion rate are one important index for evaluating the path planning scheme and the batch scheme, the input information of the narrow-channel congestion analysis operator comprises the congestion rate of the current congestion monitoring point and the optimal path loops of a plurality of pickers, the output result comprises the optimal path combination scheme of the plurality of pickers and the increasing amplitude of the congestion rate under the optimal path combination scheme (comprising the increasing amplitude of the congestion rate of each congestion monitoring point and the increasing amplitude of the overall congestion rate of the warehouse), and referring to fig. 14, the method comprises the following steps:

And step 1, inputting the optimal picking loop schemes of all the pickers related to the current batch scheme and the congestion information of the current warehouse.

And 2, acquiring all possible paths which each worker can walk according to the input optimal picking loop scheme of all the pickers. And randomly generates a path combining scheme.

And step 3, identifying whether congestion is generated in the path combination scheme obtained in the step 2. The judgment of the congestion condition is mainly by judging whether the time conflict and the space conflict exist simultaneously. A time conflict refers to that at least 2 pickers in the path combination scheme cross through the time window of the same narrow channel, and a space conflict refers to that the pickers cross the paths of the narrow channel, as shown in fig. 15, including the following three cases: in case (1), two employees or more employees pick goods with each other in the same narrow channel, and need to pass through each other to reach their respective target pick points; case (2), where two or more employees pick in the same narrow aisle, reaching the same target pick point at approximately the same time; in case (3), two or more employees pick in the same direction in the same narrow lane, but one employee stops picking at a pick point, blocking the pick path of the other employee, resulting in waiting.

Step 4, if there are time conflict and space conflict in the path combination scheme at the same time, congestion will occur, judge whether the path has space adjusted, if yes, carry out step 5; if there is no conflict or only one of the conflicts, the congestion forming condition cannot be constituted, or there is no room for adjustment of the path combining scheme although there is a conflict, step 6 is performed.

And 5, adjusting a path combination scheme to solve the conflict.

In this embodiment, a picking path of one employee is fixed first, a picking path direction of another employee is adjusted, if the path or direction of another employee cannot be adjusted, the employee with the fixed route is replaced until a feasible path combination scheme is finally obtained, and then steps 3 to 4 are repeatedly executed.

Step 6, calculating the congestion rate of the optimal path combination scheme according to the analysis result in the step 3, wherein the congestion rate is defined as the ratio of congestion states of all congestion monitoring points in the whole picking period, and the calculation formula of the congestion rate is as follows:

。

and 7, outputting an optimal path combination scheme and a congestion rate.

The embodiment also provides a picking order allocation planning device based on narrow channel blocking avoidance, which is used for realizing the embodiment and the preferred implementation mode, and the description is omitted. As used below, the terms "module," "unit," "sub-unit," and the like may be a combination of software and/or hardware that implements a predetermined function. While the means described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and contemplated.

Fig. 16 is a block diagram of a narrow channel unblocking-based pick order allocation planning device, as shown in fig. 16, including a pre-planning module 161, a detection module 162, a correction module 163, and a processing module 164, wherein,

the pre-planning module 161 is configured to detect real-time running state information corresponding to a target warehouse during order picking based on first picking planning information, and acquire first running state information corresponding to the target warehouse from the first picking planning information, where the first picking planning information is generated by periodically performing picking allocation processing on multi-scale fusion data corresponding to a target picking order by using a pre-planning model, the pre-planning model is constructed based on a preset optimization strategy target and an improved nested genetic algorithm, the picking allocation processing includes picking batch and path planning, and a corresponding warehouse channel is formed in the target warehouse by deploying narrow channel shelves.

The detection module 162 is coupled to the pre-planning module 161, and is configured to determine whether a deviation exists between the real-time operation state information and the first operation state information, and determine a target correction policy according to the initial channel congestion state information corresponding to the warehouse channel when the deviation exists, where the target correction policy includes a target picking correction plan.

The correction module 163 is coupled to the detection module 162, and obtains corrected operation state information from the corrected picking plan information, and determines first warehouse congestion state information corresponding to the corrected operation state information, where the corrected picking plan information is generated by performing target picking correction planning on the real-time operation state information and the initial channel congestion state information.

And the processing module 164 is coupled to the correction module 163, and is configured to perform reprofiling update on the corrected picking plan information according to the first warehouse congestion status information according to the target picking correction plan, and perform allocation planning on the target picking order according to the target picking plan information generated by the reprofiling update, where when order picking is performed based on the target picking plan information, the target warehouse relieves warehouse aisle congestion.

According to the narrow-channel blocking-avoidance-based picking order allocation planning device, real-time running state information corresponding to a target warehouse is detected in the order picking process based on first picking planning information, and first running state information corresponding to the target warehouse is obtained from the first picking planning information, wherein the first picking planning information is generated by periodically picking allocation processing of multi-scale fusion data corresponding to the target picking order by using a pre-planning model, the pre-planning model is constructed based on a preset optimization strategy target and an improved nested genetic algorithm, the picking allocation processing comprises picking batch and path planning, and a corresponding warehouse channel is formed in the target warehouse by deploying narrow-channel shelves; determining whether deviation exists between the real-time running state information and the first running state information, and determining a target correction strategy according to initial channel congestion state information corresponding to a warehouse channel under the condition that the deviation exists, wherein the target correction strategy comprises a target picking correction plan; acquiring correction operation state information from correction picking planning information, and determining first warehouse congestion state information corresponding to the correction operation state information, wherein the correction picking planning information is generated by carrying out target picking correction planning on real-time operation state information and initial channel congestion state information; according to the first warehouse congestion state information, re-planning and updating modified picking planning information according to target picking correction planning, and carrying out distribution planning on target picking orders according to target picking planning information generated by the re-planning and updating, wherein when order picking is carried out based on the target picking planning information, a target warehouse releases warehouse channel congestion, the problem that narrow channel congestion is ignored by an e-commerce warehouse picking planning method in the related art, picking planning disorder, labor cost and order fulfillment cost are easy to cause is solved, and the beneficial effects of improving warehouse picking operation smoothness and reducing labor and picking operation cost are achieved by periodically pre-planning and dynamically planning and correcting the picking orders in real time and executing a narrow channel blocking avoidance mechanism.

In some of these embodiments, the detection module 162 further includes: the first detection unit is used for detecting a first congestion rate corresponding to a congestion detection node positioned on the warehouse channel in the initial warehouse channel congestion state information, wherein the first congestion rate is used for representing the congestion degree of the corresponding warehouse channel; the first computing unit is coupled with the first detecting unit and is used for computing first total congestion rates corresponding to all warehouse channels in the target warehouse based on the first congestion rates, and determining channel congestion levels in the target warehouse according to the first total congestion rates; the first acquisition unit is coupled with the first calculation unit and is used for acquiring a target picking correction plan from a preset picking correction plan based on the channel congestion level, wherein the picking correction plan at least comprises one of the following steps: a single picking object picking plan adjustment, a linkage adjustment of the picking plans of a plurality of picking objects, the addition and deletion of the picking plans, and the correction of the first picking plan information.

In some of these embodiments, the correction module 163 further includes: the second detection unit is used for detecting first picking sub-planning information from the real-time running state information, wherein the first picking sub-planning information comprises a first picking object and a first picking sub-planning; the first determining unit is coupled with the second detecting unit and is used for determining a target picking correction plan according to the congestion state information of the initial warehouse channel; the first correction unit is coupled with the first determination unit and is used for carrying out correction processing on the first picking sub-plan based on the target picking correction plan to generate candidate picking sub-plans, wherein the correction processing comprises adjustment of the first picking sub-plan and/or adjustment of the first picking object, and the candidate picking sub-plan comprises a second picking object after adjustment of the first picking object and a second picking sub-plan after adjustment of the first picking object; the first selecting unit is coupled with the first correcting unit and is used for determining a deviation value of a first picking cost corresponding to each candidate picking sub-plan and a preset cost threshold value, determining a second congestion rate of a warehouse channel corresponding to each candidate picking sub-plan, selecting the candidate picking sub-plan with the smallest deviation value and the lowest second congestion rate, and obtaining corrected picking planning information, wherein the corrected picking planning information comprises the candidate picking sub-plans with the smallest deviation value and the lowest second congestion rate.

In some of these embodiments, the processing module 164 further includes: the third detection unit is used for detecting a third congestion rate corresponding to a congestion detection node positioned on the warehouse channel in the first warehouse congestion state information, wherein the third congestion rate is used for representing the congestion degree of the warehouse channel when the picking operation is performed according to the corrected picking planning information; the second calculation unit is coupled with the third detection unit and is used for calculating third total congestion rates corresponding to all warehouse channels in the target warehouse based on the third congestion rates and judging whether the target warehouse releases the warehouse channel congestion according to the third total congestion rates; the first judging unit is coupled with the second calculating unit and is used for determining that the target picking planning information comprises modified picking planning information under the condition that the warehouse channel congestion is judged to be relieved by the target warehouse; and the second judging unit is coupled with the second calculating unit and is used for carrying out correction planning on the correction picking planning information according to the target picking correction planning until the target picking planning information is generated under the condition that the warehouse channel congestion is not relieved by the target warehouse.

In some of these embodiments, the improved nested genetic algorithm includes a genetic algorithm and a culling path optimization ratlif & Rosenthal algorithm, and the pre-planning module 161 further includes: the first distribution unit is used for carrying out coding and population initialization processing on the picking object information and the order information in the multi-scale fusion data by utilizing a genetic algorithm to generate a plurality of first coding arrays corresponding to the picking batch tasks, wherein the first coding arrays comprise a plurality of first codes, the gene values of the first codes are used for representing a third picking object corresponding to the picking object information, the position numbers of the first codes are used for representing a target picking order corresponding to the order information, and the first coding arrays are also related to path parameters corresponding to the picking batch tasks; the first planning unit is coupled with the first distribution unit and is used for solving a first optimal picking path corresponding to each first coding array by utilizing Ratliff & Rosenthal algorithm based on path parameters, carrying out narrow-channel congestion analysis and path adjustment processing on the first optimal picking path to obtain a second optimal picking path and second congestion rate amplification corresponding to the second optimal picking path, and taking the second optimal picking path as the optimal picking path corresponding to the first coding array currently, wherein the second congestion rate amplification is used for representing congestion amplification of congestion detection points in a target warehouse caused by the second optimal picking path; the first operation unit is coupled with the first planning unit and used for inspiring the first coding array to perform preset genetic evolution operation based on the second congestion rate increase so as to generate a second coding array, wherein the genetic evolution operation comprises fitness calculation, tournament selection, genetic crossover and genetic variation; the first updating unit is coupled with the first operation unit and is used for sequentially carrying out Ratliff & Rosenthal algorithm-based optimal picking path solving, narrow-channel congestion analysis and path adjustment processing and genetic evolution operation processing on the solved corresponding optimal picking path, and determining a target coding array from the generated candidate coding arrays, wherein the first picking planning information comprises the target coding array, the gene value of target coding of the target coding array represents a corresponding target picking object, the position number of the target coding represents a corresponding target picking order, and the target coding array is related to the path parameters corresponding to target picking batch tasks.

In some embodiments, the first planning unit is further configured to detect, in a first optimal picking path, a first picking sub-path corresponding to each third picking object, and detect, from the first picking sub-paths, a second picking sub-path in which a path conflict exists, where the path conflict includes that when at least two third picking objects pick orders in one warehouse aisle, a time window has a cross and the picking paths have a cross; carrying out path adjustment on the second picking sub-path, carrying out path conflict recognition on the second picking sub-path subjected to the path adjustment, and determining a third picking sub-path with path conflict and a fourth picking sub-path without conflict in the second picking sub-path, wherein a warehouse channel corresponding to the third picking sub-path is congested; and determining a second congestion rate increase according to the ratio of the third picking sub-path to all the first picking sub-paths, and determining that the second optimal picking path comprises the first picking sub-paths, the third picking sub-paths and the fourth picking sub-paths except the second picking sub-path in all the first picking sub-paths.

In some embodiments, the first updating unit is further configured to calculate, using a genetic algorithm, a fitness corresponding to the candidate coding array, and select, according to the fitness, a candidate coding array with a highest fitness from the candidate coding arrays, so as to obtain the target coding array.

In some of these embodiments, the narrow channel blocking-based pick order allocation plan device is further configured to obtain user requirements of the target user from a preset platform prior to generating the first pick plan information, wherein the user requirements include order information purchased by the target user; based on order information, generating corresponding picking demand information in a target warehouse, and fusing the picking demand information with full-flow real-time data corresponding to the target warehouse to generate multi-scale fused data, wherein the full-flow real-time data at least comprises one of the following: the method comprises the steps of picking task execution progress information, planned picking path information and warehouse channel congestion state information.

The present embodiment also provides an electronic device comprising a memory having stored therein a computer program and a processor arranged to run the computer program to perform the steps of any of the method embodiments described above.

Optionally, the electronic apparatus may further include a transmission device and an input/output device, where the transmission device is connected to the processor, and the input/output device is connected to the processor.

Alternatively, in this embodiment, the processor may be configured to be executed by a computer program to perform the steps of any of the method embodiments described above.

It should be noted that, specific examples in this embodiment may refer to examples described in the foregoing embodiments and alternative implementations, and this embodiment is not repeated herein.

In addition, in combination with the picking order allocation planning method based on narrow channel blocking avoidance in the above embodiment, the embodiment of the application can be realized by providing a storage medium. The storage medium has a computer program stored thereon; the computer program, when executed by a processor, implements any of the narrow channel blocking-avoidance based pick order allocation planning methods of the above embodiments.

The foregoing examples represent only a few embodiments of the present application, which are described in more detail and are not to be construed as limiting the scope of the invention. It should be noted that it would be apparent to those skilled in the art that various modifications and improvements could be made without departing from the spirit of the present application, which would be within the scope of the present application. Accordingly, the scope of protection of the present application is to be determined by the claims appended hereto.

Claims

1. A picking order allocation planning method based on narrow channel blocking avoidance is characterized by comprising the following steps:

in order picking process based on first picking planning information, detecting corresponding real-time running state information in a target warehouse, and acquiring first running state information corresponding to the target warehouse from the first picking planning information, wherein the first picking planning information is generated by periodically picking allocation processing of multi-scale fusion data corresponding to a target picking order by using a pre-planning model, the pre-planning model is constructed based on a preset optimizing strategy target and an improved nested genetic algorithm, the picking allocation processing comprises picking batch and path planning, and a corresponding warehouse channel is formed in the target warehouse by deploying narrow channel shelves;

Determining whether the real-time running state information and the first running state information have deviation, and determining a target correction strategy according to initial channel congestion state information corresponding to the warehouse channel under the condition that the deviation exists, wherein the target correction strategy comprises a target picking correction plan;

acquiring correction operation state information from correction picking planning information, and determining first warehouse congestion state information corresponding to the correction operation state information, wherein the correction picking planning information is generated by carrying out target picking correction planning on the real-time operation state information and the initial channel congestion state information;

and carrying out rescheduling updating on the corrected picking planning information according to the first warehouse congestion state information and carrying out distribution planning on target picking orders according to target picking planning information generated by rescheduling updating, wherein when order picking is carried out based on the target picking planning information, the target warehouse releases warehouse channel congestion.

2. The method of claim 1, wherein determining a target correction policy based on initial warehouse aisle congestion status information corresponding to the warehouse aisle, comprises:

Detecting a first congestion rate corresponding to a congestion detection node positioned on the warehouse channel in the initial warehouse channel congestion state information, wherein the first congestion rate is used for representing the congestion degree of the corresponding warehouse channel;

calculating first total congestion rates corresponding to all warehouse channels in the target warehouse based on the first congestion rates, and determining channel congestion levels in the target warehouse according to the first total congestion rates;

based on the aisle congestion level, acquiring the target picking correction plan from a preset picking correction plan, wherein the picking correction plan at least comprises one of the following steps: a single picking object picking plan adjustment, a linkage adjustment of the picking plans of a plurality of picking objects, the addition and deletion of the picking plans, and the correction of the first picking plan information.

3. The method of claim 2, wherein the objective picking correction planning of the real-time operational status information and the initial warehouse aisle congestion status information comprises:

detecting first picking sub-plan information from the real-time running state information, wherein the first picking sub-plan information comprises a first picking object and a first picking sub-plan;

Determining the target picking correction plan according to the initial warehouse channel congestion state information;

correcting the first picking sub-plan based on the target picking correction plan to generate a candidate picking sub-plan, wherein the correction process comprises adjustment of the first picking sub-plan and/or adjustment of a first picking object, and the candidate picking sub-plan comprises a second picking object after adjustment of the first picking object and a second picking sub-plan after adjustment of the first picking object;

determining a deviation value of a first cable casting cost corresponding to each candidate picking sub-plan and a preset cost threshold value, determining a second congestion rate of the warehouse channel corresponding to each candidate picking sub-plan, and selecting the candidate picking sub-plan with the minimum deviation value and the lowest second congestion rate to obtain modified picking planning information, wherein the modified picking planning information comprises the candidate picking sub-plans with the minimum deviation value and the lowest second congestion rate.

4. The method of claim 1, wherein rescheduling updates the modified pick plan information in accordance with the target pick correction plan based on the first warehouse congestion status information, comprising:

Detecting a third congestion rate corresponding to a congestion detection node positioned on the warehouse channel in the first warehouse congestion state information, wherein the third congestion rate is used for representing the congestion degree of the corresponding warehouse channel when the picking operation is performed according to the corrected picking planning information;

calculating third total congestion rates corresponding to all warehouse channels in the target warehouse based on the third congestion rates, and judging whether the target warehouse relieves warehouse channel congestion according to the third total congestion rates;

determining that the target picking plan information includes the modified picking plan information if it is determined that the target warehouse relieves warehouse aisle congestion;

and under the condition that the warehouse channel congestion is not relieved by the target warehouse, carrying out correction planning on the correction picking planning information according to the target picking correction planning until the target picking planning information is generated.

5. The method of claim 1, wherein the modified nested genetic algorithm comprises a genetic algorithm and a pick path optimization ratlif & Rosenthal algorithm, wherein periodically picking allocation processing is performed on the multiscale fused data corresponding to the target pick order using a pre-planning model, and generating the first pick planning information comprises:

Coding and population initializing the picking object information and the order information in the multi-scale fusion data by utilizing the genetic algorithm to generate a plurality of first coding arrays corresponding to picking batch tasks, wherein the first coding arrays comprise a plurality of first codes, the gene values of the first codes are used for representing a third picking object corresponding to the picking object information, the position numbers of the first codes are used for representing the target picking order corresponding to the order information, and the first coding arrays are also associated with path parameters corresponding to the picking batch tasks;

based on the path parameters, solving a first optimal picking path corresponding to each first coding array by utilizing the Ratliff & Rosenthal algorithm, carrying out narrow-channel congestion analysis and path adjustment processing on the first optimal picking path to obtain a second optimal picking path and second congestion rate amplification corresponding to the second optimal picking path, and taking the second optimal picking path as the optimal picking path currently corresponding to the first coding array, wherein the second congestion rate amplification is used for representing congestion amplification of a congestion detection point in the target warehouse caused by the second optimal picking path;

Inspiring the first coding array to perform a preset genetic evolution operation based on the second congestion rate increase to generate a second coding array, wherein the genetic evolution operation comprises fitness calculation, tournament selection, genetic crossover and genetic variation;

and sequentially carrying out the processes of solving an optimal picking path based on the Ratliff & Rosenthal algorithm, carrying out the narrow channel congestion analysis and path adjustment processing and the genetic evolution operation on the solved corresponding optimal picking path, and determining a target coding array from the generated candidate coding arrays, wherein the first picking planning information comprises the target coding array, the gene value of the target coding array represents a corresponding target picking object, the position number of the target coding represents a corresponding target picking order, and the target coding array is related to the path parameter corresponding to the target picking batch task.

6. The method of claim 5, wherein performing narrow-path congestion analysis and path adjustment processing on the first optimal picked path to obtain a second optimal picked path and a second congestion rate increase corresponding to the second optimal picked path, comprises:

Detecting a first picking sub-path corresponding to each third picking object in the first optimal picking paths, and detecting a second picking sub-path with path conflict from the first picking sub-paths, wherein the path conflict comprises that at least two third picking objects cross and the picking paths cross when picking orders in one warehouse channel;

carrying out path adjustment on the second picking sub-path, carrying out path conflict recognition on the second picking sub-path subjected to the path adjustment, and determining a third picking sub-path with path conflict and a fourth picking sub-path without conflict in the second picking sub-path, wherein the warehouse channel corresponding to the third picking sub-path is congested;

determining the second congestion rate increase according to the ratio of the third culling sub-path to all the first culling sub-paths, and determining that the second optimal culling path includes the first culling sub-path, the third culling sub-path, and the fourth culling sub-path of all the first culling sub-paths except the second culling sub-path.

7. The method of claim 5, wherein determining the target coding array from the generated candidate coding arrays comprises: and calculating the adaptability of the candidate coding array by using the genetic algorithm, and selecting the candidate coding array with the highest adaptability from the candidate coding arrays according to the adaptability to obtain the target coding array.

8. The method of claim 1, wherein prior to generating the first pick plan information, the method further comprises:

acquiring user requirements of a target user from a preset platform, wherein the user requirements comprise order information purchased by the target user;

based on the order information, generating corresponding picking demand information in the target warehouse, and fusing the picking demand information with full-flow real-time data corresponding to the target warehouse to generate the multi-scale fused data, wherein the full-flow real-time data at least comprises one of the following: the method comprises the steps of picking task execution progress information, planned picking path information and warehouse channel congestion state information.

9. An electronic device comprising a memory and a processor, characterized in that the memory has stored therein a computer program, the processor being arranged to run the computer program to perform the steps of the narrow channel unblocking-based pick order allocation planning method of any one of claims 1 to 8.

10. A storage medium having stored thereon a computer program, which when executed by a processor, implements the narrow channel unblocking-based pick order allocation planning method of any one of claims 1 to 8.