Optimization method and device for human-machine collaborative picking system

Abstract

The present invention discloses an optimization method for a human-machine collaborative picking system, comprising the following steps: S1. Modeling the three human-machine collaborative picking modes ("bundling," "picker-led," and "robot-led") based on a semi-open queuing network embedded in a Fork-Join queuing network to construct a queuing network model; S2. Based on the queuing network model, converting the Fork-Join queuing network into a load-based composite node; S3. Obtaining an analytical solution for performance indicators based on the composite node; and S4. Optimizing the picking system layout based on this. The present invention can effectively predict the performance of a human-machine collaborative picking system and help managers better optimize, select, and make scientific decisions about the human-machine collaborative picking system, thereby improving work efficiency and reducing operating costs.

Description

Technical Field

[0001] The present invention relates to the field of robot warehousing systems, and in particular to an optimization method and device for a human-machine collaborative picking system. Background Art

[0002] As a crucial fulcrum of the logistics network and a key link in the supply chain, the revitalization of the logistics industry depends on the innovative development of warehousing systems. Warehousing systems handle essential operations such as receiving, storing, picking, and shipping, making them an essential component of any supply chain. Picking is the most labor-intensive and expensive part of warehousing operations, characterized by high repetitiveness, complexity, and poor ergonomics. Statistics show that up to 55% of operating costs within distribution centers are attributable to picking costs, making it crucial for warehousing systems and process automation.

[0003] More and more practices and research indicate that the role of humans in the picking process cannot be ignored. To balance the flexibility of manual picking with the efficiency of automated picking, human-machine collaborative picking systems have emerged. Currently, this system has been implemented and applied by multiple companies, such as Fetch Robotics (a picking robot company), Locus Robotics (a warehouse robot company), 6River Systems (a logistics and warehouse robot company), and Toyota. According to a research report by ABI Research, a global technology intelligence company, by 2025, more than 50,000 warehouses worldwide will have installed 4 million robots, the majority of which will be collaborative robots.

[0004] In a human-robot collaborative picking system, pickers and robots work closely together to complete order picking. Pickers only need to complete the picking process, while robots load multiple transfer boxes and transport them to distribution points after picking batches of orders. This significantly reduces pickers' unproductive walking time, enabling them to complete their work more efficiently and ergonomically. From an investment perspective, human-robot collaborative picking systems offer good scalability and can quickly adapt to different work needs. Companies can gradually transition from manual picking to a collaborative picking system, significantly reducing both investment costs and the risks associated with automation. Furthermore, by adjusting strategies such as the human-robot ratio and the number of robots, the system can better cope with the fluctuating demand currently present in e-commerce warehouses.

[0005] At present, the academic community lacks systematic research on the modeling, simulation and optimization technologies of human-machine collaborative picking systems. There is little research on how to carry out the layout design, resource allocation, and operation strategy of human-machine collaborative picking systems. The travel time method in the existing technology cannot characterize the queuing phenomenon, and the simulation method is too time-consuming. Therefore, the existing technology lacks an accurate and fast method for optimizing the performance of human-machine collaborative picking systems to help managers make quick decisions. Summary of the Invention

[0006] The purpose of the present invention is to overcome the shortcomings of the prior art in the above-mentioned background technology that the prior art lacks an accurate and fast method for optimizing the performance of a human-machine collaborative picking system, and to provide an optimization method for a human-machine collaborative picking system.

[0007] To achieve the above object, the present invention adopts the following technical solutions:

[0008] An optimization method for a human-machine collaborative picking system comprises the following steps:

[0009] S1. Based on the semi-open queuing network embedded in the Fork-Join queuing network, the three human-machine collaborative picking modes of "bundling", "picker-led" and "robot-led" are modeled to construct a queuing network model;

[0010] S2. Based on the queuing network model, convert the Fork-join queuing network into a load-based composite node;

[0011] S3. Obtaining an analytical solution of a performance indicator based on the composite node;

[0012] S4. On this basis, optimize the picking system layout.

[0013] In some embodiments, step S1 includes:

[0014] S11. Divide the system into several service nodes based on the system operation processes of the three human-machine collaborative picking modes: "bundling," "picker-led," and "robot-led," and obtain the number of service desks and service rate of each service node;

[0015] S12. Calculate the first-order moment and second-order moment of the service time of each service node based on the actual system layout and size;

[0016] S13. According to the first-order moment and second-order moment of the service time of each service node, a queuing network model is constructed by embedding the Fork-Join queuing network in the semi-open queuing network.

[0017] In some embodiments, in step S11, the system operation process of the three human-machine collaborative picking modes is as follows:

[0018] The system operation process of the "bundled" human-machine collaborative picking mode is as follows: when a batch order arrives, the robot loads the transfer box or pallet and follows the human to pick the items. After the current order is picked, the robot returns to the distribution point, and the human continues to match the idle robot to complete the picking of the next batch;

[0019] The system operation process of the "robot-led" human-machine collaborative picking model is as follows: the picking area is divided into multiple blocks, each of which is managed by multiple pickers. When a batch order arrives, it is bound to an idle robot. When it passes through each block via an S-path, it converges with the picker in that block at the first picking point in the block and collaboratively completes the remaining picking within the block.

[0020] The "picker-led" human-robot collaborative picking model operates as follows: the picking area is divided into multiple blocks, each of which is handled by multiple robots. When a batch order arrives, it is assigned to a free picker. As the picker moves through each block along an S-shaped path, it rendezvouses with the robot assigned to that block at the first picking point within the block, collaborating to complete the remaining picking within the block.

[0021] In some embodiments, in step S13, the data required to calculate the first-order moment and the second-order moment of the service time include batch order data, picking area layout data, robot data, and picker data.

[0022] In some embodiments, step S2 includes the following steps:

[0023] S21. Calculate the number of states of the Fork-join queuing network according to the number of mobile resources in the half-open queuing network;

[0024] S22: removing waiting queues in each branch of the Fork-join queuing network, and converting the Fork-join queuing network into a load-based composite node using an approximate average value analysis method.

[0025] In some embodiments, step S3 includes the following steps:

[0026] S31. Define state variables according to system characteristics;

[0027] S32. Solve the semi-open queueing network model to obtain the steady-state probability value corresponding to each state variable;

[0028] S33. Based on the steady-state probability value, find an analytical solution for the performance indicator.

[0029] In some embodiments, the performance indicators include expected system throughput, expected mobile resource utilization, and expected length of an external queue.

[0030] In some embodiments, in step S32, the half-open queuing network model is solved using a matrix geometry method.

[0031] The present invention also provides an optimization device for a human-machine collaborative picking system, which specifically includes a processor and a storage medium for storing a computer program. The processor can implement the above-mentioned optimization method for the human-machine collaborative picking system by executing the computer program.

[0032] The present invention also provides a computer-readable storage medium for storing a computer program, wherein the computer program, when executed by the processor, can implement the above-mentioned optimization method of the human-machine collaborative picking system.

[0033] The present invention has the following beneficial effects: the optimization method of the human-machine collaborative picking system of the present invention is based on queuing network technology. By fully characterizing the characteristics of various types of job queues, it can effectively predict the performance of the human-machine collaborative picking system and realize the optimization of the human-machine collaborative picking system, thereby improving work efficiency and reducing operating costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 is a flow chart of an optimization method for a human-machine collaborative picking system according to an embodiment of the present invention;

[0035] Figure 2 is a plan view of a “bundled” human-machine collaborative picking mode in an embodiment of the present invention;

[0036] Figure 3 is a plan view of a “robot-led” human-machine collaborative picking mode in an embodiment of the present invention;

[0037] Figure 4 It is a plan view of the "picker-led" human-machine collaborative picking mode in an embodiment of the present invention;

[0038] Figure 5 Schematic diagram of a semi-open queuing network model of a "bundled" human-machine collaborative picking mode in an embodiment of the present invention;

[0039] Figure 6 Schematic diagram of a semi-open queuing network model of a "robot-led" human-machine collaborative picking mode in an embodiment of the present invention;

[0040] Figure 7 Schematic diagram of a semi-open queuing network model of a "picker-led" human-machine collaborative picking mode in an embodiment of the present invention;

[0041] Figure 8 is a schematic diagram of the relative error distribution of the analytical model relative to the simulation model with respect to the expected throughput time of the system in an embodiment of the present invention;

[0042] Figure 9is a schematic diagram of the relative error distribution of the analytical model in an embodiment of the present invention relative to the simulation model regarding the expected number of batch orders waiting in the external queue;

[0043] Figure 10 is a schematic diagram of the relative error distribution of the analytical model relative to the simulation model regarding the expected utilization of mobile resources in an embodiment of the present invention;

[0044] Figure 11 is a sub-matrix B of the generator matrix of the queuing network of the human-machine collaborative picking system in the embodiment of the present invention. 00 ;

[0045] Figure 12 is a submatrix B of the generator matrix of the queuing network of the human-machine collaborative picking system in an embodiment of the present invention;

[0046] The following are the descriptions of the reference numerals:

[0047] 1-Shelf, 2-Robot, 3-Workstation, 4-Picker, 5-Transfer box, 6-Block, 7-Conveyor belt. DETAILED DESCRIPTION

[0048] The following is a detailed description of the embodiments of the present invention. It should be emphasized that the following description is only exemplary and is not intended to limit the scope of the present invention and its application.

[0049] The terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of such features. In the description of the embodiments of the present invention, "multiple" and "several" mean two or more, unless otherwise specifically defined.

[0050] like Figure 1 As shown, an embodiment of the present invention proposes an optimization method for a human-machine collaborative picking system, which specifically includes the following steps:

[0051] S1. Based on the semi-open queuing network embedded in the Fork-Join queuing network, the three human-machine collaborative picking modes of "bundling", "picker-led" and "robot-led" are modeled to construct a queuing network model;

[0052] S11. Divide the system into several service nodes based on the system operation processes of the three human-machine collaborative picking modes ("bundling," "picker-led," and "robot-led"), and obtain the number of service desks and service rate of each service node:

[0053] refer to Figure 2 、 Figure 3 、 Figure 4The human-robot collaborative picking system includes: shelf 1, robot 2 (a collaborative robot), workstation 3, picker 4, transfer boxes 5, blocks 6, and conveyor belt 7. In the human-robot collaborative system, various inventory items are stored on shelf 1. The distribution point is located at the bottom of the block. The distribution point is represented by workstation 3 in the "bundling" and "robot-led" modes and by the conveyor belt in the "picker-led" mode. When a batch order arrives, picker 4 works closely with robot 2 to complete the order picking. The picker only needs to complete the picking process, while robot 2 loads multiple transfer boxes and transports them to the distribution point after the batch order is picked. Workstation 3 is responsible for packaging and labeling the goods in transfer boxes 5. A buffer is provided for the picking robots to queue. Robot 2 places transfer boxes 5 at workstation 3 and waits to be matched with the next batch order. The conveyor belt is used to sort the goods in the transfer boxes. The system operation process of the three human-robot collaborative picking modes is as follows:

[0054] like Figure 2 As shown, the system operation process of the "bundled" human-machine collaborative picking mode is that when a batch order arrives, robot 2 loads the transfer box 5 or pallet and follows the person to pick until the current order is picked. Then robot 2 returns to the distribution point, and the person continues to match the idle robot to complete the picking of the next batch.

[0055] like Figure 3 As shown, the system operation process of the "robot-led" human-machine collaborative picking mode is that the picking area is divided into multiple blocks 6, and each block is in charge of multiple pickers. When a batch order arrives, it is bound to an idle robot, and robot 2 starts picking from the leftmost side of the area. When it passes through each block via the S path, it converges with the picker 4 of the block at the first picking point in the block, and collaborates to complete the remaining picking in the block. Since the picker will stay at the last picking point of the current block when completing the picking task of the block he is responsible for, in order to reduce the time for human-machine convergence, the starting aisle for the robot to pick in the block is changed according to the position distribution of people in the aisle. In addition, since the walking path is an S path, picker 4 has p r The probability of staying in the rightmost lane of the block is 1-p r Stay in the leftmost lane, so robot 2 has p r The probability of crossing the block to the rightmost lane to start block picking.

[0056] like Figure 4As shown, the system operation process of the "picker-led" human-machine collaborative picking mode is that the picking area is divided into multiple blocks, and each block is in charge of multiple robots. When a batch order arrives, it is bound to an idle picker. When it passes through each block via the S path, it converges with the robot of the block at the first picking point in the block and collaborates to complete the remaining picking in the block. After completion, picker 4 goes to the next block, and robot 2 transports the items to the conveyor belt located below the area and stays in the buffer zone of the corresponding block. It is also stipulated that the order picking in the block starts from the position closest to the picker. Since picker 4 returns directly after completing the batch order picking, there is a p idle robot. l The probability is that the picker is not in the same lane, so it is necessary to cross the block to the lane where picker 4 is located to start block picking.

[0057] According to the system operation process of the three human-machine collaborative picking modes, the human-machine collaborative process can be divided into three stages: (1) When the batch order is assigned to the picker and the robot by the warehouse management system, the picker or robot respectively arrives at the first picking point of the batch order from its current location; (2) The robot and the picker meet at the first picking point and bind together to complete the picking of the remaining goods of the batch order; (3) After completing the batch order, the robot returns to the distribution point alone, and the picker stays at the last picking point of the batch order. Based on this, these three processes are abstracted as service nodes, and the number of service stations (i.e., the number of robots or pickers) and the service rate (i.e., the time required for the picker and robot to spend in each stage) of the service node are determined.

[0058] S12. Calculate the first-order moment and second-order moment of the service time of each service node based on the actual system layout and size;

[0059] When calculating the first-order moment and second-order moment of the service time of each service node, the required data mainly include: batch order data (arrival rate λ, number of orders n), picking area layout data (number of lanes L, number of shelves per lane M, shelf width f, distance between lanes w c , the distance from the shelf to the aisle w l , number of partitions N z ), collaborative robot data (number k, moving speed v r ), picker data (quantity p, moving speed v p , picking time t pIn the simulation, four batch order scenarios and three picking zone layouts were considered: a batch order arrival rate of 10 items per hour, no partitioning, 2 partitions, and 4 partitions; a batch order arrival rate of 20 hours, no partitioning, 2 partitions, and 4 partitions; a batch order arrival rate of 40, no partitioning, 2 partitions, and 4 partitions; and a batch order arrival rate of 60, no partitioning, 2 partitions, and 4 partitions. Based on these 12 basic scenarios, the number of pickers and robots was varied, resulting in a total of 60 scenarios. Table 1 below shows the basic warehouse information:

[0060] Table 1

[0061] n L M <![CDATA[v r ]]> <![CDATA[v p ]]> f <![CDATA[w c ]]> <![CDATA[w l ]]> <![CDATA[t p ]]> 6 1 3 3 0.7 1 1 3 3

[0062] S13. Based on the first-order moment and second-order moment of the service time of each service node, a queuing network model is constructed by embedding the Fork-Join queuing network in the semi-open queuing network:

[0063] Based on the semi-open queueing network and the Fork-join queueing network, draw the queueing network models of three human-machine collaborative picking modes, and mark and explain each service node, such as Figures 2 to 5 shown.

[0064] Figure 5 、 Figure 6 、 Figure 7 The semi-open queue network models for the three-machine collaborative picking modes of "bundling", "picker-led", and "robot-led" are shown respectively. The relevant nodes involved are detailed in Table 2:

[0065] Table 2

[0066]

[0067] S2. Based on the constructed queuing network model, the approximate average value analysis method is used to transform the Fork-join queuing network into a load-based composite node;

[0068] Based on the queueing network models of three human-machine collaborative picking modes, a Fork-join queueing network is nested in a semi-open queueing network. The Fork-join queueing network needs to be short-circuited to eliminate the waiting queues of each task branch. The approximate average value analysis method is used to aggregate the Fork-join queueing network into a load-based composite node, and the service rate of the review node is calculated. Converting the Fork-join queueing network into a load-based composite node includes:

[0069] S21. Calculate the number of states in the Fork-join queueing network based on the number of mobile resources in the half-open queueing network:

[0070] A closed queuing network containing N·k tasks is constructed by a short-circuited Fork-join queuing network. Where N is the number of branches in the Fork-join queuing network, and k is the number of robots flowing in the original network. If N represents the number of parallel branches, and each branch contains Q i nodes and k tasks, then for each branch, in Q i +1 node (including queue Q i The number of possible combinations of assigning k tasks in ) is calculated as follows:

[0071]

[0072] Therefore, the calculation method for allocating N·k tasks to all nodes is as follows:

[0073]

[0074] If each join queue Q i If there is only one robot, it will immediately go to the first station. If this is not possible, the number of states in the Fork-join queueing network is calculated as follows.

[0075]

[0076] S22. Remove the waiting queues in each branch of the Fork-join queuing network, so that the number of states of the transformed network is approximately equal to the number of mobile resources of the original network. Use the approximate average value analysis method to transform the Fork-join queuing network into a load-based composite node:

[0077] The queuing network obtained in S21 is further transformed to construct a closed queuing network with product solution, whose number of nodes is Determine the number of jobs to be K so that the number of states in the network is approximately equal to the number of states in the Fork-join queuing network. Because when the number of states in the two networks is approximately equal, the throughput of the two networks is λ PF (K) and λ FJ (k) is also approximately equal. Therefore, we determine the number of jobs K in a closed queueing network with a product solution by the following formula.

[0078] K:|Z FJ (k)-Z PF (K)|=m l in|Z FJ (k)-Z PF (l)|

[0079] Among them, Z PF (K) = z M-1 (K).

[0080] Furthermore, the approximate average method is used to solve the throughput λ of the closed queuing network. PF (K), and approximate it as the throughput λ of the Fork-join queueing network FJ (k).

[0081] S3. Based on the composite nodes aggregated from the Fork-join queuing network, we use the matrix geometry method to solve the semi-open queuing network containing two nodes and obtain the analytical solution of the indicator:

[0082] S31. Define state variables according to system characteristics;

[0083] The state variables of the semi-open queueing network after aggregation can be expressed as S j =(a,b,c), where a represents the number of jobs waiting in the external queue of the queuing network, i.e., the number of batch orders, b represents the number of jobs at the composite site after the Fork-join queuing network is transformed, and c represents the number of jobs at the second site. m is the number of mobile resources in the queuing network. The mobile resources are robots in the “bundling” and “robot-led” modes and pickers in the “picker-led” mode. The subscripts of the state variables are calculated using the following formula.

[0084]

[0085] Π0 represents the steady-state probability when there is no waiting in the external queue, Π i It represents the steady-state probability when the number of jobs waiting in the external queue is i.

[0086]

[0087] S32. Solve the semi-open queueing network model based on matrix geometry method to obtain the steady-state probability value corresponding to each state variable;

[0088] The specific form of the generation matrix for a human-machine collaborative picking system is as follows. Figure 11 、 Figure 12 Amplify B 00 Matrix and B matrix:

[0089]

[0090]

[0091] Where λ is the order arrival rate of the system, μ1(i) represents the service rate when there are i mobile resources in the first integrated node obtained by the approximate mean analysis method, μ2(i) represents the service rate when there are i mobile resources in the second integrated node, and the matrix B00 represents the transfer matrix from the steady-state probability vector Π0 to Π0, B 01 represents the transfer matrix from Π0 to the steady-state probability vector Π1, B 10 represents the transfer matrix from Π1 to Π0, and C is the transition matrix from Π i+1 To the steady-state probability vector Π i The transfer matrix, B is the probability vector Π from the steady state i to π i The transfer matrix, A is from π i to π i+1 The transfer matrix, B 00 is a dimension of ((N m +1)*(N m +2)) / 2×((N m +1)*(N m +2)) / 2 square matrix, B 01 and B 10 The dimensions are ((N m +1)*(N m +2) / 2)×(N m +1) and (N m +1)×((N m +1)(N m +2) / 2), B, C and A are all of dimension (N m +1)×(N m +1) square matrix.

[0092] R=-(C+R 2 A)B -1

[0093] R is the transfer matrix, and the pseudo code of its iterative solution algorithm is as follows. k is the value of the R matrix at the kth iteration, and ε is a sufficiently small threshold.

[0094]

[0095] That is, S1: initial assignment R0 = 0, R1 = -(C + R 2 A)B -1 , k=0, S2: Let K=K+1, calculate R k+1 The value of S3: judge || R k+1 -R k || ∞ Is it greater than ε, when ||R k+1 -R k || ∞ >ε, execute step S2, otherwise jump out of the loop and execute S4, S4: output R = R k value.

[0096] By solving the following system of equations, we can find Π0 and Π1, where e is a 1×(N m +1)(N m +2) / 2, where all elements are 1, and e' is a column vector of dimension 1×(N m +1)*(N m +2) is a column vector whose elements are all 1, and I is a column vector with dimension (N m +1)×(N m +1) of the identity matrix.

[0097]

[0098] The other steady-state probability vectors can be calculated as follows: i =Π i-1 R.

[0099] S33. Based on the matrix geometry method, the steady-state probability value of each state variable is obtained and the analytical solution of the performance index is obtained:

[0100] L eq = Π1(IR) -2 e

[0101]

[0102] Among them, the random variable L eq represents the expected number of batch orders waiting in the external queue, L DC represents the expected number of busy mobile resources, represents the expected waiting time of mobile resources in the external queue, THT DC Indicates the system's expected throughput time, represents the expected utilization of mobile resources, π(0,b,c) represents the steady-state probability of the state variable when the number of jobs waiting in the external queue represented by (0,ib,c) is 0, the number of jobs at the composite site after the Fork-join queuing network transformation is b, and the number of jobs at the second site is c.

[0103] S4. On this basis, optimize the picking system layout.

[0104] By solving the aforementioned performance evaluation indicators, performance predictions can be provided for the layout design, resource allocation, and operational strategy optimization of the human-robot collaborative picking system, thereby optimizing the layout of the human-robot collaborative picking system. Based on queuing network technology, the analysis method proposed in this embodiment of the present invention can quickly and accurately solve the system performance under the three human-robot collaborative picking modes, and based on this, perform layout optimization. It can also predict and optimize resource allocation and operational strategies, providing decision support tools and methods.

[0105] In the simulation software MATLAB R2019a, a system simulation model of three human-machine collaborative picking modes was constructed. By changing the batch order data (4 arrival rates), the layout of the picking area (3 layouts), and the number of pickers and robots (several pairings), 60 scenarios were generated. In each scenario, the running warm-up time of each simulation scenario was 100 hours and the running time was 1000 hours. The relative error was used to calculate the error. To evaluate the accuracy of the analytical model in this embodiment, R a and R s Corresponding to analytical results and simulation results, respectively. Figure 8 、 Figure 9 、 Figure 10 The expected throughput time THT of the system under 60 scenarios DC , the expected number of batch orders waiting in the external queue L eq , expected utilization of mobile resources Frequency plot of the relative error distribution intervals of the three indicators. The simulation results can verify the accuracy of the system performance prediction of the three human-machine collaborative picking modes in the embodiment of the present invention, and further verify the reliability of using the predicted performance in this embodiment to optimize the human-machine collaborative picking system.

[0106] An embodiment of the present invention further provides a processor, which can implement the above-mentioned optimization method of the human-machine collaborative picking system by executing a computer program.

[0107] An embodiment of the present invention further provides a computer-readable storage medium for storing a computer program. The computer program can be executed by a processor, and the computer program can implement the above-mentioned optimization method of the human-machine collaborative picking system during the execution of the computer program by the processor.

[0108] An embodiment of the present invention further provides an analysis device, comprising a processor and a storage medium for storing a computer program. The processor can implement the above-mentioned optimization method of the human-machine collaborative picking system by executing the computer program.

[0109] To address the lack of professional decision-making support tools and methods during the design and selection phase of human-machine collaborative picking warehousing systems, the present invention proposes a highly promising method for optimizing human-machine collaborative picking systems by using queuing network technology to predict and optimize the performance of these systems. This method analyzes the performance of systems encompassing three human-machine collaborative picking modes, evaluates the system, and optimizes its layout and resource allocation based on this performance. It also enables strategic design for efficient resource operation, thereby improving the efficiency and reducing the cost of the human-machine collaborative picking system.

[0110] The performance prediction method for the human-machine collaborative picking system proposed in this invention can not only quickly and accurately solve a variety of human-machine collaborative picking systems, but also explore spatial morphological layout and resource allocation solutions that match business needs through multi-scenario analysis.

[0111] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0112] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0113] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0114] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0115] The above description further details the present invention in conjunction with specific / preferred embodiments, and the specific implementation of the present invention should not be construed as being limited to these descriptions. Persons skilled in the art will appreciate that, without departing from the spirit of the present invention, they may make various substitutions or modifications to the described embodiments, and these substitutions or modifications should be considered to fall within the scope of protection of the present invention. Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "preferred embodiments," "examples," "specific examples," or "some examples" indicates that the specific features, structures, materials, or characteristics described in conjunction with such embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of these terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Persons skilled in the art may combine and assemble the different embodiments or examples described in this specification, as well as features of different embodiments or examples, without conflicting opinions. Although the embodiments of the present invention and their advantages have been described in detail, it should be understood that various changes, substitutions, and modifications may be made herein without departing from the scope of protection of the patent application.

Claims

1. An optimization method for a human-machine collaborative picking system, characterized in that: The following steps are involved: S1. Based on the semi-open queuing network embedded in the Fork-Join queuing network, the system is modeled for three human-machine collaborative picking modes: "bundling", "picker-led", and "robot-led", and a queuing network model is constructed; S2. Based on the queuing network model, convert the Fork-join queuing network into a load-based composite node; S3. Obtaining an analytical solution of a performance indicator based on the composite node; S4. Optimize the picking system layout based on this; in, Step S2 includes the following steps: S21. Calculate the number of states of the Fork-join queuing network according to the number of mobile resources in the half-open queuing network; S22, remove the waiting queues in each branch of the Fork-join queuing network, and use the approximate average analysis method to transform the Fork-join queuing network into a load-based composite node; further transform the queuing network obtained in S21 to construct a closed queuing network with a product solution, the number of nodes of which is The number of jobs is determined to be K, so that the number of states in the network is approximately equal to the number of states in the Fork-join queuing network. The number of jobs K in the closed queuing network with product solution is determined by the following formula: Among them, Z PF (K) = z M-1 (K), N is the number of branches in the Fork-join queuing network, Z FJ (k) is the number of states in the Fork-join queueing network.

2. The optimization method of a human-machine collaborative picking system according to claim 1, characterized in that: Step S1 includes: S11. Divide the system into several service nodes based on the system operation processes of the three human-machine collaborative picking modes: "bundled," "picker-led," and "robot-led." Obtain the number of service desks and service rate for each service node. S12. Calculate the first-order moment and second-order moment of the service time of each service node based on the actual system layout and size; S13. According to the first-order moment and second-order moment of the service time of each service node, a queuing network model is constructed by embedding the Fork-Join queuing network in the semi-open queuing network.

3. The optimization method of a human-machine collaborative picking system according to claim 2, characterized in that: In step S11, the system operation process of the three human-machine collaborative picking modes is as follows: The system operation process of the "bundled" human-robot collaborative picking mode is as follows: when a batch of orders arrives, the robot loads the transfer box or pallet and follows the human to pick the items. After the current order is picked, the robot returns to the distribution point, and the human continues to match the idle robot to complete the next batch of picking. The system operation process of the "robot-led" human-machine collaborative picking model is as follows: the picking area is divided into multiple blocks, each of which is managed by multiple pickers. When a batch order arrives, it is bound to an idle robot. As it passes through each block via an S-path, it converges with the picker in that block at the first picking point within the block and collaboratively completes the remaining picking within the block. The system operation process of the "picker-led" human-machine collaborative picking mode is as follows: the picking area is divided into multiple blocks, each block is managed by multiple robots. When a batch order arrives, it is bound to an idle picker. When it passes through each block via the S path, it converges with the robot of the block at the first picking point in the block and collaborates to complete the remaining picking in the block.

4. The optimization method of a human-machine collaborative picking system according to claim 2, characterized in that: In step S13, the data required for calculating the first-order moment and the second-order moment of the service time include batch order data, picking area layout data, robot data, and picker data.

5. The optimization method of a human-machine collaborative picking system according to claim 1, characterized in that: Step S3 includes the following steps: S31. Define state variables according to system characteristics; S32. Solve the semi-open queueing network model to obtain the steady-state probability value corresponding to each state variable; S33. Based on the steady-state probability value, find an analytical solution for the performance indicator.

6. The optimization method of a human-machine collaborative picking system according to claim 5, characterized in that: In step S33, the performance indicators are the expected throughput time of the system, the expected number of batch orders waiting in the external queue, and the expected utilization rate of mobile resources.

7. The optimization method of a human-machine collaborative picking system according to claim 5, characterized in that: In step S32, the half-open queue network model is solved using a matrix geometry method.

8. An optimization device for a human-machine collaborative picking system, characterized in that , specifically including a processor and a storage medium for storing a computer program. The processor can implement the optimization method of the human-machine collaborative picking system according to any one of claims 1 to 7 by executing the computer program.

9. A computer-readable storage medium for storing a computer program, characterized in that , the computer program can implement the optimization method of the human-machine collaborative picking system described in any one of claims 1-7 when being executed by the processor.

Images