Method for order picking optimization of man-machine cooperation in movable shelf intelligent warehouse
By encoding shelves and robots and combining tabu search and neighborhood search algorithms, the joint optimization problem of human-machine collaboration in mobile shelf smart warehouses is solved, thereby improving order picking efficiency.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- NORTHEASTERN UNIV CHINA
- Filing Date
- 2023-09-15
- Publication Date
- 2026-07-21
Smart Images
Figure CN117273327B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of operations research and optimization technology, and more particularly to an order picking optimization method for human-machine collaboration in a mobile shelving smart warehouse. Background Technology
[0002] With the widespread penetration and popularization of the internet, online shopping has become an indispensable part of people's daily lives. This trend has led to a continuous increase in orders for a large number of diverse, small-batch goods, thus bringing a series of order picking optimization problems to logistics and warehousing companies. The main task of order picking is to retrieve the goods required by an order from the warehouse, which is a typical labor-intensive task and one of the main sources of warehouse operating costs. At the same time, the shift in customer demand from large-volume, low-batch orders to small-batch, personalized orders further increases the difficulty of order picking. How to efficiently complete order picking tasks is of great significance for reducing warehouse operating costs. With the continuous development of robotics technology, mobile robots are increasingly being used in order handling and delivery. They have the advantages of requiring little space, high flexibility, and unrestricted working hours, making them very suitable for e-commerce operations.
[0003] Mobile robot-assisted mobile shelving intelligent warehousing systems transform labor-intensive manual sorting into intelligent sorting through human-machine collaboration with robot assistance, achieving order picking efficiency 2-3 times that of traditional warehousing picking systems. Current research on operational optimization for this system primarily focuses on the rational arrangement of shelving distribution, order batching, robot task allocation, and path planning within a given warehouse environment. It lacks consideration of joint optimization involving shelving selection, robot task assignment, and picking worker task arrangement, and also lacks high-performance solutions for such joint optimization problems.
[0004] In view of this, the present invention proposes an optimized order picking method for human-machine collaboration in a mobile shelf smart warehouse. Summary of the Invention
[0005] To address the shortcomings mentioned above, this invention provides a human-machine collaborative order picking optimization method in a mobile shelving intelligent warehouse. This invention primarily achieves joint optimization of three sub-problems: shelving selection, robot task assignment, and order sorting through problem modeling, encoding / decoding design, and an improved tabu search algorithm.
[0006] The technical means employed in this invention are as follows:
[0007] This invention provides a human-machine collaborative order picking optimization method in a mobile shelving smart warehouse, comprising:
[0008] Encode P shelves, where P is a positive integer;
[0009] Calculate the valid sequence;
[0010] Encode X robots, where X is a positive integer;
[0011] When the valid sequence includes P1 shelf codes and X1 robots are in an idle state, if P1 equals X1, the X1 robots move the shelves sequentially according to the valid sequence, where P1 ≤ P and P1 is a positive integer, and X1 ≤ X and X1 is a positive integer;
[0012] If P1 is less than X1, sort the X1 robots in ascending order of task volume to obtain a robot sequence. Select the first P1 robots in the robot sequence, and the P1 robots move the shelf in sequence according to the effective sequence.
[0013] If P1 is greater than X1, the effective sequence includes a first part and a second part. X1 robots first move the shelf in sequence according to the first part, and then find an idle robot in sequence to move it in the second part. Among the X1 robots, the robot that has completed the moving task and is returning to its original position is the idle robot.
[0014] Based on the order, the picking process is carried out using a greedy picking method, following the order in which the shelves arrive at the picking station.
[0015] Further, calculating the effective sequence includes:
[0016] Establish mathematical models for the shelf, the robot, and the order;
[0017] Initialize the taboo list, initialize long-term memory, determine long-term memory penalty factors, and determine the maximum and minimum taboo times.
[0018] The initial solution of the mathematical model is obtained based on the tabu list, the long-term memory, the long-term memory penalty factor, the maximum tabu time, and the minimum tabu time, and the fitness function of the initial solution is calculated.
[0019] At least one candidate solution is obtained by using a neighborhood search mechanism on the initial solution;
[0020] Determine whether the optimal solution among the candidate solutions is better than the current global optimal solution;
[0021] If so, select the optimal solution from the candidate solutions and update the optimal solution from the candidate solutions to the current solution, wherein the current solution includes the valid sequence and the invalid sequence.
[0022] Furthermore, when calculating the initial solution, the distance from the picking station to each of the shelves is obtained, the types of goods required for each order and their corresponding quantities, and the types of goods contained in each shelf and their corresponding quantities.
[0023] Furthermore, a neighborhood search mechanism is applied to the initial solution to obtain at least one candidate solution, including:
[0024] In the P shelf codes, if at least one of the shelf codes in any two adjacent shelf codes belongs to the shelf to be moved, the two adjacent shelf codes are swapped using 2-opt, so that the shelf code corresponding to the shelf to be moved is close to or enters the valid area, and the shelf code corresponding to the shelf to be moved does not need to be close to or enter the invalid area.
[0025] When the effective region and the invalid region remain unchanged, a candidate solution is generated.
[0026] Furthermore, a mathematical model is established for the shelf, the robot, and the order, calculated in the following manner:
[0027]
[0028] in, To complete the total picking time, To complete the picking task for the p-th shelf, The time point at which the picking task begins from the p-th shelf. The time required to complete the picking task corresponding to the p-th shelf, The time point at which the picking task for the (p-1)th shelf is completed, The time point at which the p-th shelf is delivered to the picking station; Let p be the start time for the robot to move the p-th shelf. The time required for the robot to complete the handling of the p-th shelf. The robot that is moving the p-th shelf will remember the time point at which the shelf was last transported to the picking station.
[0029] Furthermore, based on the order, picking is performed using a greedy picking method according to the order in which the shelves arrive at the picking station, including:
[0030] For each of the aforementioned shelves that arrives at the picking station, all the orders and the corresponding product types and quantities for each shelf are located, and the products are picked according to the product types and quantities.
[0031] Furthermore, all X robots have the same moving speed and carrying capacity, and each robot moves one shelf at a time.
[0032] Compared with the prior art, the present invention has the following advantages:
[0033] 1. The order picking optimization method for human-machine collaboration in a mobile shelf intelligent warehouse provided by this invention realizes the joint optimization of three sub-problems: shelf selection, robot task assignment, and order sorting.
[0034] 2. The order picking optimization method for human-machine collaboration in a mobile shelf intelligent warehouse provided by the present invention overcomes the difficulty of solving large-scale problems in traditional methods based on shelf selection strategy, robot task assignment strategy and sorting strategy based on greedy picking.
[0035] 3. The human-machine collaborative order picking optimization method in the mobile shelf intelligent warehouse provided by the present invention adopts a neighborhood search mechanism, which can more accurately locate and adjust key elements, and avoid the impact of invalid exchanges on the algorithm search efficiency through the neighborhood search mechanism.
[0036] 4. The human-machine collaborative order picking optimization method in the mobile shelf intelligent warehouse provided by this invention demonstrates a strong search capability in the early stage of iteration by calculating the effective sequence, and has the ability to quickly find a better solution, showing strong optimization capability and stable performance. Attached Figure Description
[0037] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0038] Figure 1 This is a flowchart illustrating a human-machine collaborative order picking optimization method in a mobile shelf intelligent warehouse provided by the present invention.
[0039] Figure 2 This is a flowchart illustrating the workflow of the mobile robot-assisted mobile shelving intelligent warehousing system of the present invention.
[0040] Figure 3 This is a three-stage decomposition diagram of the collaborative optimization problem of the present invention.
[0041] Figure 4 This is a diagram illustrating the three-stage analysis and strategy construction for the collaborative optimization problem in this invention.
[0042] Figure 5 This is a flowchart of the improved tabu search algorithm of the present invention.
[0043] Figure 6 The coding structure diagram for the collaborative optimization problem. Detailed Implementation
[0044] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.
[0045] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0046] Combination Figure 1 , Figure 1 This is a flowchart illustrating a human-machine collaborative order picking optimization method in a mobile rack intelligent warehouse provided by the present invention, to illustrate a specific embodiment of the human-machine collaborative order picking optimization method in a mobile rack intelligent warehouse provided by the present invention, including:
[0047] S1: Encode P shelves, where P is a positive integer;
[0048] S2: Calculate the valid sequence;
[0049] S3: Encode X robots, where X is a positive integer;
[0050] S41: When the valid sequence includes P1 shelf codes and X1 robots are in an idle state, if P1 equals X1, the X1 robots move the shelves in sequence according to the valid sequence, P1≤P and P1 is a positive integer, X1≤X and X1 is a positive integer;
[0051] S42: If P1 is less than X1, sort the X1 robots in ascending order of task volume to obtain a robot sequence. Select the first P1 robots in the robot sequence, and the P1 robots move the shelves in sequence according to the valid sequence.
[0052] S43: If P1 is greater than X1, the valid sequence includes the first part and the second part. X1 robots first move the shelves in order according to the first part, and then find idle robots in order to move them in the second part. Among the X1 robots, the robot that has completed the moving task and is returning to its position is the idle robot.
[0053] S5: Based on the order, pick the goods using a greedy picking method according to the order in which the shelves arrive at the picking station.
[0054] It is understandable that P shelves are encoded using an adaptive metaheuristic sequence encoding method, with each shelf receiving a unique integer encoding. The codes of all shelves are then arranged to form a single sequence encoding [I1, I2, I3, ..., I...]. n Based on the single-sequence encoding, a shelf selection strategy is implemented, sequentially selecting [I1], [I1,I2], ..., until a set of shelves [I1,I2,I3,...,I4] that satisfies the order constraints is found. m If m ≤ n, then the first m movable shelves are the shelves that the robot needs to handle. By shuffling the shelf coding sequence, different sets of shelves that meet the requirements can be obtained. The robot is then assigned to handle the first m movable shelves. Due to different transport distances, the order in which the shelves arrive at the picking station is [I1″, I2″, I3″, ... I...]. m "] and the initial shelf arrangement order [I1′,I2′,I3′,…I m The order sequence may differ from the given sequence. To assign picking tasks to picking workers, the order sequence is first encoded as [o1, o2, o3, ..., o...]. p Pickers proceed with picking items from the movable shelves in the order they arrive at the picking station. For each arriving shelf, pickers use a greedy picking method, taking as many items as possible from the shelf that meet the order requirements, until the shelf is empty. Picking then ends for that shelf, and the process moves on to the next shelf. This continues until all shelves have been processed, completing the picking task for the pickers.
[0055] In some alternative embodiments, refer to Figure 2 , Figure 3 , Figure 4 and Figure 5 , Figure 2 This is a flowchart illustrating the workflow of the mobile robot-assisted mobile shelving intelligent warehousing system of the present invention. Figure 3 This is a three-stage decomposition diagram of the collaborative optimization problem of this invention. Figure 4 This diagram illustrates the three-stage analysis and strategy construction for the collaborative optimization problem according to the present invention. Figure 5 The flowchart of the improved tabu search algorithm of this invention is shown below. Calculating the valid sequence includes:
[0056] Establish mathematical models for shelves, robots, and orders;
[0057] Initialize the taboo list, initialize long-term memory, determine long-term memory penalty factors, and determine the maximum and minimum taboo times.
[0058] The initial solution of the mathematical model is obtained based on the taboo list, long-term memory, long-term memory penalty factor, maximum taboo time and minimum taboo time, and the fitness function of the initial solution is calculated.
[0059] At least one candidate solution is obtained by using a neighborhood search mechanism on the initial solution;
[0060] Determine whether the optimal solution among the candidate solutions is better than the current global optimal solution;
[0061] If so, select the optimal solution from the candidate solutions and update the optimal solution from the candidate solutions to the current solution, wherein the current solution includes the valid sequence and the invalid sequence.
[0062] Understandably, the current solution is calculated using a dynamic change mechanism for the tabu list length, such as... Figure 5 As shown, the dynamic tabu list length change mechanism improves upon the traditional tabu search algorithm's fixed tabu list length. It sets a maximum and minimum tabu list length, and in each iteration, randomly sets the tabu list length based on these constraints. By dynamically changing the tabu list length, the algorithm can ensure that solutions trapped in local optima are tabulated, while simultaneously reducing the total tabu list length.
[0063] After the number of iterations reaches the tabu length, the algorithm removes the tabu to the previously tabu solution space. However, this strategy can still lead to local optima under certain circumstances. To avoid this and enhance the algorithm's search capability, this invention adds long-term memory to the short-term memory of the tabu table to constrain the solution space. Long-term memory restricts the solution space by storing search actions from the start of the algorithm's iterations. When evaluating the domain solution, a penalty F×f(m) is added, where f(m) is the total number of moves since the algorithm's inception, and the penalty factor F is a new parameter that needs to be adjusted in this method. In this mechanism, the algorithm prioritizes movement strategies with fewer moves from the start of the search, thereby increasing the algorithm's ability to escape local optima.
[0064] In some optional embodiments, when calculating the initial solution, the distance from the picking station to each shelf, the types of goods required for each order and their corresponding quantities, and the types of goods contained in each shelf and their corresponding quantities are obtained.
[0065] In some alternative embodiments, refer to Figure 5 andFigure 6 , Figure 6 To encode the structure graph of the collaborative optimization problem, a neighborhood search mechanism is used to obtain at least one candidate solution from the initial solution, including:
[0066] In P shelf codes, if at least one shelf code in any two adjacent shelf codes belongs to the shelf to be moved, perform a 2-opt swap on the two adjacent shelf codes to make the shelf code corresponding to the shelf to be moved close to or enter the valid area, and the shelf code corresponding to the shelf to be moved does not need to be moved close to or enter the invalid area.
[0067] When the effective region and the invalid region remain unchanged, a candidate solution is generated.
[0068] It is understandable that a single sequence is encoded as [I1′, I2′, I3′, ... I... n Perform 2-opt swapping on the encoded sequence [I1′, I2′, I3′, ..., I′], and a portion of the solution becomes the valid region [I1′, I2′, I3′, ..., I′]. m [′] indicates the shelf that needs to be moved, while the other part is an invalid area. m+1 ′, I m+2 ′, I m+3 ′,…I n When the neighborhood operator modifies only the invalid region in the latter half, no new feasible solution can be generated. To avoid this, this paper uses the 2-opt exchange (i.e., pairwise exchange) operator as a basis. When both endpoints of the exchange are in the invalid region, no new solution can be generated. Therefore, this invention stipulates that during the neighborhood search process, at least one endpoint must be in the valid region before the current neighborhood search is considered. This neighborhood search design avoids the impact of invalid exchanges on the algorithm's search efficiency.
[0069] In some optional embodiments, mathematical models of the shelves, robots, and orders are established and calculated in the following manner:
[0070]
[0071] in, To complete the total picking time, To complete the picking task for the p-th shelf, This is the time point at which the picking task begins from the p-th shelf. The time required to complete the picking task corresponding to the p-th shelf, The time point at which the picking task for the (p-1)th shelf is completed, This represents the time point at which the p-th shelf is delivered to the picking station. Let p be the start time for the robot to move the p-th shelf. The time required for the robot to complete the movement of the p-th shelf. The robot that moves the p-th shelf will remember the time when it last delivered the shelf to the picking station.
[0072] Understandable This represents the minimum time required for a picker to complete all shelf picking tasks. This means that the time it takes for the picker to complete the picking task of the p-th shelf is equal to the total time it takes for the picker to start picking from the p-th shelf and the time consumed in picking from the p-th shelf. This indicates that the time when the picker starts picking from the p-th shelf is no earlier than the time when the picker completes the picking task from the (p-1)-th shelf; This means that the time when the picker starts picking from the p-th shelf is no earlier than the time when the robot delivers the p-th shelf to the picking station; This means that the time it takes for the robot to deliver the p-th shelf to the picking station is equal to the sum of the time when the robot starts transporting the p-th shelf and the robot's transfer time. The robot's transfer time represents the total time required for the robot to travel from the picking station to the location of the p-th shelf and then transport it to the picking station. This means that each robot can only transfer one shelf at a time, meaning that it can only start transferring the next shelf after completing the transfer of the previous shelf.
[0073] In some alternative embodiments, picking is performed according to the order in which the shelves arrive at the picking station, using a greedy picking method, including:
[0074] When a shelf arrives at the picking station, locate all orders and their corresponding product types and quantities, and then pick the items according to the product types and quantities.
[0075] In some alternative embodiments, X robots move at the same speed and have the same carrying capacity, and each robot moves one shelf at a time.
[0076] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.
[0077] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.
[0078] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.
Claims
1. A method for optimizing order picking through human-machine collaboration in a mobile shelving smart warehouse, characterized in that, include: Encode P shelves, where P is a positive integer; Calculate the valid sequence; including: Establish a mathematical model for the shelf, the robot, and the order, and calculate it in the following manner: ； in, To complete the total picking time, To complete the picking task for the p-th shelf, The time point at which the picking task begins from the p-th shelf. The time required to complete the picking task corresponding to the p-th shelf, The time point at which the picking task for the (p-1)th shelf is completed, This refers to the time point at which the p-th shelf is delivered to the picking station; Let p be the start time for the robot to move the p-th shelf. The time required for the robot to complete the handling of the p-th shelf. The time point at which the robot carrying the p-th shelf last delivered the shelf to the picking station; Initialize the taboo list, initialize long-term memory, determine long-term memory penalty factors, and determine the maximum and minimum taboo times. The initial solution of the mathematical model is obtained based on the tabu list, the long-term memory, the long-term memory penalty factor, the maximum tabu time, and the minimum tabu time, and the fitness function of the initial solution is calculated. At least one candidate solution is obtained by using a neighborhood search mechanism on the initial solution; Determine whether the optimal solution among the candidate solutions is better than the current global optimal solution; If so, select the optimal solution from the candidate solutions and update the optimal solution from the candidate solutions to the current solution, wherein the current solution includes the valid sequence and the invalid sequence; Encode X robots, where X is a positive integer; When the valid sequence includes P1 shelf codes and X1 robots are in an idle state, if P1 equals X1, the X1 robots move the shelves sequentially according to the valid sequence, where P1 ≤ P and P1 is a positive integer, and X1 ≤ X and X1 is a positive integer; If P1 is less than X1, sort the X1 robots in ascending order of task volume to obtain a robot sequence. Select the first P1 robots in the robot sequence, and the P1 robots move the shelf in sequence according to the effective sequence. If P1 is greater than X1, the effective sequence includes a first part and a second part. X1 robots first move the shelf in sequence according to the first part, and then find an idle robot in sequence to move it in the second part. Among the X1 robots, the robot that has completed the moving task and is returning to its original position is the idle robot. Based on the order, the picking process is carried out using a greedy picking method, following the order in which the shelves arrive at the picking station.
2. The method for optimizing order picking through human-machine collaboration in a mobile shelving intelligent warehouse according to claim 1, characterized in that, When calculating the initial solution, the distance from the picking station to each shelf is obtained, the types of goods required for each order and their corresponding quantities, and the types of goods contained in each shelf and their corresponding quantities.
3. The method for optimizing order picking through human-machine collaboration in a mobile shelving intelligent warehouse according to claim 1, characterized in that, At least one candidate solution is obtained by applying a neighborhood search mechanism to the initial solution, including: In the P shelf codes, if at least one of the shelf codes in any two adjacent shelf codes belongs to the shelf to be moved, the two adjacent shelf codes are swapped using 2-opt, so that the shelf code corresponding to the shelf to be moved is close to or enters the valid area, and the shelf code corresponding to the shelf to be moved does not need to be close to or enter the invalid area. When the effective region and the invalid region remain unchanged, a candidate solution is generated.
4. The method for optimizing order picking through human-machine collaboration in a mobile shelving intelligent warehouse according to claim 1, characterized in that, According to the order, the picking process is performed using a greedy picking method, following the order in which the shelves arrive at the picking station, including: For each of the aforementioned shelves that arrives at the picking station, all the orders and the corresponding product types and quantities for each shelf are located, and the products are picked according to the product types and quantities.
5. The method for optimizing order picking through human-machine collaboration in a mobile shelving intelligent warehouse according to claim 1, characterized in that, X robots have the same moving speed and carrying capacity, and each robot moves one shelf at a time.