Order processing decision method and device based on medical and pharmaceutical distribution industry and storage medium

Abstract

The application relates to an order processing decision method and device based on a medicine circulation industry and a storage medium, and is applied to the technical field of medicine logistics automation warehousing, and comprises the following steps: the similarity coefficients between orders are calculated, orders with high coupling degrees are distributed to the same picking table, and the possibility of box reuse is enhanced; for the internal orders of each picking table, a single-picking-table order sorting algorithm based on neighborhood search is adopted, neighborhood operations such as local exchange and segment flipping are used to optimize the order sequence, and the quality and stability of order sorting are improved; on the basis of the completion of internal order sorting of each picking table, the global synchronization and conflict-free execution logic of each stage is taken as the core, the picking demands of multiple picking tables are combined for state expansion, a global candidate state set is constructed, a Cartesian product combination and conflict detection are performed, the box conflict states are effectively removed, and through global box sorting, parallel and uncoupled efficient picking is realized.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of medical logistics automation warehousing, and particularly relates to an order processing decision method and device based on a medical circulation industry and a storage medium. BACKGROUND

[0002] Compared with other industries, the medical logistics is extremely strict in timeliness, and how to efficiently complete a large number of picking operations of disassembled orders in a limited time becomes a great challenge faced by the industry. The “goods-to-person” picking system (SBS / RS) based on a multi-layer shuttle vehicle is widely used in the disassembled picking of the medical industry due to its efficient in-out warehouse frequency and rapid order processing capacity. The system takes out the bins from the high-density storage shelves through the multi-layer shuttle vehicle and the elevator and transports them to the picking table through the conveying line, and the picker picks the goods at the picking table. However, in the medical industry, as a special commodity, the flow process of the medicine is strictly limited by the batch number management specification. According to the requirements of the “Pharmaceutical Quality Management Specification”, the in-out warehouse of the medicine must follow the principle of first-in first-out, and different batches of medicine cannot be stored in the same bin. These rules mean that for the same batch of medicine, the front batch number must be completed before the rear batch number, and when the last bin of a batch is needed by orders of multiple picking tables, the bin coupling situation will occur, which will reduce the picking efficiency to some extent, and the high single-row characteristics of the order in the medical industry will also exacerbate this situation to some extent.

[0003] In the SBS / RS, the system extracts the required bin from the high-density shelf through the multi-layer shuttle vehicle, transports it to the picking table through the conveying line, and completes the order picking operation by the picker. After completing the picking of each bin, the system decides whether to temporarily store it in the cache or directly return to the warehouse according to whether the next order still needs the goods in the bin. The system usually configures multiple picking tables, and a single picking table adopts a vs b mode, that is, a bin picking position and b order box positions are set. However, in actual operation, a single picking table needs to supplement the next batch of orders after completing the current batch of order picking, and unreasonable order sequence will lead to frequent bin replacement, causing long waiting time and significantly reducing efficiency. For the multi-picking table parallel operation scene, when different picking tables pick orders, the situation that two different orders of two picking tables need the same bin at the same time may occur, that is, order coupling occurs. Once the bin is locked by a picking table, only after the picking table releases the bin, other picking tables can pick the bin. SKU SKU Before the bin is released, the remaining picking tables can only be in a waiting state. This not only causes waste of picking resources, but also causes low efficiency of the entire picking system. SKU

[0004] ​​However, the existing order allocation research mainly focuses on the optimization of picking station processing speed and the minimization of shelf access volume, but has not considered the influence of order overlapping characteristics on order allocation. SKU Secondly, the existing order sequencing research mostly focuses on the optimization of shelf / position level sequencing, and lacks consideration of bin replacement frequency and waiting bottleneck. Thirdly, the joint optimization of order allocation, order sequencing and storage unit sequencing mostly focuses on resource coordination of the RMFS system, and lacks research on the bin coupling problem in the SBS / RS system. In particular, in the pharmaceutical industry, due to the high-frequency demand and high-precision picking characteristics of medicines, bin coupling is more significant, and without reasonable sequencing strategies, frequent waiting and system congestion will be caused. SUMMARY

[0005] Therefore, the present application aims to provide an order processing decision method, device and storage medium based on the pharmaceutical distribution industry, to solve the problem that the existing order allocation research mainly focuses on the optimization of picking station processing speed and the minimization of shelf access volume, but has not considered the influence of order overlapping characteristics on order allocation. SKU Secondly, the existing order sequencing research mostly focuses on the optimization of shelf / position level sequencing, and lacks consideration of bin replacement frequency and waiting bottleneck. Thirdly, the joint optimization of order allocation, order sequencing and storage unit sequencing mostly focuses on resource coordination of the RMFS system, and lacks research on the bin coupling problem in the SBS / RS system. In particular, in the pharmaceutical industry, due to the high-frequency demand and high-precision picking characteristics of medicines, bin coupling is more significant, and without reasonable sequencing strategies, frequent waiting and system congestion will be caused.

[0006] According to a first aspect of an embodiment of the present application, an order processing decision method based on the pharmaceutical distribution industry is provided, the method comprising:

[0007] For all orders in a batch to be processed, a seed order is assigned to each picking station according to the similarity coefficient between any two orders;

[0008] For the remaining orders that are not assigned as seed orders, the number of orders of the same goods category as the seed order assigned to each picking station is calculated for each order as a coupling factor, and each remaining order is assigned to the picking station with the maximum coupling factor, thereby completing the order allocation;

[0009] For the multiple orders assigned to each picking station, a random method is used to generate multiple candidate order sequences to form an initial candidate set, and a local optimization is performed on each candidate order sequence in the initial candidate set using neighborhood search to obtain a preliminary local optimal order sequence;

[0010] The preliminary local optimal order sequence is subjected to non-adjacent order exchange and local segment flip operations in sequence to obtain a final candidate set, and the order sequence with the least bin replacement frequency is selected from the final candidate set as the optimal order sequence of the picking station.

[0011] In each picking station, an initial state set of each picking station is obtained according to the optimal order sequence; a local candidate state set of each picking station is generated by performing a picking operation independently by each picking station;

[0012] The local candidate state sets of each picking station are combined into a global candidate state combination through Cartesian product; each global candidate state generated is subjected to bin conflict detection, and if there is no bin conflict between picking stations, the global candidate state is legal; if there is a bin conflict, the global candidate state is excluded;

[0013] The legal global candidate states are scored by using a comprehensive evaluation function, and the scores are sorted from low to high, and the first K global candidate states with the lowest scores are retained;

[0014] The retained K global candidate states are subjected to state expansion and pruning operations in a loop until all order picking tasks are completed according to the optimal order sequence of each picking station or the maximum number of iterations is reached, and a final global candidate state set is obtained;

[0015] The state with the lowest score in the final global candidate state set is selected as the global optimal state according to the comprehensive evaluation function, and a bin sorting sequence corresponding to the global optimal state is output.

[0016] Preferably,

[0017] The step of assigning a seed order to each picking station according to the similarity coefficient between any two orders comprises:

[0018] The similarity coefficient between any two orders is calculated according to the number of repeated ordered goods species between any two orders, and a similarity data set is obtained;

[0019] The two orders with the largest similarity coefficient in the similarity data set are selected as the seed orders of a picking station, and all similarity coefficients related to the selected two orders are deleted in the similarity data set;

[0020] In the remaining similarity coefficients, the two orders with the largest similarity coefficient are selected as the seed orders of the next picking station, and all similarity coefficients related to the selected two orders are deleted; this step is repeated until a seed order is assigned to each picking station.

[0021] Preferably, it further comprises:

[0022] The difference between the number of orders assigned to adjacent two picking stations is set to be no more than a preset threshold;

[0023] When the order quantity allocated to any picking station just meets the difference of the order quantity of the picking station adjacent to it, and the difference is not more than a preset threshold value; for the order to be allocated to the picking station, the order is allocated to the picking station with the second largest coupling factor according to the coupling factor sorting, and if the picking station with the second largest coupling factor also meets the order quantity, the order is allocated to the picking station with the third largest coupling factor; so that the order quantity under each picking station is relatively balanced.

[0024] Preferably,

[0025] The local optimization of each candidate order sequence in the initial candidate set is performed by using neighborhood search to obtain a preliminary local optimal order sequence, and the preliminary local optimal order sequence is obtained by using neighborhood search.

[0026] For each candidate order sequence in the initial candidate set, all adjacent order pairs in the sequence are enumerated in turn;

[0027] According to the order of adjacent order pairs, the order of adjacent order pairs is exchanged in turn, and after each order of adjacent order pairs is exchanged, the minimum bin replacement number of the new sequence after the order of the exchanged order pairs is calculated by using the bin replacement number determination algorithm based on bundle search;

[0028] If the minimum bin replacement number of the new sequence after the exchange is less than the minimum bin replacement number of the sequence before the exchange, the new sequence is accepted;

[0029] The updated sequence is continuously exchanged for adjacent orders until the order of all adjacent orders cannot be reduced to the minimum bin replacement number, and a preliminary local optimal order sequence is obtained.

[0030] Preferably,

[0031] The non-adjacent order exchange and local segment flip operation are performed on the preliminary local optimal order sequence in turn to obtain a final candidate set, and the final candidate set is obtained by performing the non-adjacent order exchange and local segment flip operation on the preliminary local optimal order sequence in turn.

[0032] The order of any two non-adjacent order pairs in the preliminary local optimal order sequence is exchanged, and the minimum bin replacement number of the sequence after the order exchange is obtained; if the minimum bin replacement number of the sequence after the exchange is less than the minimum bin replacement number of the sequence before the exchange, the exchange is accepted; this process is repeated until the order exchange of any two non-adjacent order pairs cannot reduce the minimum bin replacement number;

[0033] For the order sequence after the non-adjacent order exchange, any continuous subsequence is selected, and reverse order operation is performed on all orders in the subsequence to obtain the minimum bin replacement number of the sequence after the reverse order operation. If the minimum bin replacement number of the sequence after the reverse order operation is less than the minimum bin replacement number of the sequence before the reverse order operation, the exchange is accepted. The process is repeated until the reverse order operation of any continuous subsequence cannot reduce the minimum bin replacement number, and a final candidate set is obtained.

[0034] Preferably, further comprising:

[0035] In the process of performing local optimization on the order sequence by neighborhood search, or performing non-adjacent order exchange operation on the order sequence, or performing local segment flip operation on the order sequence, a probability of acceptance is set, and the probability of acceptance is gradually reduced in steps with the number of iterations;

[0036] When the minimum bin replacement number of the updated order sequence is greater than or equal to the minimum bin replacement number of the order sequence before updating, the updated order sequence is accepted with the set probability of acceptance, and the local optimization is exited.

[0037] Preferably,

[0038] The minimum bin replacement number calculated by the bin replacement number determination algorithm based on the cluster search includes:

[0039] The total number of orders allocated to any picking station is obtained, each order is allocated to an order position of the picking station, and the order sequence of each order position is obtained; c

[0040] The initial state set of the picking station is set, and the initial state of the picking station includes: the to-be-picked list of the current order of each order position of the picking station, the current bin, the index of the next candidate order, the cumulative bin replacement number initialized to 0, and the record of the bin arrival sequence; SKU

[0041] It is judged whether each state in the initial state set of the picking station satisfies a preset completion state, and the preset completion state includes: all order positions are picked and the candidate order is exhausted; SKU

[0042] If any state satisfies the preset completion state, the state is saved to the completion state list; otherwise, the state is expanded, and a candidate state set of the next stage is obtained by enumerating all possible picking operations. The state change of the state in the candidate state set compared with the state before expansion includes:

[0043] ​​​The order line value of the order position corresponding to the picked order line is reduced by 1, and when the value is reduced to 0, it indicates that the current order on the order position has been completely picked, and a new order needs to be supplemented from the backup order queue, the order line value corresponding to the new order is updated to the line number of the new order, and the order number is increased by 1; according to whether the switching of the picking operation occurs SKU , so as to determine whether the magazine needs to be replaced, if the magazine used in the new state is different from the last state, it indicates that the magazine replacement occurs, and the cumulative magazine replacement number is increased by 1;

[0044] The states in the candidate state set are sorted from low to high according to the score calculated by the comprehensive evaluation function, and for the states with the same score, the states are sorted according to the generation order; the comprehensive evaluation function includes: obtaining the comprehensive evaluation function score by the cumulative magazine replacement number, the remaining picking workload, the remaining amount of backup orders and the preset weight index in the state;

[0045] The states with the top K comprehensive evaluation function scores are selected as the state set of the next stage;

[0046] If there is a state in the new state set that meets the completion condition, the state is saved to the completion state list, and the expansion is ended; if there is no state in the new state set that meets the completion condition,

[0047] the expansion of each state in the new state set is continued, and the states with the top K comprehensive evaluation function scores are selected as the state set of the next stage; it is judged again whether there is a state in the new state set that meets the completion condition; this step is repeated until there is a state in the new state set after the expansion that meets the completion condition, or the preset iteration number is reached, or all states in the new state set after the expansion cannot be expanded, and the expansion is ended;

[0048] If there is a completion state, the state with the lowest comprehensive evaluation function score is selected from the completion state as the final optimal state;

[0049] If there is no completion state, the state with the lowest comprehensive evaluation function score is selected from the state set obtained in the last iteration as the final optimal state;

[0050] The order sequence corresponding to the final optimal state is the optimal order sequence, and the cumulative magazine replacement number corresponding to the final optimal state is the minimum magazine replacement number of the picking station.

[0051] Preferably,

[0052] The state expansion and pruning operation performed on the reserved K global candidate states in cycles includes:

[0053] For the reserved K Each global candidate state performs a picking operation once according to the optimal order sequence of each picking station, resulting in an expanded set of global candidate states;

[0054] For each global candidate state in the expanded global candidate state set, a bin conflict detection is performed. If there is no bin conflict between picking stations, the global candidate state is valid; if there is a bin conflict, the global candidate state is excluded.

[0055] A comprehensive evaluation function is used to score the expanded legal global candidate states, and the scores are sorted from low to high, retaining the lowest-scoring states. K One global candidate state;

[0056] Repeat the above steps until all order picking tasks are completed or the maximum number of iterations is reached, to obtain the final global candidate state set.

[0057] According to a second aspect of the present invention, an order processing decision-making apparatus based on the pharmaceutical distribution industry is provided, the apparatus comprising:

[0058] Seed order allocation module: used to allocate a seed order to each picking station based on the similarity coefficient between any two orders for all orders in a batch to be processed;

[0059] The All Orders Assignment Module: For the remaining orders that were not assigned as seed orders, it calculates the quantity of the same type of goods ordered by each order and the seed orders already assigned to each picking station as a coupling factor, and assigns each remaining order to the picking station with the largest coupling factor, thus completing the order assignment.

[0060] Preliminary order sequence acquisition module: It is used to generate multiple candidate orders for each picking station using a random method, sort them to form an initial candidate set, and perform local optimization on each candidate order sequence in the initial candidate set using neighborhood search to obtain a preliminary local optimal order sequence;

[0061] Final order sequence acquisition module: used to sequentially perform non-adjacent order exchange and local segment flipping operations on the preliminary local optimal order sequence to obtain a final candidate set, and select the order sequence with the fewest bin replacements in the final candidate set as the optimal order sequence for the picking station;

[0062] Local candidate state acquisition module: used to obtain the initial state set of each picking station according to the optimal order sequence within each picking station; each picking station independently performs a picking operation to generate the local candidate state set of each picking station;

[0063] The global candidate state acquisition module is configured to form a global candidate state combination through a Cartesian product of the local candidate state sets of the respective picking stations, and to perform bin conflict detection on each generated global candidate state, wherein if there is no bin conflict between the picking stations, the global candidate state is legal, and if there is a bin conflict, the global candidate state is excluded;

[0064] The global state screening module is configured to score the legal global candidate states by using a comprehensive evaluation function, to sort the scores from low to high, and to retain the top K global candidate states with the lowest scores; K

[0065] The cycle expansion module is configured to cyclically perform state expansion and pruning operations on the retained K global candidate states until all order picking tasks are completed according to the optimal order sequence of each picking station or the maximum iteration number is reached, and to obtain a final global candidate state set; K

[0066] The global optimal state acquisition module is configured to select a state with the lowest score in the comprehensive evaluation function from the final global candidate state set as a global optimal state, and to output a bin sorting sequence corresponding to the global optimal state.

[0067] According to a third aspect of an embodiment of the present application, a storage medium is provided, which stores a computer program, and when the computer program is executed by a host, each step in the above method is implemented.

[0068] The technical solution provided by the embodiments of the present application can have the following beneficial effects:

[0069] The present application assigns orders with a high coupling degree to the same picking station by calculating the similarity coefficient between orders, enhances the possibility of bin reuse, and lays a foundation for subsequent optimization. For the internal orders of each picking station, a single-picking-station order sorting algorithm based on neighborhood search is used to minimize the number of bin replacements, and neighborhood operations such as local exchange and segment flipping are used to optimize the order sequence. A bin replacement number determination algorithm based on bundle search is used to accurately evaluate the cost of the sorting scheme, and the quality and stability of the order sorting are improved. On the basis of the order allocation and the internal order sorting of each picking station, the picking requirements of multiple picking stations are combined to perform state expansion, the global candidate state set is constructed, the Cartesian product combination and conflict detection are performed, the bin conflict states are effectively excluded, and the global bin sorting is performed to realize efficient picking in parallel and without coupling.

[0070] It should be understood that the foregoing general description and the following detailed description are only exemplary and explanatory, and cannot limit the present application. BRIEF DESCRIPTION OF DRAWINGS

[0071] ​​The accompanying drawings, which are incorporated herein and constitute part of this specification, illustrate embodiments consistent with the application and, together with the description, further serve to explain the principles of the application.

[0072] Figure 1 is a flowchart of a method for order processing decision based on the pharmaceutical distribution industry according to an exemplary embodiment;

[0073] Figure 2 is a flowchart of a multi-picking station bin sorting decoupling algorithm according to another exemplary embodiment;

[0074] Figure 3 is a picking station intra-order similarity comparison diagram according to another exemplary embodiment;

[0075] Figure 4 is a picking station inter-order similarity comparison diagram according to another exemplary embodiment;

[0076] Figure 5 is a cumulative bin replacement frequency comparison diagram according to another exemplary embodiment;

[0077] Figure 6 is a cumulative bin replacement frequency comparison diagram of a single-picking station order sorting algorithm based on neighborhood search and a random order sorting strategy under different order schemes according to another exemplary embodiment;

[0078] Figure 7 is a verification solution result diagram according to another exemplary embodiment;

[0079] Figure 8 is a system diagram of an order processing decision device based on the pharmaceutical distribution industry according to another exemplary embodiment;

[0080] In the drawings: 1-seed order allocation module, 2-all order allocation module, 3-initial order sequence acquisition module, 4-final order sequence acquisition module, 5-local candidate state acquisition module, 6-global candidate state acquisition module, 7-global state screening module, 8-cyclic expansion module, 9-global optimal state acquisition module. DETAILED DESCRIPTION

[0081] The exemplary embodiments will be described in detail herein with reference to the attached drawings. In the following description, like reference numerals refer to like elements throughout the description. The following exemplary embodiments are not representative of all embodiments consistent with the present application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of the present application as detailed in the appended claims.

[0082] Embodiment one

[0083] Figure 1 is a flowchart of an order processing decision method based on the pharmaceutical distribution industry according to an exemplary embodiment, as shown in Figure 1 , the method comprises:

[0084] S1, for all orders in a batch to be processed, a seed order is assigned to each picking station according to the similarity coefficient between any two orders;

[0085] S2, for the remaining orders not assigned as seed orders, the number of orders of the same goods category as the seed order assigned to each picking station is calculated for each order as a coupling factor, and each remaining order is assigned to the picking station with the maximum coupling factor, completing the order assignment;

[0086] S3, for multiple orders assigned to each picking station, a random method is used to generate multiple candidate order sequences to form an initial candidate set, and a neighborhood search is used for local optimization for each candidate order sequence in the initial candidate set to obtain a preliminary local optimal order sequence;

[0087] S4, the preliminary local optimal order sequence is sequentially subjected to non-adjacent order exchange and local segment flip operation to obtain a final candidate set, and the order sequence with the least number of bin replacements in the final candidate set is selected as the optimal order sequence of the picking station;

[0088] S5, in each picking station, the initial state set of each picking station is obtained according to the optimal order sequence; a picking operation is independently performed by each picking station to generate a local candidate state set of each picking station;

[0089] S6, the local candidate state sets of each picking station are combined through Cartesian product to form a global candidate state combination; each global candidate state generated is subjected to bin conflict detection, if there is no bin conflict between picking stations, the global candidate state is legal; if there is a bin conflict, the global candidate state is excluded;

[0090] S7, a comprehensive evaluation function is used to score the legal global candidate states, and the scores are sorted from low to high, and the first K global candidate states with the lowest scores are retained;

[0091] S8, the K global candidate states retained are subjected to state expansion and pruning operations in a loop until the optimal order sequence of each picking station completes all order picking tasks or reaches the maximum number of iterations, obtaining a final global candidate state set;

[0092] S9, selecting a state with the lowest score of the comprehensive evaluation function in the final global candidate state set as a global optimal state, and outputting a bin sorting sequence corresponding to the global optimal state;

[0093] It can be understood that in the medical and pharmaceutical circulation scene, a typical goods-to-person picking system usually includes multiple parallel picking stations. In order to improve the overall efficiency of the system and alleviate the resource competition problem between the picking stations, it is necessary to consider the "order allocation strategy", "order sorting in the picking station", and "bin sorting and conflict decoupling between picking stations" three key links in the order execution process. Based on this, the embodiment constructs a multi-picking station order execution decision algorithm framework, and the core goal is to maximize the bin reuse rate, minimize the bin replacement times, and eliminate the bin conflicts between multiple picking stations.

[0094] Mainly includes the following three stages:

[0095] Order similar allocation strategy: first, according to the order structure information, the initial allocation is performed. The traditional dynamic allocation strategy allocates orders to each picking station in turn according to the order receiving time, but this method may lead to highly similar orders being scattered, increasing the fragmentation of SKU distribution. Therefore, the order similar allocation strategy is proposed, which calculates the SKU overlap degree between orders, allocates orders with high coupling degree to the same picking station, enhances the possibility of bin reuse, and lays the foundation for subsequent optimization.

[0096] Single picking station order sorting algorithm based on neighborhood search: for the internal orders of each picking station, after allocation, the order processing sequence is further optimized. The single picking station order sorting algorithm based on neighborhood search is designed to minimize the bin replacement times. Through local exchange, segment flipping and other neighborhood operations, the order sequence is optimized, and the bin replacement times determination algorithm based on bundle search is used to accurately evaluate the cost of the sorting scheme, improving the quality and stability of the order sorting.

[0097] Multi-picking station bin sorting decoupling algorithm: based on the allocation of orders and the completion of internal order sorting of each picking station, the multi-picking station bin sorting decoupling algorithm is further proposed. The algorithm takes global synchronization and conflict-free execution logic as the core, combines the picking demands of multiple picking stations for state expansion, effectively eliminates bin conflict states through global bin sorting, and realizes parallel and decoupled efficient picking.

[0098] Among them,

[0099] I. The order similar allocation strategy specifically includes:

[0100] Dynamic order allocation is a relatively traditional allocation method, where orders are sequentially distributed to each picking station according to their receipt time. Combining the results of order batching, the specific steps of the existing dynamic allocation strategy are as follows:

[0101] Orders within a batch are allocated to different picking stations based on their receiving time, with the number of picking stations being [number missing]. n The total number of orders is q When the picking station number is less than or equal to At that time, allocate to the picking station For each order, when the picking station number is greater than [number]... At that time, allocate to the picking station For each order, each picking station picks orders sequentially according to their assigned order. If any order is completed during the picking process, the next order is added to continue picking until all orders in that batch have been picked.

[0102] However, in reality, there is a certain correlation between the internal structure of orders. If highly correlated orders are assigned to the same picking station, a single raw material bin can meet the needs of multiple orders. Therefore, this embodiment proposes an order similarity allocation strategy, which allocates orders based on their correlation coefficient, ensuring that highly correlated orders are assigned to the same picking station. This, in turn, improves overall picking efficiency by increasing the re-picking rate of individual raw material bins. The specific allocation steps are as follows:

[0103] (1) Assume that the number of orders that a single picking station can process at the same time is The total number of orders is The number of picking stations is Calculate the similarity coefficient between any two orders within a batch. This refers to the number of duplicate order items between two orders, which is added to the dataset. Sort them in descending order and select... The two corresponding orders are used as a seed order for a picking station and are removed from the dataset. All corresponding correlation coefficients, then in the dataset Continue to select the one with the largest similarity coefficient As the seed order for the next picking station, repeat this process until... n Each picking station is assigned a seed order (each seed order corresponds to two similar orders);

[0104] (2) After each picking station has determined its seed order, it is assigned the remaining other orders. The remaining orders that were not assigned as seed orders form an order pool. For each order, calculate the quantity of the same type of goods ordered as the seed order already assigned to the picking station as a coupling factor. For each order Allocate it to The corresponding picking station will repeat this step cyclically until all orders have been assigned to a picking station. The maximum number of orders that can be assigned to each picking station is as follows:

[0105] When the picking station number is less than or equal to At that time, allocate to the picking station For each order, when the picking station number is greater than [number]... At that time, allocate to the picking station One order; simply put, the number of orders between two adjacent picking stations cannot exceed 1. If the number of orders at a certain picking station has already met the quantity requirement, then for the next order assigned to that picking station, it will be assigned to the picking station with the second highest coupling factor. If the number of orders at the picking station with the second highest coupling factor has also met the requirement, then the order will be postponed to the picking station with the third highest coupling factor, and so on.

[0106] II. Order sorting algorithm for a single picking station based on neighborhood search:

[0107] In a "goods-to-person" pharmaceutical picking system, although orders have been allocated to various picking stations according to certain rules, each order involves... Differences in type and quantity mean that if the order processing sequence is not arranged properly during the actual picking process, it may lead to frequent replacement of material bins. Each replacement of material bins is accompanied by equipment switching operations and may also cause problems such as conveyor line blockage and picking waiting, thereby significantly reducing the system operating efficiency.

[0108] Therefore, after completing order allocation, this embodiment further focuses on the order sequencing problem within a single picking station. The core objective is to achieve this by rationally sequencing the order execution order. Minimize the number of bin replacements, improve the continuous operation capability of the picking station and the system throughput efficiency. Specifically, reasonable order sequencing can reduce the number of identical items in multiple orders. Centralized picking maximizes the reuse rate and continuous use time of bins without affecting order accuracy.

[0109] Based on the above motivations, this embodiment designs an order ranking optimization algorithm based on neighborhood search. Starting with an initial order sequence, it progressively evaluates the local perturbations and global search, and uses feedback from the number of bin replacements to filter the ranking scheme. Specifically, it includes:

[0110] (1) Generate initial candidate solutions

[0111] Randomly generated for orders within the current batch Sort the candidates to form an initial candidate set:

[0112] For the current batch of orders of the picking station, first generate candidate order sequences using random method, and denote any candidate sequence as:

[0113] (1)

[0114] To ensure the diversity of the solution space, generate candidate order sequences randomly, and form the initial candidate set:

[0115] (2)

[0116] (2) Preliminary optimization of adjacent order exchange

[0117] For each candidate solution, exchange adjacent orders pair by pair and calculate the number of bin replacements (evaluate by bin replacement number determination algorithm based on cluster search), if the exchange reduces the number of bin replacements, accept the new sequence and continue optimization, until adjacent exchange cannot further improve:

[0118] To reduce the number of bin replacements, for each candidate sequence, use neighborhood search for local optimization, for any candidate sequence , enumerate all adjacent order pairs in the sequence in turn, exchange positions one by one to generate new sequence , for the sequence obtained after exchange , calculate its minimum bin replacement number using bin replacement number determination algorithm based on cluster search , if:

[0119] (3)

[0120] accept the new sequence , update the current candidate sequence, let:

[0121] (4)

[0122] and continue to exchange adjacent orders for the updated sequence, until all adjacent exchanges cannot reduce , at this time, the preliminary local optimal sequence is obtained, denoted as ;

[0123] (3) Neighborhood expansion optimization

[0124] ​Based on the local optimal solution obtained by preliminary optimization, two kinds of expansion operations are further explored. The first kind of expansion is non-adjacent order exchange, and any two non-adjacent orders in the sequence are selected to exchange positions, and the number of magazine replacements is re-evaluated. If it is improved, it is accepted. The second kind of expansion is local segment flip, and the selected continuous sub-sequence in the sequence is processed in reverse order, and the number of magazine replacements is re-evaluated. If a better result is obtained, it is accepted. Specifically, it includes:

[0125] 1. Non-adjacent order exchange

[0126] For any order pair in that satisfies , perform position exchange to generate a new sequence , calculate for each exchanged sequence, and determine whether to accept it.

[0127] 2. Local segment flip

[0128] Select any continuous sub-sequence in the sequence from position to (where ), perform reverse operation on all orders in the segment to form a new sequence , enumerate all possible combinations of and , and for each flip result, calculate and determine whether to accept it.

[0129] (4) Acceptance criteria and escape mechanism

[0130] In the iteration process, in order to avoid falling into local optimum, set a mechanism to accept slightly worse solutions with a certain probability, and gradually reduce the acceptance probability as the number of iterations increases, so that the algorithm can escape from local traps and also converge to better solutions in the later stage. Specifically, it includes:

[0131] To avoid the expansion operation being limited by the local optimum, set an acceptance probability (the probability gradually decreases with the number of iterations), for a certain exchange operation:

[0132] If , still accept the update with a probability of to escape from the local optimum, otherwise strictly accept the operation with a reduced value.

[0133] (5) Selection and output of optimal sequence

[0134] ​​After multiple iterations and expansion operations, the sequence with the minimum number of bin replacements from the final retained sequence is selected as the output, i.e., the optimal arrangement scheme of the single picking station order, which specifically includes:

[0135] After completing the expansion neighborhood operation, the updated candidate sequence is denoted as ;

[0136] After neighborhood search, the initial candidate set is updated to:

[0137] (5)

[0138] Each undergoes local improvement of order arrangement, so that under a given decision, the number of bin replacements is reduced to a lower level, and the order sequence with the minimum number of bin replacements is selected from the set candidate set as the final order sorting scheme, i.e.:

[0139] (6)

[0140] The sorting scheme is the optimal order sorting scheme for the single picking station, and the corresponding number of bin replacements is the theoretical minimum value.

[0141] It should be emphasized that the above-mentioned bin replacement number determination algorithm based on cluster search calculates the minimum bin replacement number, which specifically includes:

[0142] In order to accurately evaluate the bin replacement number under any order sorting, it is necessary to consider multiple factors such as "order position number" and "arrival of candidate orders". Therefore, the bin replacement number determination algorithm based on cluster search is adopted in this embodiment. In the multi-stage picking process simulation, the approximate optimal bin arrival order and the corresponding minimum replacement number are efficiently solved through dynamic state transition and cluster pruning.

[0143] Main idea:

[0144] The picking process on a single picking station is divided into several consecutive time stages (each stage corresponds to the picking of one order line). Initially, each order position is assigned an order in sequence. When the orders on a certain order position are all picked, the next order is supplemented from the candidate order queue.

[0145] In order to avoid exponential state explosion at each picking stage, the algorithm uses a comprehensive evaluation function to evaluate the state at each stage, and retains the top states (controlled by the width of cluster search). When the retained states are expanded to the next stage, pruning is performed. The comprehensive evaluation function considers the cumulative bin replacement number The remaining order quantity and the remaining workload to ensure that the search direction is mainly to reduce the core goal of the replacement of the box.

[0146] State description and initial definition:

[0147] A batch of orders are allocated to a picking station for picking, and the picking process is divided into multiple consecutive time stages (denoted as ), each stage corresponds to a picking time period of an order line, and a picking station is set to have order positions, each order position is bound to at most one order to be picked at any time, it should be noted that an order corresponds to multiple order lines, each order line corresponds to a SKU , and each SKU corresponds to a box;

[0148] In the initial stage, the first orders are allocated to each order position in turn to form the initial state of the picking station, and the state vector is used to represent the state of stage , which is composed of four parts and is denoted by symbol:

[0149] (7)

[0150] Among them, the set represents the number of order lines in the current order that have not been picked on each order position, and if there is no order allocated to a certain order position, it is defined as ; represents the number of the box used by the current picking station, which is used to judge whether the box replacement occurs in the picking process; is the number of box replacements accumulated from the initial state to the current state; represents the position index of the candidate order queue to be filled into the next order position, and the candidate order sequence is denoted as:

[0151] (8)

[0152] Therefore, the initial stage (i.e. ) can be written as:

[0153] (9)

[0154] Among them, represents the total number of order lines in the ;

[0155] State extension mechanism:

[0156] Only one order line picking operation is completed in each stage, and the state changes after the picking operation is completed, mainly in two aspects:

[0157] Order state update: for the order position where the order line is picked, its value is reduced by 1; when the value is reduced to 0, it means that the order on the order position has been completed, at this time a new order needs to be supplemented from the candidate order queue, and the corresponding to the number of rows of the new order is updated, and is increased by 1;

[0158] Bin replacement judgment: according to whether the switching of occurs in this picking operation, it is judged whether the bin needs to be replaced, if the bin used in the new state is different from the last state, it means that the bin replacement has occurred, .

[0159] The update of the two aspects is abstractly described by the state transition formula:

[0160] (10)

[0161] Among them,

[0162]

[0163] Each time a picking operation is completed, it goes from one stage to the next stage, and each stage corresponds to a picking operation of an order line and an update of the state. After each stage, a set of candidate states is generated;

[0164] The entire state transition process can be summarized into three situations according to the different conversion of the picking task, and Table 1 below describes the three situations, showing the state transition process and the influence of each operation on the state update:

[0165] Table 1 State transition description

[0166]

[0167] State pruning strategy of bundle search:

[0168] Although the picking operation in a single stage only corresponds to the processing of an order line on the current order position, due to the existence of multiple decision branches, let the state set reserved in the current stage be , then each state , through enumerating all possible picking operations, the candidate state set of the next stage is obtained:

[0169] (11)

[0170] Among them Indicates the state The process of state transition generates a set of all possible successor states. Since each state may generate multiple different state branches in a single picking operation, the number of stages increases accordingly. As the number of candidate states increases, the number of candidate states often grows exponentially. To suppress the computational burden caused by this exponential growth, a cluster search strategy is used for pruning to accelerate the efficiency of the algorithm.

[0171] After state expansion at each stage, the candidate states are ranked using a comprehensive evaluation function, and only the top-ranked states are retained. The state moves to the next stage, thus limiting the maximum number of states retained in each stage;

[0172] The comprehensive evaluation function is constructed as follows:

[0173] (12)

[0174] in This represents the cumulative number of times the hopper has been replaced. Indicates the remaining picking workload; () indicates the remaining quantity of waitlist orders; The weights are positive. The goal is to minimize... Although the evaluation function considers multiple factors, the objective of this embodiment is to minimize the cumulative number of bin replacements. Therefore, in parameter settings, a higher value is adopted. Weights to ensure the search primarily follows the decreasing path Proceed in the direction of;

[0175] In the candidate state set In, for each state calculate Then, sort the results from lowest to highest score. For cases with the same score, sort them according to the order in which they were generated. Finally, select the top-ranked results from the sorted list. These states constitute the set of states for the next stage. ;

[0176] In summary, the specific process of the algorithm for determining the number of bin replacements based on cluster search is as follows:

[0177] Step 1: Initialization

[0178] (1) Calculate the total number of orders and assign initial orders to each order position (select the first order in sequence). (one order);

[0179] (2) Set the initial state, including: the current order positions to be picked. List; Current bin; Index of the next waiting order; Cumulative bin replacement count Initialize to 0; Record the bin arrival sequence.

[0180] Step 2: State expansion and candidate state generation

[0181] For each state in the current state set, do the following:

[0182] (1) If the state has met the completion condition (all order positions have been picked and all candidate orders have been exhausted), save the state to the completed state list;

[0183] (2) Otherwise, expand the state to generate all possible successor states (each state corresponds to a picking operation), and collect the candidate state set.

[0184] Step 3: Pruning and updating the state set

[0185] (1) Sort the candidate state set according to the comprehensive evaluation function (considering the cumulative replacement number, remaining candidate order number, and remaining picking workload);

[0186] (2) Select the top states as the next stage state set;

[0187] (3) If there is a state in the new state set that meets the completion condition, save it to the completed state list and end the expansion early;

[0188] (4) Otherwise, repeat "Step 2" and "Step 3" until the maximum number of iterations is reached or all states cannot be expanded.

[0189] Step 4: Result selection

[0190] If there is a completed state, i.e., when all order positions in the list are empty, and all candidate orders have been supplemented and picked, in this state, there is no picking operation that can be executed because all tasks have been completed, then select the state with the lowest comprehensive evaluation score from it as the final optimal state; if there is no completed state, select the state with the lowest score from the current state set (if the bin sequence is unreasonable, it may lead to the fact that a certain order sequence cannot be completed, such as the current order row requiring a bin that has already been used in front).

[0191] Step 5: Return the final state

[0192] Return the best state, which contains the best bin arrival sequence and the minimum cumulative bin replacement number B .

[0193] Three, multi-picking station bin sorting decoupling algorithm: ​

[0194] In a real-world multi-picking station parallel operation environment, if different picking stations simultaneously require the same bin, coupling conflicts are likely to occur, severely impacting picking efficiency. To effectively address this issue, this embodiment further proposes a multi-picking station bin sorting decoupling algorithm. This algorithm considers the needs of all picking stations simultaneously at each picking stage. By constructing and expanding the global state space and leveraging the pruning mechanism of cluster search, a conflict-free and efficient bin sorting scheme is quickly selected to decouple multiple picking stations.

[0195] As attached Figure 2 As shown, the specific steps of the algorithm include:

[0196] (1) Global state initialization

[0197] Within each picking station, the orders to be picked are initialized according to the existing order sorting scheme. Information such as the list, current bin, waiting list orders, and bin replacement count.

[0198] (2) Parallel phased state expansion

[0199] In each picking stage, each picking station independently performs a "single picking station state transition," generating its own local candidate state set. Each state represents the update result after a bin picking operation, specifically including:

[0200] In a single picking station scenario, each stage only requires updating the picking status of a single order line. However, when extended to multiple picking stations, each stage becomes a parallel process where all picking stations proceed synchronously. That is, all picking stations must execute their picking actions simultaneously, and each picking station has its own local state at a certain stage. :

[0201] (13)

[0202] Among them, set Indicates picking station No. The number of order lines in the current order that have not yet been picked at each order position. If no order has been assigned to a certain order position, then... ; Indicates the current picking station The bin number used is used to determine whether a bin has been changed during the picking process; This represents the cumulative number of bin replacements from the initial state to the current state. This represents the position index of the next order in the queue of pending orders.

[0203] For each picking station in the system Based on the current state , independently execute a state transition process of a single sorting table logic, according to the single sorting table state transition scenario, for the sorting table Generate a set of next-stage feasible local candidate states:

[0204] (14)

[0205] Each local candidate state has the following information:

[0206] (15)

[0207] (3) Global state combination and conflict detection

[0208] The local states generated by each sorting table form a global state combination through Cartesian product. For each global state generated, perform bin conflict detection. If there is no bin conflict between sorting tables, the state is legal and proceeds to the next step of evaluation. If there is a bin conflict, the state is excluded. Specifically, it includes:

[0209] After each sorting table generates a local candidate state, perform Cartesian product operation on the candidate state set of all sorting tables to construct a complete global candidate state set. Each global candidate state generated can be represented as:

[0210] (16)

[0211] Where the global cumulative bin replacement number is the sum of the replacement number in the current candidate state and the previous cumulative replacement number of each sorting table. Specifically, it is represented as:

[0212] (17)

[0213] Where is the number of sorting tables.

[0214] In the multi-sorting table parallel operation scenario defined in this algorithm, "conflict" specifically means that when and only when the same bin is selected by different sorting tables (i.e., the same SKU is sorted) in the same stage, a conflict is considered to have occurred. Specifically, it is expressed as:

[0215] (18)

[0216] In order to avoid coupling problems, the conflict needs to be filtered:

[0217] For each global candidate state , extract the bins determined by each sorting table in this stage to form a global bin set:

[0218] (19)

[0219] Then, check the duplicated set. Number of elements and number of picking stations Are they equal?

[0220] 1. If the number of elements equals ,Right now If the condition is met, it indicates that the bins selected by each picking station are unique in this stage. This global candidate state satisfies the conflict-free constraint and can proceed to the next evaluation stage.

[0221] 2. If the number of elements is less than ,Right now If the condition is true, it indicates that at least two picking stations have selected the same bin. In this case, the global state will be immediately and completely excluded from the candidate state set and will not participate in the subsequent state evaluation and pruning process.

[0222] Ensure the status of all candidates entering the evaluation stage All of them strictly satisfy the global constraint of no conflict among multiple picking stations, and finally form a valid global candidate set. .

[0223] (4) Cluster search pruning optimization

[0224] A comprehensive evaluation function is used to score the legitimate global candidate states. The evaluation function takes into account the number of bin replacements, remaining picking quantity, and number of waiting orders. Based on the scores, only the top-scoring states are retained. Each state proceeds to the next stage to avoid searching for too many states. Specifically, this includes:

[0225] After conflict filtering, the valid global candidate states are scored and ranked according to the comprehensive evaluation function. The comprehensive evaluation function is as follows:

[0226] (20)

[0227] in, This represents the total number of times the hopper has been replaced globally. This indicates that all picking stations have not yet been completed. Total number of rows; This indicates the remaining quantity of all waiting orders at all picking stations; The weights are positive, and the objective is to minimize... ;

[0228] Then, a cluster search mechanism is used to sort the valid global candidate states, retaining only the lowest-scoring states. Each state moves to the next stage, thereby controlling the search scale and effectively improving search efficiency.

[0229] (5) Iterative progress and termination judgment

[0230] The state expansion and pruning process is executed repeatedly until all order picking tasks are completed or the maximum number of iterations is reached.

[0231] (6) Output the optimal solution

[0232] Select the state with the lowest score from all completed states, and output the corresponding bin sorting sequence and the cumulative number of bin replacements.

[0233] To verify the feasibility and effectiveness of the above-mentioned solution in this application, the following specific implementation examples are also disclosed in this embodiment:

[0234] Analysis of the Order Similarity Allocation Strategy Experiment:

[0235] To verify the effectiveness of the order similarity allocation strategy, random order data with different similarity levels (0.1 to 0.9) were generated, and the proposed similarity allocation strategy was compared with the traditional dynamic random allocation strategy. The experiment was set with 80 orders and 4 picking stations.

[0236] After running the above strategies, record the mean similarity of orders within the picking station under different similarity parameters (measuring the similarity between orders on the same picking station). The results of the overlap degree and the mean similarity between picking stations (measuring the degree of overlap between orders located on different picking stations) are attached. Figure 3 and appendix Figure 4 As shown.

[0237] Appendix Figure 3 The study shows the impact of different allocation strategies on the similarity of orders within the picking station as the overall order similarity gradually increases. When the overall similarity is low (e.g., 0.1-0.2), the two strategies do not show a significant difference in the similarity of orders within the picking station. As the similarity gradually increases, the order similarity allocation strategy gradually shows a higher mean "in-station" similarity, indicating that this strategy tends to allocate orders with higher similarity within the picking station. Orders with high overlap are placed on the same picking station, thus reducing the need for picking orders in the pharmaceutical order picking scenario. The role of dispersion.

[0238] Appendix Figure 4 The display shows the trend of order similarity between picking stations. As the overall similarity increases, the traditional dynamic allocation strategy shows a relatively larger increase in order similarity between different picking stations, indicating that under this strategy, there are orders containing the same or similar items. Orders can be more dispersedly assigned to multiple picking stations; while the order similarity assignment strategy significantly reduces the dispersion, making the overlap between different picking stations smaller, and more in line with the demand for order independence from each other under the pharmaceutical batch constraint.

[0239] Overall, the order similarity assignment strategy can achieve the effect of "similar orders on the same station, different orders across stations" at a higher similarity, which is more practical for reducing interference between different batch orders in the pharmaceutical scenario.

[0240] Experimental analysis of single picking station order sorting algorithm based on neighborhood search:

[0241] Experimental design and data explanation:

[0242] In order to verify the effectiveness and stability of the single picking station order sorting algorithm based on neighborhood search proposed in this application, numerical experiments are designed, and the specific experimental settings are as follows:

[0243] Order data generation: For a picking station, 20 orders are generated, overlapping between different orders, i.e., each order contains no repetition, but repetition between different orders is allowed to simulate the demand for the same in different orders in the real environment.

[0244] Picking station configuration: Set the number of order positions that can be processed in parallel on the picking station to 4. When picking starts, the first 4 orders are assigned to 4 order positions for processing, and after completing the picking of a certain order position, new orders are supplemented to the idle order positions from the standby orders.

[0245] Algorithm parameters: For the bin replacement frequency determination algorithm based on cluster search, different parameters are set to evaluate the balance between search accuracy and computational efficiency, taking values of {1, 10, 15}.

[0246] Set the random order sorting strategy as the comparison baseline, and to ensure the stability of the random strategy, each sorted random strategy is repeated 30 times and the average value is taken. The weights in the evaluation function are (1000, 1, 1) respectively, where tends to focus on reducing bin replacement frequency, and gives a certain weight to the unfinished orders and the remaining to-be-picked numbers, which is much smaller than .

[0247] Experimental results and discussion

[0248] Impact analysis on order sorting optimization

[0249] Run the algorithm and record the different order sorting schemes under a single order. The parameter is the minimum cumulative number of bin replacements in each picking stage during the search process. The changes, and the different strategies The cumulative number of bin replacements obtained from the cluster search is as follows: Comparison details are attached. Figure 5 As shown:

[0250] Under a single order sorting scheme, different The parameters selected in each picking stage differ, which in turn affects the final cumulative number of bin replacements. There are obvious differences. Larger strategies, by evaluating more states at each stage, can select solutions with better state quality earlier, leading to a more efficient final outcome. The value decreased significantly, while Smaller strategies, due to their limited search scope, may only achieve worse states during stage selection, leading to a higher B-value at convergence. The settings directly affect the diversity and quality of state selection at each stage, which in turn determines the degree of optimization and stability of the final result.

[0251] A comprehensive comparative analysis of single-picking station order sorting algorithms based on neighborhood search and random strategies:

[0252] Appendix Figure 6 The cumulative number of bin replacements was compared between the proposed neighborhood search-based single-picking station order sorting algorithm and the random order sorting strategy under different order schemes.

[0253] From the appendix Figure 6 It is clear from the above that, regardless of the selected Regardless of how the parameters change, the single picking station order sorting algorithm based on neighborhood search consistently performs well in reducing the cumulative number of bin replacements, and its advantages far exceed those of random order sorting strategies.

[0254] When using a larger In this case, cluster search can find the optimal path in a wider state space, thus achieving a lower cumulative bin replacement count in most order scenarios. Order sorting by value, when using a smaller value When relying entirely on random strategies, the algorithm is more prone to getting trapped in local optima due to the limited local search range, leading to... The value fluctuates significantly across different order arrangements, resulting in poor overall optimization performance.

[0255] Therefore, the single picking station order sequencing algorithm based on neighborhood search can not only stably obtain low values under different parameter settings, but also further improves the results of the final sequencing scheme by increasing the number of search states.

[0256] Stability analysis of single picking station order sequencing algorithm based on neighborhood search on random initial sequence:

[0257] To further verify the stability of the algorithm, the same data configuration (20 orders, each order with 4 order positions) was used for multiple independent experiments, each experiment randomly generated a different initial order sequence, and the proposed single picking station order sequencing algorithm based on neighborhood search was applied to solve it;

[0258] As shown in the accompanying Figure 7 figure, the optimal value line obtained in each run, with the average value and (standard deviation) error band, can be found that the values of most runs fluctuate around a relatively small interval, indicating that the algorithm has good convergence effect on different random initial solutions under the same parameter setting; a few runs have relatively high values, which usually means that the search did not successfully transition to a better solution during the early local optimal or inferior solution acceptance process, but overall this is a rare case. It can be concluded that the algorithm has relatively small dependence on the initial solution, showing certain stability and robustness.

[0259] Experimental analysis of multi-picking station bin sequencing decoupling algorithm:

[0260] Experimental design and data description

[0261] To evaluate the actual performance of the multi-picking station bin sequencing decoupling algorithm proposed in this application, 9 groups of multi-picking station parallel experiment scenarios were constructed, each scenario containing 4 independent picking stations, each picking station with 4 order positions, and the order data was generated in the same way and scale as above to ensure consistency and comparability of the experimental conditions.

[0262] The three strategies for specific comparison and analysis are as follows:

[0263] (1) Random strategy: independently select bins randomly for each picking station to obtain the bin picking order of each picking station;

[0264] (2) Single picking station cluster search strategy: perform single picking station bin replacement frequency determination algorithm based on cluster search for each picking station (without considering conflicts between different picking stations) to obtain the bin picking order of each picking station;

[0265] (3) Decoupling strategy for multi-picking station bin sorting: In each parallel stage, the local state of each picking station is expanded, and a global candidate is formed by combining them through Cartesian product. Then, combinations with "duplicate SKUs in the same stage" are filtered out, thereby ensuring that the final picking sequence is conflict-free, and setting A value of 10 is used to filter the states and count the final number of bin replacements and the degree of coupling.

[0266] To quantify the overall performance of each algorithm, the following key metrics were recorded:

[0267] (1) Coupling frequency: This indicates the number of conflicts caused by each picking station using the same SKU in the same stage;

[0268] (2) Number of bin replacements: The total number of bin replacements across the four picking stations, reflecting picking efficiency;

[0269] (3) Algorithm running time: The time (in seconds) taken for the algorithm to start and generate the final sorting scheme.

[0270] Experimental Results and Discussion:

[0271] Three strategies were applied to nine sets of test data. The experimental results showed that the random strategy cannot avoid issues in the same stage of a multi-pickup parallel scenario. Conflicts lead to a relatively high number of couplings, and due to random selection It is more difficult to control the frequency of bin replacement, resulting in a higher number of bin replacements. Although it has the shortest running time, the overall picking efficiency and conflict level are not good.

[0272] Single-station cluster search optimizes the picking order within a single picking station, significantly reducing the number of bin changes. However, it does not coordinate between picking stations, so conflicts cannot be avoided during the parallel phase. Compared to random strategies, the coupling is reduced, but some repetition may still occur. .

[0273] In the multi-picking station bin sorting decoupling strategy, because "same bins in the same stage" are filtered at each stage... SKU The candidate combination of "" can ideally reduce the number of couplings to 0 (i.e., no conflict). Experimental results also show that the number of couplings of this strategy is significantly lower than the previous two strategies, and there is no conflict. At the same time, it also has a significant advantage in reducing the number of bin replacements. However, due to the need for larger-scale state combination (Cartesian product extension) and conflict detection and evaluation, its algorithm has the longest running time among the three.

[0274] In summary, the multi-picking station bin sorting decoupling algorithm achieves the lowest coupling frequency and a better bin replacement frequency in most test scenarios, demonstrating a significant advantage in picking quality. If the actual application environment has high requirements for conflict reduction and bin scheduling quality, the algorithm is even more competitive.

[0275] Example 2

[0276] Figure 8 This is a schematic diagram of an order processing decision-making device for the pharmaceutical distribution industry, illustrated according to another exemplary embodiment. The device includes:

[0277] Seed order allocation module 1: Used to allocate a seed order to each picking station based on the similarity coefficient between any two orders for all orders in a batch to be processed;

[0278] All Orders Assignment Module 2: For the remaining orders that were not assigned as seed orders, this module calculates the quantity of the same type of goods ordered by each order and the seed orders already assigned to each picking station as a coupling factor, and assigns each remaining order to the picking station with the largest coupling factor, thus completing the order assignment.

[0279] Preliminary order sequence acquisition module 3: It is used to generate multiple candidate orders for each picking station using a random method, sort them to form an initial candidate set, and perform local optimization on each candidate order sequence in the initial candidate set using neighborhood search to obtain a preliminary local optimal order sequence;

[0280] Final order sequence acquisition module 4: is used to sequentially perform non-adjacent order exchange and local segment flipping operations on the preliminary local optimal order sequence to obtain a final candidate set, and select the order sequence with the fewest bin replacements in the final candidate set as the optimal order sequence for the picking station;

[0281] Local candidate state acquisition module 5: used to obtain the initial state set of each picking station according to the optimal order sequence within each picking station; and to generate the local candidate state set of each picking station by performing a picking operation independently.

[0282] Global Candidate State Acquisition Module 6: This module is used to combine the local candidate state sets of each picking station into a global candidate state combination through Cartesian product; it performs bin conflict detection on each generated global candidate state. If there is no bin conflict between picking stations, the global candidate state is valid; if there is a bin conflict, the global candidate state is excluded.

[0283] Global State Filtering Module 7: Used to score the valid global candidate states using a comprehensive evaluation function, sort the scores from low to high, and retain the lowest-scoring states.K a global candidate state;

[0284] a cycle extension module 8 configured to perform state extension and pruning operations on the reserved K a cycle extension module 8 configured to perform state extension and pruning operations on the reserved

[0285] a global optimal state acquisition module 9 configured to select a state with the lowest score of the comprehensive evaluation function from the final global candidate state set as the global optimal state, and output a bin sorting sequence corresponding to the global optimal state.

[0286] Embodiment Three

[0287] The storage medium stores a computer program, and the computer program is executed by a host to implement each step in the above method.

[0288] It can be understood that the storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0289] It can be understood that the same or similar parts in the above embodiments can be mutually referred to, and the contents not described in detail in some embodiments can be referred to the same or similar contents in other embodiments.

[0290] It should be noted that in the description of the present application, the terms "first", "second", etc. are only for the purpose of description, and cannot be understood as indicating or implying relative importance. In addition, in the description of the present application, unless otherwise specified, the meaning of "a plurality of" is at least two.

[0291] Any process or method descriptions in flow charts or otherwise described herein can be understood as representing code modules, segments, or portions of code that include one or more executable instructions for performing specific logic functions or steps in the process, and the various embodiments of the present application include additional implementations in which the order of execution is not necessarily the same as the order shown or discussed, including the implementation of functions according to the functionality involved, in substantially simultaneous manner or in reverse order, which should be understood by those skilled in the art of the embodiments of the present application.

[0292] It should be understood that each part of the present application can be realized by hardware, software, firmware or a combination thereof. In the above-mentioned embodiments, a plurality of steps or methods can be realized by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if realized by hardware, and as in another embodiment, any one or a combination of the following technologies known in the art can be used: discrete logic circuit with logic gate circuit for implementing logic functions on data signals, application specific integrated circuit with suitable combination logic gate circuit, programmable gate array (PGA), field programmable gate array (FPGA), etc.

[0293] Those skilled in the art of the present technology can understand that all or part of the steps carried out by the above-mentioned embodiment methods can be completed by programs instructing related hardware, and the programs can be stored in a computer readable storage medium, and when executed, include one or a combination of steps of the method embodiments.

[0294] In addition, each functional unit in each embodiment of the present application can be integrated into one processing module, or each unit can exist physically alone, or two or more units can be integrated into one module. The above-mentioned integrated module can be realized in the form of hardware or in the form of a software functional module. The integrated module, if realized in the form of a software functional module and sold or used as an independent product, can also be stored in a computer readable storage medium.

[0295] The above-mentioned storage medium can be a read-only memory, a magnetic disk or an optical disk, etc.

[0296] In the description of the present specification, the description of the terms "one embodiment", "some embodiments", "an example", "a specific example" or "some examples" means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In the present specification, the illustrative description of the above-mentioned terms does not necessarily mean the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0297] Although the embodiments of the present application have been shown and described above, it should be understood that the above-mentioned embodiments are exemplary and cannot be understood as limiting the present application, and those skilled in the art can make changes, modifications, replacements and variations to the above-mentioned embodiments within the scope of the present application.

Claims

1. An order processing decision-making method based on the pharmaceutical distribution industry, characterized in that, The method includes: For all orders within a batch to be processed, assign a seed order to each picking station based on the similarity coefficient between any two orders; For the remaining orders that were not assigned as seed orders, calculate the quantity of the same type of goods ordered by each order and the seed orders already assigned to each picking station as a coupling factor, and assign each remaining order to the picking station with the largest coupling factor to complete the order allocation; For each picking station, multiple orders are assigned to a random number of candidate orders to form an initial candidate set. Then, a neighborhood search is used to perform local optimization on each candidate order sequence in the initial candidate set to obtain a preliminary locally optimal order sequence. The preliminary locally optimal order sequence is subjected to non-adjacent order swapping and local segment flipping operations in sequence to obtain the final candidate set. The order sequence with the fewest bin replacements in the final candidate set is selected as the optimal order sequence for the picking station. Within each picking station, an initial state set for each picking station is obtained based on the optimal order sequence; each picking station performs a picking operation independently, generating a local candidate state set for each picking station. The local candidate states of each picking station are combined into a global candidate state combination through Cartesian product. For each generated global candidate state, a bin conflict detection is performed. If there is no bin conflict between picking stations, the global candidate state is valid; if there is a bin conflict, the global candidate state is excluded. A comprehensive evaluation function is used to score the valid global candidate states, and the scores are sorted from low to high, retaining the lowest-scoring states. K One global candidate state; For the reserved K The global candidate state is repeatedly expanded and pruned until all order picking tasks are completed according to the optimal order sequence of each picking station or the maximum number of iterations is reached, and the final global candidate state set is obtained. The state with the lowest comprehensive evaluation function score in the final global candidate state set is selected as the global optimal state, and the bin sorting sequence corresponding to the global optimal state is output.

2. The method according to claim 1, characterized in that, The step of assigning a seed order to each picking station based on the similarity coefficient between any two orders includes: Calculate the similarity coefficient between any two orders based on the number of duplicate order types between them, and obtain the similarity dataset. In the similar dataset, select the two orders with the highest similarity coefficient as seed orders for a picking station, and delete all similarity coefficients related to the two selected orders in the similar dataset; From the remaining similarity coefficients, select the two orders with the highest similarity coefficients as the seed orders for the next picking station, and delete all similarity coefficients related to the two selected orders; repeat this step until a seed order is assigned to each picking station.

3. The method according to claim 2, characterized in that, Also includes: The difference in the number of orders assigned to two adjacent picking stations shall not exceed a preset threshold. When the number of orders assigned to any picking station is exactly such that the difference between the number of orders to its adjacent picking station does not exceed a preset threshold, then for orders that need to be assigned to that picking station, the orders are sorted according to the coupling factor and assigned to the picking station with the second largest coupling factor. If the picking station with the second largest coupling factor also has enough orders, then the orders are assigned to the picking station with the third largest coupling factor, so that the number of orders under each picking station is relatively balanced.

4. The method according to claim 3, characterized in that, The step of performing local optimization on each candidate order sequence in the initial candidate set using neighborhood search to obtain a preliminary locally optimal order sequence includes: For each candidate order sequence in the initial candidate set, enumerate all adjacent order pairs in the sequence in turn; Following the order of adjacent order pairs, the order of adjacent order pairs is swapped sequentially. After each swap, the minimum number of bin replacements is calculated for the new sequence after the swap. If the minimum number of bin replacements in the new sequence after the swap is less than the minimum number of bin replacements in the original sequence, then the new sequence is accepted. Continue swapping adjacent orders in the updated sequence until the order swapping of all adjacent orders can no longer reduce the minimum number of bin replacements, thus obtaining a preliminary locally optimal order sequence.

5. The method according to claim 4, characterized in that, The preliminary locally optimal order sequence is subjected to non-adjacent order swapping and local segment flipping operations in sequence to obtain the final candidate set, which includes: For any two non-adjacent order pairs in the preliminary local optimal order sequence, perform sequential swapping and obtain the minimum number of box replacements in the sequence after the sequential swap. If the minimum number of box replacements in the sequence after the swap is less than the minimum number of box replacements in the sequence before the swap, then accept the swap. Repeat this process until the sequential swapping of any two non-adjacent order pairs can no longer reduce the minimum number of box replacements. For the order sequence after non-adjacent order swaps, select any continuous subsequence and reverse all orders in the subsequence. Obtain the minimum number of box replacements in the reversed sequence. If the minimum number of box replacements in the reversed sequence is less than the minimum number of box replacements in the sequence before the reversed sequence, then accept the swap. Repeat this process until the reversed sequence of any continuous subsequence cannot reduce the minimum number of box replacements, and obtain the final candidate set.

6. The method according to claim 5, characterized in that, Also includes: In the process of performing neighborhood search for local optimization of the order sequence, or performing non-adjacent order swap operation on the order sequence, or performing local segment flip operation on the order sequence, an acceptance probability is set, which decreases stepwise with the number of iterations. If the minimum number of bin replacements in the updated order sequence is greater than or equal to the minimum number of bin replacements in the original order sequence, the update is accepted with the set acceptance probability, thus escaping the local optimum.

7. The method according to claim 6, characterized in that, The algorithm for determining the minimum number of bin replacements using a cluster search-based algorithm includes: Get the total number of orders assigned to any picking station, and then assign each order to the picking station. c Each order position is represented by an order number, and the order sequence for each order position is obtained. Set the initial state set for the picking station, which includes: the current order to be picked at each order position on the picking station. SKU The list, current bin, index of the next waiting order, and cumulative bin replacement count are initialized to 0; the bin arrival sequence is recorded. Determine whether each state in the initial state set of the picking station meets a preset completion state, wherein the preset completion state includes: all order positions. SKU All orders have been picked and all standby orders have been used up; If any state satisfies a preset completion state, then that state is saved to the completion state list; otherwise, the state is expanded by enumerating all possible picking operations to obtain a set of candidate states for the next stage; the state changes of the states in the candidate state set compared to the states before expansion include: For the order position corresponding to the picked order line, its order line value is decremented by 1. When this value reaches 0, it indicates that all orders at that position have been picked, and a new order needs to be added to the waiting order queue. The corresponding order line value is then updated to the new order's line number, and the order count is incremented by 1. This process is repeated depending on whether the current picking operation has occurred. SKU The system switches between states to determine whether a material bin needs to be replaced. If the material bin used in the new state is different from the one used in the previous state, it means that a material bin has been replaced, and the cumulative number of material bin replacements is incremented by 1. Each state in the candidate state set is sorted from low to high according to the score calculated by the comprehensive evaluation function. For states with the same score, they are sorted according to the order in which they were generated. The comprehensive evaluation function is calculated by: obtaining the comprehensive evaluation function score through the cumulative number of bin replacements, the remaining picking workload, the remaining quantity of backup orders, and a preset weight index for that state. Before selecting the comprehensive evaluation function score K Each state serves as the state set for the next stage; If a state that satisfies the completion condition exists in the new state set, it is saved to the completed state list, and the expansion ends; otherwise, if no state that satisfies the completion condition exists in the new state set. Then, for each state in the new state set, the process is expanded, and the top-scoring states in the comprehensive evaluation function are selected. K Each state continues as the state set for the next stage; it is checked again whether there is a state in the new state set that satisfies the completion condition; this step is repeated until there is a state in the new state set after a certain expansion that satisfies the completion condition, or the preset number of iterations is reached, or all states in the new state set after a certain expansion cannot be expanded, at which point the expansion ends; If there are completed states, then the state with the lowest comprehensive evaluation function score among the completed states is selected as the final optimal state; If there is no completed state, then select the state with the lowest comprehensive evaluation function score from the state set obtained in the last iteration as the final optimal state; The order sequence corresponding to the final optimal state is the optimal order sequence, and the cumulative number of bin replacements corresponding to the final optimal state is the minimum number of bin replacements for the picking station.

8. The method according to claim 7, characterized in that, The retention K The state expansion and pruning operations are performed iteratively on each global candidate state, including: For the reserved K Each global candidate state performs a picking operation once according to the optimal order sequence of each picking station, resulting in an expanded set of global candidate states; For each global candidate state in the expanded global candidate state set, a bin conflict detection is performed. If there is no bin conflict between picking stations, the global candidate state is valid; if there is a bin conflict, the global candidate state is excluded. A comprehensive evaluation function is used to score the expanded legal global candidate states, and the scores are sorted from low to high, retaining the lowest-scoring states. K One global candidate state; Repeat the above steps until all order picking tasks are completed or the maximum number of iterations is reached, to obtain the final global candidate state set.

9. An order processing decision-making device for the pharmaceutical distribution industry, characterized in that: The device includes: Seed order allocation module: used to allocate a seed order to each picking station based on the similarity coefficient between any two orders for all orders in a batch to be processed; The All Orders Assignment Module: For the remaining orders that were not assigned as seed orders, it calculates the quantity of the same type of goods ordered by each order and the seed orders already assigned to each picking station as a coupling factor, and assigns each remaining order to the picking station with the largest coupling factor, thus completing the order assignment. Preliminary order sequence acquisition module: It is used to generate multiple candidate orders for each picking station using a random method, sort them to form an initial candidate set, and perform local optimization on each candidate order sequence in the initial candidate set using neighborhood search to obtain a preliminary local optimal order sequence; Final order sequence acquisition module: used to sequentially perform non-adjacent order exchange and local segment flipping operations on the preliminary local optimal order sequence to obtain a final candidate set, and select the order sequence with the fewest bin replacements in the final candidate set as the optimal order sequence for the picking station; Local candidate state acquisition module: used to obtain the initial state set of each picking station according to the optimal order sequence within each picking station; each picking station independently performs a picking operation to generate the local candidate state set of each picking station; Global Candidate State Acquisition Module: This module is used to combine the local candidate state sets of each picking station into a global candidate state combination through Cartesian product; it performs bin conflict detection on each generated global candidate state. If there is no bin conflict between picking stations, the global candidate state is valid; if there is a bin conflict, the global candidate state is excluded. Global State Filtering Module: This module uses a comprehensive evaluation function to score valid global candidate states, sorts the scores from lowest to highest, and retains the lowest-scoring states. K One global candidate state; Loop extension module: used for reserved... K The global candidate state is repeatedly expanded and pruned until all order picking tasks are completed according to the optimal order sequence of each picking station or the maximum number of iterations is reached, and the final global candidate state set is obtained. Global optimal state acquisition module: used to select the state with the lowest comprehensive evaluation function score as the global optimal state from the final global candidate state set, and output the bin sorting sequence corresponding to the global optimal state.

10. A storage medium, characterized in that, The storage medium stores a computer program, which, when executed by the main controller, implements each step of the order processing decision-making method based on the pharmaceutical distribution industry as described in any one of claims 1-8.

Images