A stereoscopic warehouse aisle stacking machine cooperative access and AGV docking scheduling method

Abstract

The present application relates to the field of automated stereoscopic warehouse material handling and storage scheduling, and particularly relates to a stereoscopic warehouse roadway stacker cooperative storage and AGV docking scheduling method. The method comprises: for the docking time window mismatch caused by the composite cycle interruption, the small number of outbound cache bits, establishing a one-way reverse causal relationship of the next round of task pairing and the current release deadline and a completion time ladder mapping table, adjusting the task pairing and AGV service sequence, identifying the deadline cluster and iteratively adjusting, so that the AGV predicted taking time of all roadways is not later than the respective release deadline. The present application realizes the dynamic coordination of the docking time window under the condition of multi-lane shared AGV, and finally improves the reliability of the roadway docking, significantly reduces the risk of stacker waiting and cross-lane congestion, and enhances the continuity and operation efficiency of the automated stereoscopic warehouse inbound and outbound operation.

Description

Technical Field

[0001] This invention relates to the field of automated warehouse material handling and storage scheduling technology, and in particular to a method for collaborative storage and retrieval of stacker cranes and AGV docking and scheduling in automated warehouse aisle. Background Technology

[0002] Stacker cranes in automated warehouses are devices that move horizontally along aisle tracks and vertically lift and lower loading platforms to handle pallets in and out of the warehouse. Collaborative storage and retrieval systems, along with AGV (Automated Guided Vehicle) (AGV) docking and scheduling, allow the stacker crane to simultaneously complete both inbound and outbound operations in a single cycle, with AGVs handling pallet handover at the aisle entrance. Existing technologies employ fixed schedules, first-come-first-served, hierarchical scheduling, and heuristic allocation to improve inbound and outbound efficiency and reduce empty-trip distances.

[0003] However, existing methods have shortcomings when there are very few outbound buffer slots, uninterrupted compound cycles, and fluctuating AGV arrival times. The pairing of subsequent compound cycle tasks affects the release deadline of the current outbound pallet, and the actual AGV retrieval time is affected by path conflicts and congestion. Strict matching is difficult under zero buffer conditions, easily causing stacker crane waiting and subsequent task interruptions. When multiple aisles share AGVs, the release deadlines of multiple aisles are concentrated in the same time period. Traditional hierarchical scheduling struggles to coordinate task pairing and AGV service order, leading to cross-aisle congestion.

[0004] Therefore, it is urgent to study a joint scheduling method for minimal buffer space and shared AGVs to improve the reliability and operational efficiency of roadway intersection docking. Summary of the Invention

[0005] To overcome the drawback of mismatched time windows in cyclic docking, this invention provides a method for scheduling collaborative storage and retrieval of stacker cranes and docking with AGVs in automated warehouse aisles.

[0006] The technical implementation scheme of the present invention is as follows: a method for collaborative storage and retrieval of stacker cranes and AGV docking and scheduling in an automated warehouse aisle, comprising the following steps:

[0007] S1: Obtain the AGV's predicted pickup time and the task pairing for the next round of compound cycle, establish a one-way reverse causal relationship, and generate a completion time ladder mapping table;

[0008] S2: Based on the completion time ladder mapping table, adjust the task pairing or AGV service order for the next round of compound cycle, and recalculate the release deadline of the current outbound pallet for each lane;

[0009] S3: Based on the release deadline of the current outbound pallets in each lane and the AGV's travel time across lanes, identify the deadline time clusters, determine the degree of jump, and adjust the priority score;

[0010] S4: Jointly adjust the task pairing and AGV service order of the next round of compound cycle, with the constraint that the number of lanes in each deadline cluster is no greater than the number of AGVs, and iterate until the predicted AGV pick-up time of all lanes is no later than their respective release deadline.

[0011] Preferably, the step of obtaining the AGV's predicted pickup time and the task pairing for the next round of the compound cycle, establishing a one-way reverse causal relationship, and generating a completion time ladder mapping table includes:

[0012] Based on the location coordinates of the inbound and outbound tasks in the next round of compound cycle, establish the causal relationship between the task pairing in the next round of compound cycle and the time for the stacker crane to return to the aisle entrance in the next round.

[0013] When the time it takes for the next stacker crane to return to the aisle entrance determines the release deadline for the current outbound pallet, the causal relationship is defined as a one-way reverse causal relationship.

[0014] Preferably, determining the causal relationship as a one-way reverse causal relationship includes:

[0015] For all location coordinate combinations within the set of location coordinates that are allowed to be adjusted by the warehouse management system, which consists of the location coordinates of the inbound location corresponding to the current inbound task and the location coordinates of the outbound location corresponding to the current outbound task, calculate the next round of stacker crane return time to the aisle entrance for each location coordinate combination. Group location coordinate combinations with the same next round of stacker crane return time to the aisle entrance into the same group to obtain location coordinate combination groups. Determine the switching boundary between two adjacent location coordinate combination groups as the completion time jump boundary.

[0016] Establish a ladder mapping table of completion time between the location coordinate combination group and the discrete value of the completion time of the next round of stacker crane returning to the aisle entrance, and record the completion time jump boundary between adjacent location coordinate combination groups.

[0017] Preferably, the step of adjusting the task pairing or AGV service order for the next round of compound cycle according to the completion time ladder mapping table, and recalculating the release deadline of the current outbound pallet for each lane, includes:

[0018] For each aisle, based on the location coordinate combination in the task pairing of the next round of compound cycle, as well as the horizontal movement speed of the stacker crane, the vertical movement speed of the stacker crane, and the fork extension and retraction time, calculate the completion time of the stacker crane returning to the aisle entrance in the next round of the aisle, and determine the completion time as the release deadline of the current outbound pallet in the aisle according to the one-way reverse causal relationship.

[0019] When the predicted AGV retrieval time for a lane is later than the current release deadline for the outgoing pallet in that lane, the AGV service sequence is adjusted so that the predicted AGV retrieval time for that lane is no later than the release deadline for that lane.

[0020] If the predicted AGV pickup time of the lane is still later than the release deadline of the lane after adjusting the AGV service order, then the discrete value of the completion time that makes the sum of the start time of the next round of compound cycle and the completion time of the next round of stacker crane returning to the lane entrance no earlier than the predicted AGV pickup time of the lane is selected from the completion time ladder mapping table.

[0021] Among the selected cargo location coordinate combination groups corresponding to the discrete values ​​of completion time, select the cargo location coordinate combination with the smallest adjustment amount and replace the cargo location coordinate combination in the current task pairing of the roadway.

[0022] After the replacement, the time for the stacker crane to return to the lane entrance for the next round is recalculated for each lane.

[0023] Preferably, the recalculation of the next round of stacker crane return time to the lane entrance for each lane includes:

[0024] The time when the next stacker crane returns to the aisle entrance is taken as the release deadline for the current outbound pallet, and the release deadlines for each aisle are obtained by summing them up.

[0025] Preferably, the step of identifying deadline time clusters, determining the degree of jump, and adjusting priority scores based on the release deadline time of the current outbound pallets in each lane and the AGV's cross-lane travel time includes:

[0026] Arrange the release deadlines of the current outbound pallets in each lane according to the timeline; starting from the first release deadline, calculate the time difference between subsequent release deadlines and the earliest release deadline in the current deadline cluster; when the time difference is less than the time required for the AGV to travel between one lane entrance and another, the lane corresponding to the subsequent release deadline is added to the current deadline cluster; when the time difference is not less than the time required for the AGV to travel between one lane entrance and another, a new deadline cluster is established starting from the subsequent release deadline; determine the number of lanes in each deadline cluster.

[0027] For each lane within the same deadline time cluster, calculate the change in the completion time of the next round of stacker crane returning to the lane entrance after crossing a completion time jump boundary, and determine the jump degree of the lane.

[0028] For each roadway within the same deadline time cluster, calculate the target time difference between the release deadline time of that roadway and the boundary of the deadline time cluster; starting from the current cargo position coordinate combination group, accumulate the jump variables corresponding to the actual completion time jump boundaries crossed along the adjustment direction until the accumulated value reaches or exceeds the target time difference, and the accumulated number is taken as the number of completion time jump boundaries that need to be crossed.

[0029] Preferably, the cumulative number of times, as the number of completion time jump boundaries that need to be crossed, includes:

[0030] Based on the number of completion time jump boundaries to be crossed as required, determine the minimum cargo location coordinate adjustment amount required to cross that number of completion time jump boundaries from the current cargo location coordinate position, and use this minimum cargo location coordinate adjustment amount as the adjustment priority score for the roadway.

[0031] In the After the task pairing adjustment, if the tunnel If a time group leaves the original deadline cluster but does not enter another deadline cluster, then follow the formula. Update; if the adjusted roadway enters another cutoff time cluster, or the number of roadways in the original cutoff time cluster remains unchanged, then follow the formula. Update; Set lower limit ≥0; where In the completion time ladder mapping table, starting from the current cargo location coordinate group, along the direction that shifts the release deadline backward, the minimum cargo location coordinate adjustment amount is selected that satisfies the condition that the change in completion time is greater than 0 and not greater than half of the minimum interval between adjacent release deadlines within the same deadline time cluster. For the first After the task matching was adjusted, the alleyway Adjusting the priority score For the first After the secondary task pairing adjustment, the tunnel Adjustment priority score; after each update, roadways within the same cutoff time cluster are reordered according to the updated adjustment priority score.

[0032] Preferably, the joint adjustment of task pairing and AGV service order in the next round of compound cycle, constrained by the number of lanes within each deadline cluster not exceeding the number of AGVs, iterates until the predicted AGV pick-up time for all lanes is not later than their respective release deadlines, including:

[0033] For the roadway with the lowest adjustment priority score within the same deadline cluster, perform the next round of compound cycle task pairing adjustment;

[0034] After each task pairing adjustment, S3 is executed again; if the roadway leaves the original deadline time cluster after the adjustment and does not enter another deadline time cluster, it is determined that the deadline time cluster is eliminated, and the task pairing adjustment is retained; if the roadway leaves the original deadline time cluster and enters another deadline time cluster after the adjustment, it is determined that the deadline time cluster is transferred, and the task pairing adjustment is retained and the adjustment continues to be performed on the new deadline time cluster.

[0035] After each roadway is adjusted, it is reordered according to the updated adjustment priority score, and the roadway with the lowest adjustment priority score within the same deadline cluster is selected to perform the next task pairing adjustment.

[0036] Preferably, the step of reselecting the roadway with the lowest adjustment priority score within the same deadline cluster to perform the next task pairing adjustment includes:

[0037] After the task pairing adjustment, based on the release deadline of each lane, the order of AGV service for each lane is determined in ascending order of release deadline. After each assignment, if the estimated time for the AGV to arrive at the entrance of the next lane is later than the release deadline of the next lane, the next lane is assigned to the deadline time cluster to which its release deadline belongs, and the task pairing adjustment and AGV service order assignment are continued to be performed on the deadline time cluster.

[0038] When all lanes within the same deadline cluster have been adjusted and the number of lanes within that deadline cluster is still greater than the number of AGVs, adjust the AGV service order.

[0039] Preferably, adjusting the AGV service sequence includes:

[0040] After adjusting the AGV service order, S3 is re-executed. If a new deadline time cluster is generated, the aforementioned task pairing adjustment and AGV service order assignment are performed on the new deadline time cluster.

[0041] If the AGV predicted pick-up time for all lanes is not later than their respective release deadlines even after reaching the preset maximum number of iterations, a degradation strategy is executed, candidate scheduling schemes are output, and lanes that do not meet the conditions are marked. During repeated execution, if the number of consecutive adjustments for the same lane exceeds the number of lanes in the same deadline cluster, the adjustment priority score of that lane is increased. Alternatively, temporarily remove it from the current deadline time cluster's sorting; if the number of lanes in the original deadline time cluster does not decrease after a task pairing adjustment and the lane does not enter another deadline time cluster, then roll back the task pairing adjustment.

[0042] Beneficial Effects: This invention establishes a unidirectional reverse causal relationship between the next round of compound cycle task pairing and the current outbound pallet release deadline, generating a completion time ladder mapping table. Based on this, task pairing or AGV service order is adjusted, effectively overcoming the difficulty in matching release deadlines under conditions of extremely small buffer spaces and uninterrupted compound cycles. Furthermore, by identifying deadline clusters and quantifying the degree of jumps and adjusting priority scores, and using the constraint that the number of aisles within a cluster is no greater than the number of AGVs, the task pairing and AGV service order are iteratively adjusted, achieving dynamic coordination of docking time windows under conditions of multiple aisles sharing AGVs. Ultimately, while improving the reliability of aisle docking, it significantly reduces the risk of stacker crane waiting and cross-aisle blockage, enhancing the continuity and operational efficiency of automated warehouse inbound and outbound operations. Attached Figure Description

[0043] Figure 1 This is a flowchart of the method for coordinated storage and retrieval of stacker cranes and AGV docking and scheduling in an automated warehouse aisle according to the present invention.

[0044] Figure 2 This is a flowchart of the iterative adjustment process for deadline clusters based on priority scoring in this invention. Detailed Implementation

[0045] The present invention will be further described below with reference to the accompanying drawings and specific embodiments.

[0046] A method for coordinating the storage and retrieval of stacker cranes and the scheduling of AGVs in an automated warehouse aisle, such as... Figure 1 and Figure 2 As shown, it includes the following steps:

[0047] S1: Obtain the AGV's predicted pickup time and the task pairing for the next round of the compound cycle, establish a one-way reverse causal relationship, and generate a completion time ladder mapping table, including:

[0048] Based on the location coordinates of the inbound and outbound tasks in the next round of compound cycle, establish the causal relationship between the task pairing in the next round of compound cycle and the time for the stacker crane to return to the aisle entrance in the next round.

[0049] When the time it takes for the next stacker crane to return to the aisle entrance determines the release deadline for the current outbound pallet, the causal relationship is defined as a one-way reverse causal relationship.

[0050] The one-way reverse causal relationship is as follows: when the next round of task pairing is given, the release deadline of the current outbound pallet is uniquely determined by the next round of task pairing; the scheduling layer adjusts the next round of task pairing based on the comparison result of the AGV's predicted pick-up time and release deadline, and this adjustment belongs to the scheduling decision layer and does not change the calculation mapping direction from the next round of task pairing to the release deadline.

[0051] For all combinations of storage location coordinates within the set of storage location coordinates that are allowed to be adjusted by the warehouse management system, which consists of the storage location coordinates corresponding to the current storage task to be executed and the outbound storage location coordinates corresponding to the current outbound task to be executed, calculate the next round of stacker crane return time to the aisle entrance for each storage location coordinate combination, and group storage location coordinate combinations with the same next round of stacker crane return time to the aisle entrance into the same group to obtain storage location coordinate combination groups;

[0052] The adjacent cargo location coordinate combination group refers to the cargo location coordinate combination group corresponding to two adjacent completion time discrete values ​​after sorting the cargo location coordinate combination groups in ascending order of the completion time discrete value of the next round of stacker crane returning to the aisle entrance; all cargo location coordinate combinations corresponding to the same completion time discrete value belong to the same cargo location coordinate combination group.

[0053] The switching boundary between two adjacent cargo location coordinate combinations is defined as the completion time jump boundary; that is, the boundary between the corresponding cargo location coordinate combinations when the completion time jumps from one discrete value to an adjacent discrete value.

[0054] For example, suppose the x-coordinate of a warehouse entry location in a certain aisle can take values ​​of 1, 2, 3, 4, and 5, while the exit location remains fixed. Through enumeration and calculation, when x=1 and x=2, the stacker crane's return time to the aisle entrance is 10 seconds, belonging to location coordinate group A; when x=3 and x=4, the completion time is 12 seconds, belonging to location coordinate group B; and when x=5, the completion time is 15 seconds, belonging to location coordinate group C. The switching boundary between group A and group B lies between x=2 and x=3, which is the completion time jump boundary; the switching boundary between group B and group C lies between x=4 and x=5, which is also the completion time jump boundary. Crossing the boundary from x=2 to x=3, the completion time jumps from 10 seconds to 12 seconds; crossing the boundary from x=4 to x=5, the completion time jumps from 12 seconds to 15 seconds.

[0055] Establish a ladder mapping table of completion time between the location coordinate combination group and the discrete value of the completion time of the next round of stacker crane returning to the aisle entrance, and record the completion time jump boundary between adjacent location coordinate combination groups.

[0056] It should be noted that S1 obtains the predicted retrieval time of the automated guided vehicle (AGV) and pairs it with the next round of compound cycle tasks. The predicted retrieval time is calculated by the AGV scheduling system based on the task queue, path length, and conflict status, indicating the time when the AGV is expected to arrive at the aisle entrance and complete the pallet retrieval operation.

[0057] Specifically, for alleyways Its AGV predicts the pickup time. Calculate as follows:

[0058] ;

[0059] in, The current moment; The time required for the AGV to complete all tasks in the queue before retrieving a task in this alleyway; For AGVs to travel from their current location or the previous task point to the alleyway The travel time at the entrance is determined by the path length and the AGV's travel speed; The time required for the AGV to complete the pallet retrieval operation at the entrance of the alley; This refers to the waiting time incurred by the AGV along this path due to path conflicts or congestion. When the AGV service sequence is adjusted, it will be recalculated based on the adjusted service sequence. , , and This updates the various alleyways. The next round of task pairing is selected by the warehouse management system from the tasks to be executed, representing the combination of inbound and outbound tasks assigned in the next compound cycle.

[0060] The reason for establishing a causal relationship between the next round of task pairing and the next round of stacker crane return time to the aisle entrance is that the location coordinates of the next round of task pairing directly determine the horizontal and vertical movement distance of the stacker crane, and thus the return time. This return time is precisely the deadline before the current outbound pallet must be retrieved. Without establishing this causal relationship, it is difficult to know where the release deadline for the current outbound pallet comes from. When the completion time only determines the current release deadline and is not affected by the current release deadline in the reverse direction, this relationship is a one-way reverse causal relationship. That is, the current release deadline is determined only by the next round of task pairing, and the current release deadline does not affect the next round of task pairing in the reverse direction. Its purpose is to enable the scheduler to judge the feasibility of docking in advance, rather than discovering mismatch afterward.

[0061] Since the number of adjustable storage location coordinate combinations allowed by the warehouse management system is limited, the completion time for the next round of stacker crane to return to the aisle entrance corresponding to each storage location coordinate combination is also limited; these completion times are not continuous values, but rather a finite number of discrete values. Storage location coordinate combinations with the same completion time are grouped into the same storage location coordinate combination group, and completion time jump boundaries are determined between adjacent storage location coordinate combination groups, thereby establishing a completion time ladder mapping table between storage location coordinate combination groups and discrete completion time values.

[0062] The complete set of warehouse location coordinate combinations refers to the combinations within the set of warehouse location coordinates that are allowed to be adjusted by the warehouse management system, corresponding to the currently pending inbound and outbound tasks. It does not refer to combinations between all warehouse locations. A completion time ladder mapping table is pre-generated and stored offline. During online scheduling, only this mapping table is queried, eliminating the need for online enumeration of all warehouse location coordinate combinations. If an inbound or outbound location is constrained by the inventory strategy and cannot be adjusted, its coordinates are used as fixed values ​​in the mapping table query. Warehouse location coordinate combinations with the same completion time are grouped into the same group because these combinations produce the same release deadline. Grouping them allows identification of equivalent combinations, providing a range of options for subsequent adjustments. The switching boundary between two adjacent warehouse location coordinate combination groups is defined as the completion time jump boundary because this boundary is the critical point where the release deadline jumps. Only by crossing this boundary can the deadline be moved from one discrete value to another.

[0063] The completion time ladder mapping table takes the combination of cargo location coordinates as input and the discrete value of completion time as output, and records the completion time jump boundary. Its purpose is to allow direct lookup of the jump relationship between the next round of task pairing and the release deadline, eliminating the need for repeated enumeration during each adjustment. If the predicted retrieval time of the automated guided vehicle (AGV) is later than the release deadline, the table is consulted to reselect the cargo location coordinate combination across the completion time jump boundary, causing the release deadline to jump to a discrete value no earlier than the predicted retrieval time. This eliminates the mismatch in the cyclic docking time window, avoiding stacker crane waiting and cross-aisle blockage.

[0064] S2: Based on the completion time ladder mapping table, adjust the task pairing or AGV service order for the next round of compound loop, and recalculate the release deadline for the current outbound pallet for each lane, including:

[0065] For each aisle, based on the location coordinate combination in the task pairing of the next round of compound cycle, as well as the horizontal movement speed of the stacker crane, the vertical movement speed of the stacker crane, and the fork extension and retraction time, calculate the completion time of the stacker crane returning to the aisle entrance in the next round of the aisle, and determine the completion time as the release deadline of the current outbound pallet in the aisle according to the one-way reverse causal relationship.

[0066] When the predicted AGV retrieval time for a lane is later than the current release deadline for the outgoing pallet in that lane, the AGV service sequence is adjusted so that the predicted AGV retrieval time for that lane is no later than the release deadline for that lane.

[0067] If the predicted AGV pickup time of the lane is still later than the release deadline of the lane after adjusting the AGV service order, then the discrete value of the completion time that makes the sum of the start time of the next round of compound cycle and the completion time of the next round of stacker crane returning to the lane entrance no earlier than the predicted AGV pickup time of the lane is selected from the completion time ladder mapping table.

[0068] Among the selected cargo location coordinate combination groups corresponding to the discrete values ​​of completion time, select the cargo location coordinate combination with the smallest adjustment amount and replace the cargo location coordinate combination in the current task pairing of the roadway.

[0069] If no discrete value of completion time that meets the conditions can be selected, the cargo location coordinate combination in the current task pairing of the roadway is retained, and the roadway is marked as a roadway to be adjusted, and then the process is transferred to S4 to continue.

[0070] After replacement, recalculate the time for the stacker crane to return to the lane entrance for the next round in each lane;

[0071] The time when the next stacker crane returns to the aisle entrance is taken as the release deadline for the current outbound pallet, and the release deadlines for each aisle are obtained by summing them up.

[0072] It should be noted that in S2, the combination of storage location coordinates in the next round of compound cycle task pairing determines the inbound and outbound storage locations. Let the coordinates of the aisle entrance be... Its unit is the coordinate unit of the storage location, and the coordinates of the entrance to the same aisle are fixed. The coordinates of the storage location are... The coordinates of the outbound storage location are Both units are in the coordinate system of cargo position, and both are rounded to the nearest integer. The horizontal movement speed is... Its unit is cargo position coordinates per second, and the vertical movement speed is... Its unit is cargo position coordinates per second, and the fork extension / retraction time is... Its unit is seconds.

[0073] The next round of the compound cycle will proceed in the following order: starting from the aisle entrance, first moving to the corresponding inbound storage location for the inbound task, then moving to the corresponding outbound storage location for the outbound task, and finally returning to the aisle entrance. Horizontal and vertical movements will occur in parallel, and the larger of the horizontal and vertical movement times will be used. Fork extension / retraction time will be counted twice, once for inbound loading and once for outbound retrieval. Waiting time will be disregarded. Therefore, the time required for the stacker crane to return to the aisle entrance in the next round is... for: ;

[0074] in, This is the duration, in seconds, from the start of the next compound cycle until the stacker crane reaches the aisle entrance. This time does not include the fork extension and retraction time for the stacker crane to lower the next batch of outbound pallets after reaching the aisle entrance.

[0075] The above calculation is a simplified model at the scheduling level, ignoring the acceleration and deceleration processes of the stacker crane's horizontal and vertical movements, and assuming that each segment runs at a uniform speed. If the actual stacker crane accelerates or decelerates significantly, an acceleration / deceleration compensation time can be added to the corresponding time segment, or the equivalent speed obtained through calibration can be substituted into the above formula; this simplification does not affect the feasibility of this scheduling method.

[0076] Let the start time of the next compound cycle be... The unit is seconds, and it is determined by the warehouse management system based on the end time of the current compound cycle; therefore, the release deadline for the current outbound pallet is... For absolute times, calculate as follows: This release deadline indicates that the AGV must remove the current outbound pallet before this absolute moment to free up the outbound buffer space so that the stacker crane can place the next batch of outbound pallets after reaching the aisle entrance. If the actual system requires the release deadline to include the time it takes for the stacker crane to complete unloading at the aisle entrance, then it should be within [the specified time]. Add another fork extension / retraction time in the middle. This embodiment uses the boundary definition of "completion upon reaching the tunnel entrance". This allows us to obtain the completion time for the next round of stacker crane returning to the tunnel entrance, and generate a completion time ladder mapping table accordingly.

[0077] Based on a one-way reverse causal relationship, this completion time is the release deadline for the current outbound pallet. This is because if the current pallet still occupies the outbound buffer space when the stacker crane returns in the next round, the stacker crane will have nowhere to place the new pallet. When the AGV predicts a pickup time later than the release deadline, the AGV service sequence is first adjusted, moving the position of that aisle forward in the AGV service sequence. This ensures that the AGV prioritizes heading to that aisle after completing its current task, reducing the waiting time of that aisle in the AGV service queue and bringing the pickup time forward to no later than the release deadline, guaranteeing that the buffer space is freed up before the next round of returns. If, after adjustment, the time is still later than the release deadline, then the completion time is filtered in the completion time ladder mapping table to ensure that the time is not later than the release deadline. No earlier than the AGV's predicted pickup time The discrete value of the completion time is required because only this discrete value can meet the retrieval time requirement.

[0078] Among the selected discrete values ​​of completion time corresponding to the storage location coordinate combinations, the combination with the smallest adjustment is chosen and replaced in the current task pairing. This ensures that the completion time of the next round of stacker crane returning to the aisle entrance becomes the selected discrete value, thus shifting the release deadline backward across the transition position. Subsequently, the adjusted completion time of each aisle is calculated based on a one-way reverse causal relationship and used as the release deadline for the current outbound pallet. This data is then aggregated to identify deadline time clusters, eliminate mismatches in the cyclic docking time window, and prevent stacker crane waiting and cross-aisle blockage.

[0079] S3: Based on the release deadline of the current outbound pallets in each lane and the AGV's cross-lane travel time, identify deadline time clusters, determine the degree of transition, and adjust priority scores, including:

[0080] Arrange the release deadlines of the current outbound pallets in each lane according to the timeline; starting from the first release deadline, calculate the time difference between subsequent release deadlines and the earliest release deadline in the current deadline cluster; when the time difference is less than the time required for the AGV to travel between one lane entrance and another, the lane corresponding to the subsequent release deadline is added to the current deadline cluster; when the time difference is not less than the time required for the AGV to travel between one lane entrance and another, a new deadline cluster is established starting from the subsequent release deadline; determine the number of lanes in each deadline cluster.

[0081] For each lane within the same deadline time cluster, calculate the change in the completion time of the next round of stacker crane returning to the lane entrance after crossing a completion time jump boundary, and determine the jump degree of the lane; that is, the jump degree is the change in completion time corresponding to each crossing of a completion time jump boundary.

[0082] The change in completion time after crossing a completion time jump boundary is equal to the absolute value of the difference between the two adjacent discrete values ​​of completion time at that completion time jump boundary.

[0083] If the jump variables of adjacent completion time jump boundaries are different, the actual completion time jump boundaries crossed are accumulated one by one until the accumulated value reaches or exceeds the target time difference.

[0084] For each roadway within the same deadline time cluster, calculate the target time difference between the release deadline time of that roadway and the boundary of the deadline time cluster; starting from the current cargo position coordinate combination group, accumulate the jump variables corresponding to the actual completion time jump boundaries crossed along the adjustment direction until the accumulated value reaches or exceeds the target time difference, and the accumulated number is taken as the number of completion time jump boundaries that need to be crossed;

[0085] For example, if the target time difference is 5 seconds, and the jump variable at each completion time boundary is 2 seconds, then 5 ÷ 2 = 2.5, rounded up to 3, meaning 3 completion time boundary jumps need to be crossed. If the jump variables at each completion time boundary are 1 second, 3 seconds, and 1 second respectively, then starting from the current cargo location coordinate combination, each jump is accumulated: 1 < 5, 1 + 3 = 4 < 5, 1 + 3 + 1 = 5 ≥ 5, accumulating 3 times, meaning 3 completion time boundary jumps still need to be crossed. Therefore, when the jump variables are uniform, the ratio is directly rounded up; when the jump variables are uneven, each jump variable needs to be accumulated until the cumulative time difference reaches or exceeds the target time difference. The above division and rounding method only applies to the simplified case where the jump variables are uniform; when the jump variables are uneven, they are accumulated one by one.

[0086] The boundary of the cutoff time cluster is the boundary on one side of the earliest or latest release cutoff time within the cluster that corresponds to the adjustment direction. When it is necessary to move the release cutoff time backward away from the current cluster, the boundary is the sum of the latest release cutoff time within the cluster and the AGV's travel time across the aisle; when it is necessary to move it forward, the boundary is the earliest release cutoff time within the cluster minus the AGV's travel time across the aisle; the target time difference is the time difference between the release cutoff time of the aisle and the boundary.

[0087] Based on the number of completion time jump boundaries to be crossed as required, determine the minimum cargo location coordinate adjustment amount required to cross that number of completion time jump boundaries from the current cargo location coordinate position, and use this minimum cargo location coordinate adjustment amount as the adjustment priority score for the roadway.

[0088] In the After the task pairing adjustment, if the tunnel If a time group leaves the original deadline cluster but does not enter another deadline cluster, then follow the formula. Update; if the adjusted roadway enters another cutoff time cluster, or the number of roadways in the original cutoff time cluster remains unchanged, then follow the formula. Update; Set lower limit ≥0;

[0089] in In the completion time ladder mapping table, starting from the current cargo location coordinate group, along the direction that shifts the release deadline backward, the minimum cargo location coordinate adjustment amount is selected that satisfies the condition that the change in completion time is greater than 0 and not greater than half of the minimum interval between adjacent release deadlines within the same deadline time cluster. It has the same dimension as the priority adjustment score. After rounding... <1, and If the change in the time it takes for the next stacker crane to return to the aisle entrance corresponding to =1 is no greater than half the minimum interval between adjacent release deadlines within the same deadline time cluster, then let... =1, to ensure that each adjustment spans at least one cargo location coordinate unit; otherwise, the task pairing adjustment will not be performed this time, or the adjustment will be moved to the AGV service sequence.

[0090] The minimum interval between adjacent release deadlines within the same deadline time cluster is obtained as follows: sort the release deadline times of each roadway within the same deadline time cluster in ascending order of time axis, calculate the difference between two adjacent release deadline times, take the minimum value among all differences, and then take half of the minimum value as the threshold.

[0091] For the first After the task matching was adjusted, the alleyway Adjusting the priority score Adjusting the number of iterations for task pairing =1,2,…; The roadway number is used as the initial value, calculated based on the minimum cargo location coordinate adjustment amount before the first task pairing adjustment. , For the first After the secondary task pairing adjustment, the tunnel Adjusting priority scores; Adjusting priority scores Set a lower limit. ≥0; if updated according to the formula If <0, then let =0;

[0092] After each update, the roadways within the same cutoff time cluster are reordered according to the updated adjustment priority score.

[0093] It should be noted that in S3, after the release deadlines for each lane are arranged along the time axis, the time interval between adjacent deadlines is used to determine whether the Automated Guided Vehicle (AVR) can continuously serve two lanes within the time required to travel from one lane entrance to another. If the time interval is less than the travel time, it indicates that the AVR cannot simultaneously meet the retrieval requirements of two lanes, and these two lanes are grouped into the same deadline cluster. The number of lanes in the cluster represents the total number of lanes competing for the AVR during that time period. The change in completion time after crossing the completion time jump boundary is calculated because this change represents the amount of deadline time shift obtained by crossing each jump position, i.e., the jump degree. The time difference between the release deadline and the deadline cluster boundary is accumulated one by one with the jump degree, and the accumulated number is the number of completion time jump boundaries that need to be crossed, representing the minimum number of jumps required for the lane to leave the current deadline time cluster. The minimum amount of cargo location coordinate adjustment required to cross the time jump boundary of the current cargo location from the current cargo location coordinate position is the adjustment priority score, which represents the adjustment cost. This minimum amount of cargo location coordinate adjustment is obtained by enumerating all candidate cargo location coordinate combinations that meet the requirements of crossing the time jump boundary of the current cargo location from the time jump boundary of the current cargo location coordinate position, and taking the one with the smallest amount of cargo location coordinate adjustment between it and the current cargo location coordinate position.

[0094] Specifically, the adjustment amount for the cargo location coordinates is based on the Manhattan distance, that is, for the current cargo location coordinates... and candidate cargo location coordinates The adjustment amount of the cargo location coordinates is The adjustment amount of the storage location coordinates shall be adjusted first for the storage location coordinates corresponding to the storage task; if adjusting only the storage location coordinates is insufficient to meet the crossing requirements, then the storage location coordinates corresponding to the outbound task shall be adjusted; if further adjustment is still required, the storage location coordinates of both the storage and outbound locations shall be adjusted simultaneously, and the sum of the two adjustments shall be used as the total storage location coordinate adjustment amount.

[0095] The jump direction is determined as follows: When it is necessary to postpone the release deadline, the direction that increases the completion time of the next round of stacker crane returning to the aisle entrance is selected to cross the completion time jump boundary; when it is necessary to postpone the release deadline, the direction that decreases the completion time of the next round of stacker crane returning to the aisle entrance is selected to cross the completion time jump boundary. Crossing a completion time jump boundary means moving from the current location coordinate group along the adjustment direction to the adjacent location coordinate group, counting once for each completion time jump boundary. If, after adjustment, the aisle leaves the original deadline time cluster and does not enter another deadline time cluster, it indicates that the conflict has been eliminated, the deadline time cluster has been eliminated, and the adjustment priority score has been reduced. This is to encourage this direction; if, after adjustment, the roadway leaves the original cutoff time cluster and enters another cutoff time cluster, it indicates a conflict transfer, is identified as a cutoff time cluster transfer, and the priority score is increased. The priority of that direction is reduced; if the number of roadways in the original cutoff time cluster remains unchanged after the adjustment and that roadway does not enter another cutoff time cluster, the adjustment is invalid, and the adjustment priority score is increased. And roll back the task pairing; if the number of lanes in the original deadline cluster does not decrease after a task pairing adjustment and the lane does not enter another deadline cluster, then roll back the task pairing adjustment.

[0096] The rollback only rolls back the selected cargo location coordinate combination for this task pairing adjustment, and the adjustment priority score is retained according to the value after this update; that is, if the number of lanes in the original deadline cluster does not decrease after this adjustment and the lane does not enter another deadline cluster, the value of the adjustment priority score after adding a preset penalty step size is retained to reduce the priority of subsequent selections in the same adjustment direction.

[0097] S4: Jointly adjust the task pairing and AGV service order for the next round of compound loop, with the constraint that the number of lanes in each deadline cluster is no greater than the number of AGVs, and iterate until the predicted AGV pick-up time for all lanes is no later than their respective release deadlines, including:

[0098] For the roadway with the lowest adjustment priority score within the same deadline cluster, perform the next round of compound cycle task pairing adjustment;

[0099] After each task pairing adjustment, S3 is executed again; if the roadway leaves the original deadline time cluster after the adjustment and does not enter another deadline time cluster, it is determined that the deadline time cluster is eliminated, and the task pairing adjustment is retained; if the roadway leaves the original deadline time cluster and enters another deadline time cluster after the adjustment, it is determined that the deadline time cluster is transferred, and the task pairing adjustment is retained and the adjustment continues to be performed on the new deadline time cluster.

[0100] After each roadway is adjusted, it is reordered according to the updated adjustment priority score, and the roadway with the lowest adjustment priority score within the same deadline cluster is selected to perform the next task pairing adjustment.

[0101] After the task pairing adjustment, based on the release deadline of each lane, the order of AGV service for each lane is determined in ascending order of release deadline. After each assignment, if the estimated time for the AGV to arrive at the entrance of the next lane is later than the release deadline of the next lane, the next lane is assigned to the deadline time cluster to which its release deadline belongs, and the task pairing adjustment and AGV service order assignment are continued to be performed on the deadline time cluster.

[0102] When all lanes in the same deadline cluster have been adjusted and the number of lanes in the same deadline cluster is still greater than the number of AGVs, adjust the AGV service order.

[0103] After adjusting the AGV service order, S3 is re-executed. If a new deadline time cluster is generated, the aforementioned task pairing adjustment and AGV service order assignment are performed on the new deadline time cluster.

[0104] If the AGV predicted pick-up time for all lanes is not later than their respective release deadlines even after reaching the preset maximum number of iterations, a degradation strategy is executed, candidate scheduling schemes are output, and lanes that do not meet the conditions are marked. During repeated execution, if the number of consecutive adjustments for the same lane exceeds the number of lanes in the same deadline cluster, the adjustment priority score of that lane is increased. Or temporarily remove it from the sorting of the current deadline cluster;

[0105] The sorting of the roadway temporarily removed from the current deadline time cluster is only valid for the current iteration round. After S3 is re-executed in the next round, the deadline time cluster to which the roadway belongs is re-determined according to the updated release deadline time, and its sorting within the deadline time cluster is restored. If the number of consecutive adjustments exceeds the number of roadways in the same deadline time cluster again after restoration, the roadway is temporarily removed from the sorting of the current deadline time cluster again.

[0106] If the number of lanes in the original deadline cluster does not decrease after a task pairing adjustment and the lane does not enter another deadline cluster, then the task pairing adjustment is rolled back.

[0107] It should be noted that the reference Figure 2In S4, the roadway with the lowest adjustment priority score is prioritized for adjustment because it has the lowest adjustment cost, minimizing disturbance and allowing the release deadline to leave the current deadline cluster or reducing the number of roadways within the cluster. S3 is re-executed after each adjustment because changes in the cargo location coordinates can cause the release deadline to jump, requiring re-identification of the deadline cluster, the degree of jump, and the adjustment priority score. If the roadway leaves the original deadline cluster but does not enter another deadline cluster, the conflict is resolved, the deadline cluster is eliminated, and the task's pairing adjustment is retained. If the roadway leaves the original deadline cluster and enters another deadline cluster, the conflict is transferred, the deadline cluster is transferred, the task's pairing adjustment is retained, and adjustments continue to be performed on the new deadline cluster. If the number of roadways within the original deadline cluster remains unchanged and the roadway does not enter another deadline cluster, the task's pairing adjustment is rolled back.

[0108] After adjustment, the order is reordered and the smallest value is selected, ensuring progress is always made at the lowest cost. Then, in ascending order of release deadline, each lane is assigned to the AGV with the earliest estimated retrieval time, and the estimated retrieval time of that AGV is updated. Updating the estimated retrieval time involves recalculating the AGV's estimated retrieval time for the next lane based on the task queue time after the AGV completes its assigned tasks, the actual travel time from the current lane entrance to the next lane entrance, the retrieval operation time, and the path conflict waiting time. The service order of each AGV is the order in which the lanes to which it was assigned are formed.

[0109] The time it takes for an AGV to travel from the current lane entrance to the next lane entrance is calculated based on the actual path length between the two lane entrances and the AGV's travel speed, rather than using a fixed value. If the distance between different lane pairs is different, the actual cross-lane travel time corresponding to the current lane pair is used when identifying the deadline cluster. Specifically, if the number of AGVs in the cluster after all lane adjustments is still greater than the number of AGVs, it means that pairing adjustments are insufficient to eliminate conflicts, and the AGV service order needs to be adjusted, specifically by reordering them according to the release deadline in ascending order. After the service order is changed, S3 is executed again, and adjustments continue if a new deadline cluster appears. This process is repeated until the predicted AGV pick-up time for all lanes is no later than the release deadline. If the number of consecutive adjustments for the same lane exceeds the number of lanes in the cluster, the adjustment priority score for that lane is increased. Alternatively, the current deadline cluster may be temporarily removed from the sorting to avoid invalid loops. The step size when the priority score is increased is a preset penalty step size. The dimension of the preset penalty step size is the same as that of the priority score, which is the adjustment amount of the cargo location coordinates.

[0110] If the number of lanes within the cluster does not decrease after adjustment, the adjustment should be rolled back to prevent the conflict from escalating.

[0111] To avoid the iteration process becoming difficult to terminate, a preset maximum number of iterations, Nmax, is set. Nmax is preset based on the number of lanes, the number of AGVs, and the number of deadline clusters, for example, it is taken as an integer multiple of the product of the number of lanes and the number of AGVs. When the iteration count reaches Nmax and the predicted AGV pick-up time for all lanes is not later than their respective release deadlines, a degradation strategy is executed: outputting the current release deadlines and AGV service order of each lane as candidate scheduling schemes, and marking lanes that do not meet the conditions; or retaining the original task pairings of lanes that do not meet the conditions and issuing manual intervention prompts; or adjusting the AGV service order and restarting the iteration until the preset total iteration limit is reached. The candidate scheduling schemes do not guarantee that the predicted AGV pick-up time for all lanes will be no later than the release deadline.

[0112] It should be further noted that the parameters involved in this application, such as the horizontal movement speed of the stacker crane, the vertical movement speed, the fork extension and retraction time, the AGV travel speed, the AGV travel time across aisles, the AGV pick-up operation time, the task queue time, the path conflict waiting time, and the start time of the next compound cycle, are all obtained in real time from the equipment manual, on-site calibration, and the warehouse management system or AGV scheduling system. The completion time ladder mapping table is generated offline and stored after the warehouse layout, equipment parameters, and the adjustable storage location strategy of the warehouse management system are determined. Each threshold, such as half of the minimum interval between adjacent release deadlines, is calculated from real-time data. The maximum number of iterations is a preset integer based on the number of aisles, the number of AGVs, and the number of deadline clusters.

[0113] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for collaborative storage and retrieval of stacker cranes and scheduling of AGVs in an automated warehouse aisle, characterized in that, Includes the following steps: S1: Obtain the AGV's predicted pickup time and the task pairing for the next round of compound cycle, establish a one-way reverse causal relationship, and generate a completion time ladder mapping table; S2: Based on the completion time ladder mapping table, adjust the task pairing or AGV service order for the next round of compound cycle, and recalculate the release deadline of the current outbound pallet for each lane; S3: Based on the release deadline of the current outbound pallets in each lane and the AGV's travel time across lanes, identify the deadline time clusters, determine the degree of jump, and adjust the priority score; S4: Jointly adjust the task pairing and AGV service order of the next round of compound cycle, with the constraint that the number of lanes in each deadline cluster is no greater than the number of AGVs, and iterate until the predicted AGV pick-up time of all lanes is no later than their respective release deadline.

2. The method for collaborative storage and retrieval of stacker cranes and AGV docking and scheduling in an automated warehouse aisle as described in claim 1, characterized in that, The process of obtaining the AGV's predicted pickup time and matching tasks for the next round of the compound cycle, establishing a one-way reverse causal relationship, and generating a completion time ladder mapping table includes: Based on the location coordinates of the inbound and outbound tasks in the next round of compound cycle, establish the causal relationship between the task pairing in the next round of compound cycle and the time for the stacker crane to return to the aisle entrance in the next round. When the time it takes for the next stacker crane to return to the aisle entrance determines the release deadline for the current outbound pallet, the causal relationship is defined as a one-way reverse causal relationship.

3. The method for collaborative storage and retrieval of stacker cranes and AGV docking and scheduling in an automated warehouse aisle as described in claim 2, characterized in that, The determination of a causal relationship as a one-way reverse causal relationship includes: For all location coordinate combinations within the set of location coordinates that are allowed to be adjusted by the warehouse management system, which consists of the location coordinates of the inbound location corresponding to the current inbound task and the location coordinates of the outbound location corresponding to the current outbound task, calculate the next round of stacker crane return time to the aisle entrance for each location coordinate combination. Group location coordinate combinations with the same next round of stacker crane return time to the aisle entrance into the same group to obtain location coordinate combination groups. Determine the switching boundary between two adjacent location coordinate combination groups as the completion time jump boundary. Establish a ladder mapping table of completion time between the location coordinate combination group and the discrete value of the completion time of the next round of stacker crane returning to the aisle entrance, and record the completion time jump boundary between adjacent location coordinate combination groups.

4. The method for collaborative storage and retrieval of stacker cranes and AGV docking and scheduling in an automated warehouse aisle as described in claim 1, characterized in that, The step of adjusting the task pairing or AGV service order for the next round of compound cycles based on the completion time ladder mapping table, and recalculating the release deadline for the current outbound pallets in each lane, includes: For each aisle, based on the location coordinate combination in the task pairing of the next round of compound cycle, as well as the horizontal movement speed of the stacker crane, the vertical movement speed of the stacker crane, and the fork extension and retraction time, calculate the completion time of the stacker crane returning to the aisle entrance in the next round of the aisle, and determine the completion time as the release deadline of the current outbound pallet in the aisle according to the one-way reverse causal relationship. When the predicted AGV retrieval time for a lane is later than the current release deadline for the outgoing pallet in that lane, the AGV service sequence is adjusted so that the predicted AGV retrieval time for that lane is no later than the release deadline for that lane. If the predicted AGV pickup time of the lane is still later than the release deadline of the lane after adjusting the AGV service order, then the discrete value of the completion time that makes the sum of the start time of the next round of compound cycle and the completion time of the next round of stacker crane returning to the lane entrance no earlier than the predicted AGV pickup time of the lane is selected from the completion time ladder mapping table. Among the selected cargo location coordinate combination groups corresponding to the discrete values ​​of completion time, select the cargo location coordinate combination with the smallest adjustment amount and replace the cargo location coordinate combination in the current task pairing of the roadway. After the replacement, the time for the stacker crane to return to the lane entrance for the next round is recalculated for each lane.

5. The method for collaborative storage and retrieval of stacker cranes and AGV docking and scheduling in an automated warehouse aisle as described in claim 4, characterized in that, The recalculation of the next round of stacker crane return time to the lane entrance for each lane includes: The time when the next stacker crane returns to the aisle entrance is taken as the release deadline for the current outbound pallet, and the release deadlines for each aisle are obtained by summing them up.

6. The method for collaborative storage and retrieval of stacker cranes and AGV docking and scheduling in an automated warehouse aisle as described in claim 1, characterized in that, The process involves identifying deadline clusters based on the release deadline of the current outbound pallets in each lane and the AGV's cross-lane travel time, determining the degree of jump, and adjusting priority scores, including: Starting from the first release deadline, calculate the time difference between subsequent release deadlines and the earliest release deadline in the current deadline cluster. If the time difference is less than the time required for the AGV to travel between one lane entrance and another, the lane corresponding to the subsequent release deadline is added to the current deadline cluster. If the time difference is not less than the time required for the AGV to travel between one lane entrance and another, a new deadline cluster is established starting from the subsequent release deadline. Determine the number of lanes in each deadline cluster. For each lane within the same deadline time cluster, calculate the change in the completion time of the next round of stacker crane returning to the lane entrance after crossing a completion time jump boundary, and determine the jump degree of the lane. For each roadway within the same deadline time cluster, calculate the target time difference between the release deadline time of that roadway and the boundary of the deadline time cluster; starting from the current cargo position coordinate combination group, accumulate the jump variables corresponding to the actual completion time jump boundaries crossed along the adjustment direction until the accumulated value reaches or exceeds the target time difference, and the accumulated number is taken as the number of completion time jump boundaries that need to be crossed.

7. The method for collaborative storage and retrieval of stacker cranes and AGV docking and scheduling in an automated warehouse aisle as described in claim 6, characterized in that, The cumulative number of times, representing the number of completion time jump boundaries that need to be crossed, includes: Based on the number of completion time jump boundaries to be crossed as required, determine the minimum cargo location coordinate adjustment amount required to cross that number of completion time jump boundaries from the current cargo location coordinate position, and use this minimum cargo location coordinate adjustment amount as the adjustment priority score for the roadway. In the After the task pairing adjustment, if the tunnel If a time group leaves the original deadline cluster but does not enter another deadline cluster, then follow the formula. Update; if the adjusted roadway enters another cutoff time cluster, or the number of roadways in the original cutoff time cluster remains unchanged, then follow the formula. Update; Set lower limit ≥0; where In the completion time ladder mapping table, starting from the current cargo location coordinate group, along the direction that shifts the release deadline backward, the minimum cargo location coordinate adjustment amount is selected that satisfies the condition that the change in completion time is greater than 0 and not greater than half of the minimum interval between adjacent release deadlines within the same deadline time cluster. For the first After the task pairing adjustment, the alleyway Adjusting the priority score For the first After the secondary task pairing adjustment, the tunnel Adjustment priority score; after each update, roadways within the same cutoff time cluster are reordered according to the updated adjustment priority score.

8. The method for collaborative storage and retrieval of stacker cranes and AGV docking and scheduling in an automated warehouse aisle as described in claim 1, characterized in that, The joint adjustment of task pairing and AGV service order in the next round of compound cycle is constrained by the number of lanes in each deadline cluster not exceeding the number of AGVs. Iteration continues until the predicted AGV pick-up time for all lanes is no later than their respective release deadlines, including: For the roadway with the lowest adjustment priority score within the same deadline cluster, perform the next round of compound cycle task pairing adjustment; After each task pairing adjustment, S3 is executed again; if the roadway leaves the original deadline time cluster after the adjustment and does not enter another deadline time cluster, it is determined that the deadline time cluster is eliminated, and the task pairing adjustment is retained; if the roadway leaves the original deadline time cluster and enters another deadline time cluster after the adjustment, it is determined that the deadline time cluster is transferred, and the task pairing adjustment is retained and the adjustment continues to be performed on the new deadline time cluster. After each roadway is adjusted, it is reordered according to the updated adjustment priority score, and the roadway with the lowest adjustment priority score within the same deadline cluster is selected to perform the next task pairing adjustment.

9. The method for collaborative storage and retrieval of stacker cranes and AGV docking and scheduling in an automated warehouse aisle as described in claim 8, characterized in that, The process of reselecting the roadway with the lowest adjustment priority score within the same deadline cluster to perform the next task pairing adjustment includes: After the task pairing adjustment, based on the release deadline of each lane, the order of AGV service for each lane is determined in ascending order of release deadline. After each assignment, if the estimated time for the AGV to arrive at the entrance of the next lane is later than the release deadline of the next lane, the next lane is assigned to the deadline time cluster to which its release deadline belongs, and the task pairing adjustment and AGV service order assignment are continued to be performed on the deadline time cluster. When all lanes within the same deadline cluster have been adjusted and the number of lanes within that deadline cluster is still greater than the number of AGVs, adjust the AGV service order.

10. The method for collaborative storage and retrieval of stacker cranes and AGV docking and scheduling in an automated warehouse aisle as described in claim 9, characterized in that, The adjustment of the AGV service order includes: After adjusting the AGV service order, S3 is executed again. If a new deadline time cluster is generated, the aforementioned task pairing adjustment and AGV service order assignment are performed on the new deadline time cluster. If the AGV predicted pick-up time for all lanes is not later than their respective release deadlines even after reaching the preset maximum number of iterations, a degradation strategy is executed, candidate scheduling schemes are output, and lanes that do not meet the conditions are marked. During repeated execution, if the number of consecutive adjustments for the same lane exceeds the number of lanes in the same deadline cluster, the adjustment priority score of that lane is increased. Alternatively, temporarily remove it from the current deadline time cluster's sorting; if the number of lanes in the original deadline time cluster does not decrease after a task pairing adjustment and the lane does not enter another deadline time cluster, then roll back the task pairing adjustment.

Images