SKU Clustering for AGV Inventory Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing models for grouping inventory items in storage facilities are inefficient and inaccurate, leading to increased fuel consumption, computing resources, and time for automated guided vehicles (AGVs) due to the inefficiency in storing and retrieving items, especially when multiple aisles, racks, or shelves are involved, and simultaneous orders need to be fulfilled.
Innovation Solution
A computer-implemented method that determines processing clusters of stock keeping units (SKUs), divides them based on SKU affinities, replicates SKUs based on demand correlation, and assigns these clusters to physical locations to optimize storage and retrieval, using a SKU storing model that includes pick-cell stations for order fulfillment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If existing models group items at a single level, then the storage organization is simple, but multiple stops are required to retrieve requested items, increasing retrieval time and AGV fuel consumption
Solution Approach 1:
The patent segments the storage facility into multiple hierarchical levels (facility-level clusters, aisle-level clusters, rack-level clusters, and shelf-level clusters). This multi-level segmentation allows items to be organized and retrieved at different granularities, reducing the number of stops required by AGVs while maintaining systematic organization.
Solution Approach 2:
The patent introduces a hierarchical dimension to the storage organization structure, transforming the flat single-level grouping into a multi-level hierarchy. This dimensional change enables parallel retrieval operations at different levels and reduces traversal distance for AGVs by allowing selective access to specific cluster levels.
2Device complexity
If items are allocated to individual groups stored in single locations, then the storage arrangement is straightforward, but simultaneous orders containing the same item require multiple retrievals, increasing fuel consumption and time
Solution Approach 1:
The patent creates replicated copies of item clusters across multiple physical locations within the storage facility. When an item is allocated to a cluster, that cluster can be replicated to multiple aisles or racks, allowing AGVs to retrieve the same item from the nearest available location, thereby eliminating multiple trips for simultaneous orders.
Solution Approach 2:
The system performs preliminary clustering and replication of items based on predicted demand patterns and order history. By pre-positioning replicated clusters at strategic locations before orders arrive, the system prepares the storage structure to handle simultaneous order fulfillments efficiently, reducing actual retrieval time and AGV travel.
3Quantity of substance
If a limited number of AGVs are used, then the system is cost-effective, but fuel consumption and computational resources increase due to inefficient routing and multiple stops
Solution Approach 1:
The hierarchical clustering structure segments the storage facility into manageable units that can be independently accessed. This segmentation allows AGVs to service smaller, localized clusters rather than traversing the entire facility, reducing travel distance and fuel consumption per AGV while maintaining system capacity with fewer vehicles.
4Measurement precision
If existing models group massive numbers of items, then all items can be organized, but the grouping becomes inaccurate and inefficient, increasing computational resources and time
Solution Approach 1:
The patent divides the massive set of items into smaller hierarchical clusters at multiple levels. This segmentation reduces the computational complexity of grouping by processing items in manageable subsets rather than attempting to group all items simultaneously, while maintaining grouping accuracy through localized affinity calculations at each cluster level.
Solution Approach 2:
The hierarchical structure adds a dimensional framework to item grouping, organizing items from facility-level to shelf-level clusters. This dimensional approach improves grouping accuracy by considering local item affinities at each level while maintaining global organization, reducing computational burden compared to flat single-level grouping of all items.
Data Source
AI summary
A method determines a processing cluster including one or more stock keeping units (SKUs); divides the processing cluster into a first cluster and a second cluster based on SKU affinities between the one or more SKUs in the processing cluster; determines a first SKU of the first cluster to be replicated to the second cluster based on a demand correlation between the first SKU of the first cluster and a second SKU of the second cluster; replicates the first SKU of the first cluster to the second cluster; responsive to replicating the first SKU of the first cluster to the second cluster, determines whether the first cluster and the second cluster satisfy a defined constraint; and responsive to determining that the first cluster and the second cluster satisfy the defined constraint, assigns the first cluster to a first physical location and assigning the second cluster to a second physical location.