AGV Container Positioning Using Predictive Order Likelihoods
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage facility systems are inefficient in organizing inventory items and determining optimal container placement due to reliance on order history, which often results in frequent transportation of containers to and from picking stations, leading to inefficiencies in order fulfillment.
Innovation Solution
A method and system that determine container utilization likelihoods and optimal positions based on order likelihoods, using automated guided vehicles (AGVs) to transport containers to these positions, thereby optimizing storage and fulfillment processes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If containers are organized based on order history of item types, then frequently ordered items can be efficiently stored, but the majority of item types with low order frequencies cannot be effectively organized
Solution Approach 1:
The system changes the organizational parameter from historical order frequency to predictive order likelihood. By using machine learning models to predict future order probabilities, the system can effectively organize all item types including those with currently low order frequencies, while still optimizing for efficient fulfillment of predicted high-demand items.
2Reliability
If containers are transported frequently to and from picking stations, then order fulfillment can be maintained, but system efficiency deteriorates
Solution Approach 1:
The system performs preliminary actions by predicting which containers will be needed and transporting them to picking stations in advance, before actual orders are placed. This proactive approach reduces the frequency of reactive transportation operations while ensuring order fulfillment capability is maintained.
Solution Approach 2:
The system uses feedback from predictive models that continuously learn from actual order patterns to improve container placement decisions. By monitoring which predictions were accurate and adjusting the model accordingly, the system optimizes container positioning to minimize transportation while maintaining reliable order fulfillment.
3Productivity
If the system cannot determine optimal timing for releasing customer orders, then containers must be transported frequently, but this increases transportation overhead
Solution Approach 1:
The system releases customer orders in advance based on predictive likelihoods before they are actually needed. By using the predictive model to identify items likely to be ordered, the system can prepare and position containers proactively, avoiding the need for frequent reactive transportation operations.
Data Source
AI summary
In an example embodiment, a method determines a container containing one or more items associated with one or more item types, the container located at a current position in a storage facility; determines one or more order likelihoods of the one or more item types contained in the container; determines a container utilization likelihood of the container based on the one or more order likelihoods of the one or more item types contained in the container; determines an optimal position for the container in the storage facility based on the container utilization likelihood of the container, the optimal position being different from the current position of the container; and instructing an automated guided vehicle (AGV) to transport the container from the current position of the container to the optimal position of the container in the storage facility.