Inventory Forecasting System Using Time-Span Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large warehouses face complexity in inventory management due to diverse stock keeping units with varying characteristics, requiring efficient methods to calculate consumption forecasts and optimize ordering processes while managing resource usage effectively.
Innovation Solution
A method and system that determine forecast values and order quantities for stock keeping units across different time spans, minimizing deviation between forecast and order quantity time spans, using historical consumption data and considering influencing factors like advertising, while processing pairs of these values to optimize stock keeping and ordering costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If detailed inventory management is implemented for large warehouses with many stock keeping units, then stock keeping accuracy is improved, but calculating power resources are excessively consumed
Solution Approach 1:
The patent segments the inventory management process by dividing stock keeping units into different categories based on their characteristics (size, consumption time, stocking needs). This segmentation allows the system to apply different management strategies and calculation intensities to different segments, thereby improving accuracy for critical items while reducing computational resources for less critical items.
Solution Approach 2:
The patent changes parameters such as forecast time spans (one day, one week, one month) and adjusts the level of detail in inventory management based on the specific characteristics of each stock keeping unit. This parameter adaptation allows the system to optimize between accuracy and computational resource usage dynamically.
2Measurement precision
If forecast values are calculated for multiple different time spans, then ordering accuracy is improved, but calculating power resources are excessively consumed
Solution Approach 1:
The patent applies partial action by calculating forecast values for multiple time spans (one day, one week, one month) but only processing and analyzing the most relevant forecasts for each stock keeping unit. The system determines which time span is most appropriate based on the item's characteristics, thereby achieving good ordering accuracy without the excessive computational cost of fully processing all time spans for all items.
Solution Approach 2:
The system dynamically changes the forecast time span parameter based on the specific stock keeping unit being analyzed. For items with stable consumption patterns, shorter time spans may suffice, while items with variable patterns require longer time spans. This adaptive parameter selection improves ordering accuracy while controlling computational resource usage.
Data Source
AI summary
The present invention relates to a method, computer system and computer program for stock keeping. An embodiment of the invention determines forecast values of quantities to be consumed of a stock keeping unit for at least two different forecast time spans depending on a historical consumption data. The invention further determines order values of order quantities for each forecast value depending on a respective forecast value, stock keeping costs and ordering costs. The invention further determines an associated order quantity time span for each order value. The invention further determines a respective pair of forecast value and order value with the least deviation between forecast time span and order quantity time span from forecast values, forecast values respective forecast time spans, the order values and the order values respectively associated with order quantity time spans. The invention further processes the determined pair of forecast value and determined order value.