Inventory Planner Using Markov Chain Heuristics for Supply Chain Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electronic commerce systems face challenges in quickly determining an optimal inventory plan for a supply chain network, which is essential for minimizing human involvement and optimizing inventory management.
Innovation Solution
The system employs an inventory planner that uses a discrete demand distribution and a target service level to determine an inventory policy. This inventory planner incorporates a Markov Chain-based heuristic to evaluate and improve inventory policies, enabling efficient calculation and reducing computational time and memory usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If traditional inventory planning methods are used, then accuracy of inventory policy determination is maintained, but computational time and processing speed deteriorate
Solution Approach 1:
The patent segments the inventory planning process into distinct phases: generating candidate inventory policies, evaluating their performance metrics (service levels, costs), and selecting optimal policies. This segmentation allows parallel processing and optimization of each phase independently, significantly improving computational speed while maintaining accuracy.
Solution Approach 2:
The system dynamically adjusts parameters such as service level thresholds, cost weights, and demand distribution assumptions to balance computational efficiency with policy accuracy. By changing these parameters based on problem complexity and available computing resources, the system achieves fast results without sacrificing determination accuracy.
2Measurement precision
If detailed evaluation of inventory policies is performed, then accuracy of cost and service level calculation is improved, but computational complexity increases
Solution Approach 1:
The patent implements partial evaluation by assessing only the most critical performance metrics (service level, total cost) for each inventory policy candidate rather than进行全面 analysis of all possible outcomes. This partial action approach maintains sufficient measurement precision for decision-making while dramatically reducing computational complexity.
Solution Approach 2:
The system generates and evaluates multiple candidate inventory policies as disposable approximations rather than pursuing a single optimal solution through exhaustive analysis. Each candidate is quickly evaluated and discarded, allowing the system to find sufficiently accurate solutions through multiple inexpensive trials rather than one complex computation.
3Productivity
If inventory planning for a large number of items is performed, then comprehensiveness of inventory management is improved, but processing time increases
Solution Approach 1:
The patent segments the large-scale inventory planning problem by processing items in batches or groups, evaluating inventory policies for multiple items simultaneously where possible. This segmentation enables parallel computation across item groups, improving overall productivity while managing total processing time through efficient resource utilization.
Solution Approach 2:
The system implements universal inventory planning algorithms that can handle diverse item types, demand patterns, and constraint conditions through a single integrated framework. This multi-functionality allows the system to process large numbers of different items using the same core engine, improving throughput without requiring separate processing for each item category.
Data Source
AI summary
A system and method are disclosed for searching alternate inventory policies from an initial inventory policy of an inventory of one or more supply chain entities to a target inventory policy by receiving a current state of items in a supply chain network determining an initial inventory policy comprising a reorder point and a target quantity, identifying one or more provisional inventory policies calculating a cost, a fill rate, and a no-stockout probability, and transporting items among the one or more supply chain entities to restock the inventory.