Plan Explainer for Root Cause Analysis of Supply Chain Goal Violations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for supply chain master planning, particularly those involving multi-objective hierarchical linear optimization, fail to accurately represent complex supply chain problems and often lead to goal violations, making it difficult to identify the root cause of these violations.
Innovation Solution
A system and method that utilizes a plan explainer to parse multi-objective hierarchical linear optimization, apply heuristics to analyze logged data, and provide interactive analysis of exceptions and goal violations, enabling the determination of root cause explanations by considering decision variables, business constraints, and objectives within the supply chain network.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If hierarchical optimization is used to solve supply chain planning problems, then the plan can be generated based on multiple conflicting objectives, but the root cause analysis becomes impossible because the ultimate solution is far removed from the root cause of goal violations
Solution Approach 1:
The patent segments the hierarchical optimization process into multiple levels (first level objectives and second level objectives) with distinct cause-effect relationships. This segmentation allows the system to trace goal violations back to specific hierarchical levels and identify root causes at appropriate levels, resolving the contradiction between handling multiple objectives and enabling root cause analysis.
Solution Approach 2:
The patent introduces an intermediary analysis layer that connects hierarchical optimization results with root cause analysis. This intermediary layer processes the relationship between optimization variables, constraints, and goal violations, enabling traceability from high-level objectives to underlying causes without sacrificing the ability to handle multiple conflicting objectives.
2Ease of manufacture
If penalty weights are assigned to multiple component objective functions to represent the plan as a single objective function, then the plan can be solved using standard optimization methods, but the solution is not accurate because it relies on the comparison of non-equivalent vectors and penalties are artificially derived
Solution Approach 1:
The patent segments the objective functions into hierarchical levels with distinct priorities and constraints. This segmentation eliminates the need for artificial penalty weights by providing a structured framework where objectives are evaluated in sequence, improving both the ease of solving and the accuracy of plan representation.
Solution Approach 2:
The patent changes the parameters of the optimization problem by introducing hierarchical levels with specific constraints and priorities. This transformation allows standard optimization methods to be applied while maintaining accurate representation of the supply chain planning problem, avoiding the pitfalls of non-equivalent vector comparisons.
Data Source
AI summary
A system and method are disclosed including a supply chain network including one or more entities, one or more items and a plan explainer. The plan explainer generates an optimum inventory level based on a root cause of a goal violation by modeling the supply chain network and one or more business objectives as a hierarchy of objective functions. The plan explainer also solves the hierarchy of objective functions, stores plan explanation data and retrieves the plan explanation data. The plan explainer further generates the root cause of the goal violation by parsing the retrieved plan explanation data and calculates the optimum inventory level based on the root cause of the goal violation. In response to the calculated optimum inventory level, at least one of the one or more entities adjusts inventory levels of the one or more items according to the optimum inventory level. Other embodiments are also disclosed.