Demand Prioritization via Discrete Ranking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current manufacturing planning systems face challenges in accurately prioritizing demands due to limited mathematical accuracy and difficulty in converting customer understanding into mathematical weights, leading to a gap between user requirements and optimizer capabilities.
Innovation Solution
The technique involves receiving demands with specified parameters, ranking them based on business logic, and breaking them into sets to optimize priority fulfillment, using an objective function to minimize unfulfilled demands while ensuring earlier ranked demands are fulfilled before later ones.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If mathematical weights are used to prioritize demands in the objective function, then the optimizer can mathematically solve the planning problem, but the limited mathematical accuracy cannot handle a wide range in weights and the conversion from customer understanding to mathematical weights is difficult
Solution Approach 1:
The patent transforms the prioritization parameter from continuous mathematical weights to discrete demand ranks. Instead of using weights that require high precision (e.g., 1.0, 1.001, 1.002), the system uses integer ranks (1, 2, 3, etc.) that are naturally discrete and easier for mathematical solvers to handle accurately. This parameter transformation resolves the precision issue while maintaining automated prioritization.
Solution Approach 2:
The patent introduces demand ranks as an intermediary concept between customer business understanding and mathematical optimization. Rather than directly converting customer priorities into weights, the system first translates them into discrete ranks, which then serve as the basis for the objective function. This intermediary layer simplifies the translation process and improves mathematical handling.
2Productivity
If large weights are assigned to high priority demands in the objective function, then those demands are prioritized, but the mathematical solver's limited accuracy cannot handle the wide range in weights
Solution Approach 1:
The patent performs preliminary ranking of demands before the optimization process. By pre-establishing the priority order through discrete ranks, the system eliminates the need for the mathematical solver to handle wide weight ranges during optimization. The ranks are assigned beforehand based on business logic, and the solver only needs to work with these pre-defined discrete values.
3Extent of automation
If weights are used in the objective function, then the solver can minimize cost, but weights are difficult for non-mathematically oriented users to understand and manipulate
Solution Approach 1:
The patent creates a simplified copy of the prioritization concept that aligns with natural human understanding. Instead of using abstract mathematical weights, the system uses demand ranks that mirror the intuitive notion of priority (1st priority, 2nd priority, etc.). This copied concept is then mapped to the mathematical formulation, making it accessible to non-mathematical users while maintaining optimization capability.
Data Source
AI summary
Embodiments presented herein provide techniques for generating and optimizing a plan in a manufacturing environment. The techniques begins by receiving a plurality of demands for a plan, wherein each demand of the plurality of demands has parameters specifying a set of operations, a due date, user specified business logic and priority. The demands are ranked based on the parameters and the user specified business logic. The plurality of demands is broken into sets of demands based on the a predefined number and the demand rank. The demands in a first set of demands are optimized to generate a strategy for fulfilling the demands in the first set of demands. One or more constraints are applied to the first set of demands to ensure the first set of demands is fulfilled in preference to the remaining sets of demands.