Supply Network Flow Optimization Under Discrete Lot Size Constraints
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing software optimizers struggle to find an optimal solution for supply networks within a reasonable time due to constraints such as lot size restrictions, which require discrete values, leading to inefficiencies and suboptimal results.
Innovation Solution
A system that relaxes discrete constraints in mixed integer programming models to allow for more flexible optimization, incrementally reintroducing restrictions to find an optimal supply network configuration by prioritizing dependent operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If discrete constraints (lot size restrictions) are enforced in mixed integer programming models, then solution accuracy and customer requirement compliance are improved, but computation time increases significantly
Solution Approach 1:
The patent segments the constraint enforcement process into sequential stages. Instead of enforcing all discrete constraints simultaneously, the system progressively adds constraints in batches, allowing the optimizer to handle smaller problem sizes at each stage. This segmentation of the constraint application process reduces overall computation time while maintaining solution accuracy.
Solution Approach 2:
The patent applies preliminary relaxation of constraints before optimization. By initially removing or loosening discrete constraints to obtain a baseline solution, the system then progressively re-imposes constraints. This preliminary action allows the optimizer to quickly establish a feasible region before dealing with the complexity of discrete requirements.
2Reliability
If all constraints are enforced simultaneously in the optimization model, then solution optimality is improved, but the solver runs out of time
Solution Approach 1:
The patent segments constraint enforcement into sequential stages rather than applying all constraints simultaneously. The system divides the constraint set and applies them progressively, allowing the solver to maintain reasonable execution time at each stage while working toward the optimal solution that satisfies all constraints.
Solution Approach 2:
The patent dynamically adjusts the constraint set during the optimization process. Constraints are added or modified based on the solution progress and computational time remaining. This dynamic approach allows the system to adapt the optimization problem during execution, balancing solution optimality with time constraints.
3Reliability
If the optimization process is extended to find more options, then solution quality improves, but processing time increases
Solution Approach 1:
The patent applies partial constraint enforcement in the early stages of optimization, allowing the solver to quickly generate preliminary solutions with fewer constraints. Then, constraints are progressively added to refine the solution. This partial action approach enables the system to achieve acceptable solution quality faster than if all constraints were applied from the beginning.
Data Source
AI summary
The example embodiments are directed to methods and systems for generating a model of an optimum flow of materials within a supply network based on a relax mixed integer programming model. For example, the method may include generating, via a software application, a model of the supply network based on the data stored in the storage device, executing a MIP optimizer on the model to generate a configuration of a flow of materials within the supply network, wherein the MIP optimizer is instructed to relax discrete constraints within the supply network when generating the configuration, generating a graphical user interface that includes a plurality of graphical objects corresponding to a plurality of locations included in the configuration and directional edges between the plurality of graphical objects corresponding to the flow of the materials between the plurality of locations, and displaying the graphical user interface.