Heuristic Optimization via Weighted Objective Conversion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing optimization algorithms, particularly greedy heuristics, are limited in adapting to different objectives and multiple objectives, requiring significant time and resources to reconfigure or recalculate solutions, and often fail to provide real-time results.
Innovation Solution
An intermediary system that converts user-input objectives into a single criterion using weights, allowing existing heuristics to provide real-time solutions by selecting optimal choices based on combined criteria, even when objectives conflict, and enabling iterative adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a greedy heuristic algorithm is used for optimization, then real-time solutions can be obtained, but the system cannot adapt to different objectives or handle multiple objectives
Solution Approach 1:
The patent applies dynamics by making the heuristic algorithm adaptable to different objectives through dynamic configuration. The system allows users to define custom objectives and dynamically adjust heuristic parameters without requiring complete algorithm redesign, enabling the same heuristic framework to serve multiple optimization goals efficiently
Solution Approach 2:
The patent implements universality by designing a heuristic system that can handle multiple objectives simultaneously. The framework integrates multiple objective functions and constraints into a unified optimization process, allowing a single heuristic algorithm to address diverse sourcing, manufacturing, and logistics optimization problems
2Measurement precision
If an optimization algorithm is formulated for a new objective, then optimal solutions can be obtained, but significant time and computational resources are required
Solution Approach 1:
The patent applies preliminary action by pre-configuring heuristic rules and parameters for common optimization scenarios. The system maintains a library of pre-processed heuristic strategies that can be quickly applied to new problems without requiring complete algorithm development and testing from scratch
Solution Approach 2:
The patent employs lightweight heuristic approximations that provide sufficiently good solutions without requiring computationally expensive exact optimization methods. These simplified models deliver near-optimal results in fractions of the time required by traditional optimization algorithms
3Productivity
If a heuristic is designed for a single objective, then real-time solutions are provided, but users cannot input multiple objectives or different objectives
Solution Approach 1:
The patent applies segmentation by breaking down multiple objectives into hierarchical levels and weighted components. The system divides complex multi-objective problems into manageable segments that can be processed sequentially by the heuristic algorithm, with each segment representing a specific objective or constraint
Solution Approach 2:
The patent introduces an intermediary layer that translates user-defined multiple objectives into a format compatible with single-objective heuristics. This intermediary mechanism converts complex multi-dimensional optimization requirements into weighted single-objective formulations that the heuristic can process in real-time
Data Source
AI summary
Systems and methods that comprise: capturing business objectives from a user, including a weight of each business objective; searching possible solutions in a heuristic way, enumerating all the options; evaluating each solution against the objectives being targeted at each of the product structure; selecting a result that satisfies the given objectives the most; and outputting the result with the user.