Genetic Operator Clustering for Design Structure Matrix Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing clustering methods for design structure matrices (DSMs) face challenges in accurately predicting optimal clustering arrangements for complex systems, often resulting in oversimplification, local optimal solutions, and difficulty in representing buses and three-dimensional structures.
Innovation Solution
The method employs genetic operators, such as crossover and mutation, to optimize clustering in DSMs, using a scoring metric like Minimum Description Length (MDL) to evaluate and refine cluster structures, allowing for the identification of optimal clusterings even in complex scenarios.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional clustering methods are used for DSMs, then the clustering process is simple and fast, but the clustering accuracy and ability to represent complex structures (buses, three-dimensional structures) deteriorates
Solution Approach 1:
The patent replaces traditional mechanical/heuristic clustering algorithms with a genetic algorithm-based computational approach. The genetic algorithm uses biological evolution principles (selection, crossover, mutation) to optimize clustering solutions, enabling the system to accurately represent complex structures like buses and three-dimensional configurations while maintaining computational efficiency through automated optimization processes.
Solution Approach 2:
The patent employs the Minimum Description Length (MDL) principle to dynamically adjust clustering parameters and evaluation metrics. By changing the parameter space to include MDL scoring, the system can adaptively optimize clustering accuracy for different complex structures without requiring manual tuning of multiple parameters, thus resolving the contradiction between accuracy and method complexity.
2Productivity
If clustering methods are simplified to handle large DSMs, then computational efficiency improves, but the ability to accurately represent complex systems and avoid local optimal solutions deteriorates
Solution Approach 1:
The genetic algorithm implementation is self-service in nature, automatically performing selection, crossover, and mutation operations without requiring external intervention. The system self-optimizes clustering solutions by evaluating fitness functions and iteratively improving population quality, thereby maintaining high reliability for complex system representation while scaling efficiently to large DSMs through automated parallel processing capabilities.
Solution Approach 2:
The patent incorporates feedback mechanisms through the MDL scoring function and fitness evaluation processes. The system continuously monitors clustering quality metrics and uses this feedback to guide the evolution of population solutions, ensuring that computational efficiency does not compromise reliability. The feedback loop allows the algorithm to escape local optima by evaluating overall system description length and adjusting clustering configurations accordingly.
Data Source
AI summary
Exemplary embodiments of the present invention are directed to methods and program products for optimizing clustering of a design structure matrix. An embodiment of the present invention includes the steps of using a genetic operator to achieve an optimal clustering of a design structure matrix model. Other exemplary embodiments of the invention leverage the optimal clustering by applying a genetic operator on a module-specific basis.