Multi-objective Evolutionary Algorithm Optimization for Pareto Solution Diversity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Multi-objective evolutionary algorithms (MOEAs) face challenges in achieving diversified and converged Pareto optimal solutions in engineering design optimization, as existing methods often result in poor spread and uniformity of solutions, and require multiple simulations to achieve diverse trade-offs.
Innovation Solution
Conducting multiple MOEA-based engineering optimizations independently with varying initial generations and evolutionary algorithms, and combining the resulting Pareto optimal solutions to enhance convergence and diversity, using techniques like Nondominated Sorting Genetic Algorithm (NSGA-II) and strength Pareto evolutionary algorithm (SPEA), and evaluating design objective functions with computer-aided engineering analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If multiple MOEA-based optimizations are conducted independently with varying initial generations and evolutionary algorithms, then the diversity and convergence of Pareto optimal solutions are improved, but the computational time and complexity of the optimization process increase
Solution Approach 1:
The optimization process is divided into multiple independent MOEA-based optimizations, each with varying initial generations and evolutionary algorithms. This segmentation allows parallel execution of multiple optimization trajectories, improving solution diversity and convergence while managing computational time through structured parallel processing.
Solution Approach 2:
The Pareto optimal solutions from multiple independent optimizations are merged and combined into a unified set. This merging process consolidates the results from different optimization runs, enhancing the overall quality and diversity of the solution set while avoiding redundant computations.
2Productivity
If conventional single-objective optimization methods are used, then the optimization process is simpler and faster, but the ability to handle conflicting objectives and find diverse trade-off solutions is limited
Solution Approach 1:
The patent employs multi-objective evolutionary algorithms that can handle multiple conflicting objectives simultaneously. The optimization framework is designed to accommodate various objective functions and constraints, providing universal applicability to complex engineering design problems with multiple conflicting requirements.
Solution Approach 2:
The optimization approach transitions from single-objective to multi-objective optimization, adding an additional dimension to the optimization space. This dimensional change enables the system to explore trade-off solutions in the objective function space, providing diverse solutions that balance multiple conflicting objectives.
Data Source
AI summary
Systems and methods of obtaining a set of better converged and diversified Pareto optimal solutions in an engineering design optimization of a product (e.g., automobile, cellular phone, etc.) are disclosed. According to one aspect, a plurality of MOEA based engineering optimizations of a product is conducted independently. Each of the independently conducted optimizations differs from others with parameters such as initial generation and/or evolutionary algorithm. For example, populations (design alternatives) of initial generation can be created randomly from different seed of a random or pseudo-random number generator. In another, each optimization employs a particular revolutionary algorithm including, but not limited to, Nondominated Sorting Genetic Algorithm (NSGA-II), strength Pareto evolutionary algorithm (SPEA), etc. Furthermore, each independently conducted optimization's Pareto optimal solutions are combined to create a set of better converged and diversified solutions. Combinations can be performed at one or more predefined checkpoints during evolution process of the optimization.