Multi-Objective Evolutionary Algorithm Archive Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large-scale multi-objective optimization systems often fail to provide solutions due to 'archive explosion' and resource inefficiencies when dealing with a large number of objectives, making it difficult to manage solution archives and perform optimizations effectively.
Innovation Solution
The implementation of multi-objective evolutionary algorithms (MOEAs) with objective refinement mechanisms and archive management techniques such as n-arity non-domination sorting, gene hinting, and fine-grain mutations to partition objectives into subsets, perform sequential optimizations, and prevent archive explosion, allowing for more tractable and resource-efficient solutions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a large number of objectives are optimized simultaneously using traditional multi-objective evolutionary algorithms, then the optimization can handle complex real-world problems with multiple criteria, but the solution archive explodes and computational resources become insufficient
Solution Approach 1:
The patent divides the complete set of objectives into multiple subsets and performs sequential optimizations where each subset is optimized independently before moving to the next subset. This segmentation prevents archive explosion by limiting the number of objectives considered simultaneously, while still achieving comprehensive optimization across all objectives through progressive addition.
2Productivity
If traditional multi-objective evolutionary algorithms are used with many objectives, then comprehensive optimization can be attempted, but computational resources and time become insufficient
Solution Approach 1:
The patent performs preliminary optimizations on subsets of objectives before introducing additional objectives. By completing optimizations for smaller subsets first and using their results as starting points for subsequent optimizations, the system reduces the computational burden of handling all objectives simultaneously and accelerates the overall optimization process.
3Reliability
If heuristically solving problems with a large number of objectives, then comprehensive solution search can be performed, but the heuristic methods become burdened with a large number of archived and intermediate solutions
Solution Approach 1:
The patent segments the optimization process into multiple stages, each handling a subset of objectives. This segmentation naturally limits the number of archived solutions at any given time, as each stage only archives solutions relevant to its specific subset. The cumulative effect of these staged archiving produces a manageable and high-quality solution set without the explosion that would occur in simultaneous multi-objective optimization.
Data Source
AI summary
Systems and methods are provided for performing multi-objective optimizations with a relatively large number of objectives to which optimization is to be performed. The objectives of the optimization problem may be partitioned to two or more subsets (e.g., overlapping or non-overlapping subsets) of objectives, and partial optimization(s) may be performed using a subset or combination of subsets of the objectives. One or more of the partial optimizations may use one or more pareto-optimized chromosomes from a prior partial optimization. A final full optimization may be performed according to all of the objectives of the optimization problem and may use one or more chromosomes of any preceding partial optimization as a starting point for finding a final solution to the optimization problem. Any variety of processes may be employed to mitigate archive explosion that may be associated with relatively large objective sets.