Evolutionary Algorithm Population Augmentation via Segment-Based Search
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Evolutionary computation methods face challenges in convergence speed and consistency, often requiring evaluation of numerous candidate solutions before finding the optimal solution, with existing techniques varying in effectiveness across domains and lacking a unified approach to balance global and local search.
Innovation Solution
A method and system that augment a population of candidate solutions by randomly selecting segments from best solutions, generating populations based on these segments and their complements, and merging them to focus the search around the best solutions, thereby enhancing convergence speed and success rate.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a brute-force random search method is used, then the search covers a broad global search space, but the convergence speed is low because at least half of the possible candidate solutions are evaluated before convergence
Solution Approach 1:
The method performs preliminary action by identifying and preserving the best solution found so far during the evolutionary process. Before generating new candidate solutions, the system determines the best solution from the current population and uses it as a foundation for creating augmented populations, ensuring that valuable search information is retained and building upon previous successful explorations rather than starting from scratch each generation.
Solution Approach 2:
The method applies local quality by creating augmented populations with different focal points. Instead of treating all candidate solutions uniformly, the system generates multiple populations where each population is centered around a different best solution found in previous generations. This allows intensive local search around promising regions while maintaining diversity through multiple focal points, improving convergence speed without sacrificing global exploration capability.
2Loss of time
If a narrow directed local search is used, then the search is focused, but it lacks consistency and leads to premature convergence that provides sub-optimal solutions
Solution Approach 1:
The method merges multiple populations together, combining the strengths of different local searches. Each population focuses on a different best solution from previous generations, and by merging them, the system maintains diversity and avoids premature convergence to sub-optimal solutions. The merged population preserves the best solutions from all constituent populations, ensuring that the most promising candidates are retained while continuing the search.
Solution Approach 2:
The method applies dynamics by making the focal points of search populations dynamic rather than static. The best solutions used to center the augmented populations are determined dynamically from the current state of the evolution process. As the search progresses and better solutions are found, the focal points automatically update, allowing the search to adapt and follow the evolving landscape toward the global optimum rather than being trapped in local optima.
3Loss of time
If existing techniques combining global and local search are used, then some balance is achieved, but each technique has dependencies or assumptions about the fitness landscape that limit their applicability across different domains
Solution Approach 1:
The method achieves universality by designing an augmentation approach that works with any evolutionary computation algorithm regardless of the specific problem domain or fitness landscape characteristics. The technique of preserving best solutions and creating augmented populations around them is domain-agnostic and can be applied to any optimization problem. The method does not rely on specific assumptions about the fitness landscape structure, making it broadly applicable across different domains while maintaining effectiveness.
Data Source
AI summary
The present invention provides a system and a method for augmenting population of candidate solutions with respect to segments of one or more best solution(s) of the population to improve evolutionary computing. In operation, a segment is randomly selected from one of the one or more best solutions of a population. Further, a first population of solutions is generated with respect to randomly selected segment. Furthermore, a second population of candidate solutions is generated with respect to complement of the randomly selected segment. Yet further, the steps of randomly selecting a segment, and generating a first and a second population of solutions is repeated for other of the one or more best solutions of the population. Yet further, the first and the second population of candidate solutions generated with respect to respective one or more best solutions of the population are merged with the population to generate an augmented population.