Experience-Layered Gene Pool for Accurate Fitness Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data mining techniques using genetic algorithms face challenges in distinguishing between lucky and truly fit individuals, leading to optimization for individuals that are lucky rather than actually fit, due to the variability of fitness estimation over time.
Innovation Solution
Implementing an experience-layered gene pool system where individuals are stratified into multiple layers based on their testing experience, allowing competition only within their experience layer, and setting a minimum fitness threshold for entry into the elitist pool to ensure more accurate fitness evaluation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If fitness estimation is performed with limited testing experience, then the evaluation process is faster and requires fewer resources, but the accuracy of fitness evaluation deteriorates due to variability and luck
Solution Approach 1:
The gene pool is segmented into multiple experience layers (L0, L1, L2, etc.) based on the amount of testing experience individuals have accumulated. Each layer corresponds to a specific range of experience levels, creating a hierarchical structure where individuals compete within their layer rather than across all experience levels simultaneously. This segmentation allows the system to balance evaluation speed and accuracy by permitting early competition at lower experience levels while maintaining the option for continued refinement.
Solution Approach 2:
The system performs preliminary fitness evaluations at multiple experience layers before final selection. Individuals undergo initial evaluation at L0 with minimal testing, then progressively refined evaluation at L1, L2, and higher layers as they accumulate more testing experience. This preliminary action sequence allows the system to quickly filter out poor performers early while providing opportunities for promising individuals to demonstrate true fitness through continued testing.
2Adaptability or versatility
If the elitist pool accepts individuals with minimal testing experience, then the population diversity is maintained and evolution continues, but the reliability of selected individuals deteriorates due to lucky early performance
Solution Approach 1:
The elitist pool is segmented into experience layers where L0 contains individuals with minimal testing experience and L1, L2, and higher layers contain individuals with progressively more experience. This segmentation allows the system to maintain population diversity by continuously admitting new individuals to L0 while ensuring reliability by requiring individuals to prove their fitness across multiple experience layers before reaching the top elitist layers.
Solution Approach 2:
The system dynamically adjusts the experience requirements for elitist pool entry based on the individual's performance trajectory. Individuals who show promising early performance are not immediately accepted into the top elitist layers but are instead placed in lower experience layers where they must continue to demonstrate fitness as they accumulate more testing experience. This dynamic approach allows the system to balance diversity and reliability adaptively.
3Productivity
If competition is allowed across all experience levels, then the selection pressure is maximized and evolution is accelerated, but the fairness of evaluation deteriorates due to experience disparity
Solution Approach 1:
Competition is segmented into experience-based layers where individuals compete primarily against others with similar testing experience. This segmentation creates fair evaluation environments within each layer while maintaining overall evolutionary progress through the hierarchical structure. Individuals can advance to higher experience layers as they accumulate more testing, ensuring that competition remains both fair and productive.
Solution Approach 2:
The system implements periodic re-evaluation and layer transitions based on accumulated testing experience. Individuals are not permanently fixed in their competition level but can progress to higher experience layers as they undergo continued testing and demonstrate sustained fitness. This periodic action ensures that evaluation fairness is maintained while allowing evolution to continue at an accelerated pace.
Data Source
AI summary
Roughly described, a computer-implemented evolutionary data mining system includes a memory storing a candidate gene database in which each candidate individual has a respective fitness estimate; a gene pool processor which tests individuals from the candidate gene pool on training data and updates the fitness estimate associated with the individuals in dependence upon the tests; and a gene harvesting module providing for deployment selected ones of the individuals from the gene pool, wherein the gene pool processor includes a competition module which selects individuals for discarding from the gene pool in dependence upon both their updated fitness estimate and their testing experience level. Preferably the gene database has an elitist pool containing multiple experience layers, and the competition module causes individuals to compete only with other individuals in their same experience layer.