Genetic Programming Data Sequence Generation with Reference Table Parameter Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing techniques for generating data processing sequences using genetic programming face difficulties in efficiently generating individuals that satisfy specific criteria, particularly in adjusting parameters of image processing filters, which also applies to data processing sequences handling various data types.
Innovation Solution
A method and device for generating data processing sequences with an input layer, filter layer, and output layer, involving the sequencing of genes representing the input and filter layers, parameter allocation using a reference table, data processing, and evaluation to identify sequences that meet predetermined criteria based on similarity with target data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If parameters are added to genes expressing input layer or filter layer in genetic programming, then the ability to control and adjust parameters is improved, but the complexity of generating individuals that satisfy specific criteria increases
Solution Approach 1:
The patent segments the generation process into distinct phases: first generating the basic tree structure with genes representing input layers, filter layers, and output layers, then separately allocating parameters to filter layers through a reference table mechanism. This segmentation allows independent optimization of structure generation and parameter assignment, reducing the overall complexity.
Solution Approach 2:
The patent introduces a reference table as an intermediary component that maps genes to specific parameters of filter layers. This reference table acts as a mediator between the genetic representation and the actual parameter values, simplifying the generation process by providing a systematic lookup mechanism rather than requiring complex direct parameter assignment.
2Manufacturing precision
If multiple parameters are prepared for each type of image processing filter, then the precision of image processing is improved, but the difficulty of generating individuals with specific parameter configurations increases
Solution Approach 1:
The patent enables easy parameter changes by storing multiple parameter options in the reference table for each filter type. When generating individuals, the system can efficiently select and assign different parameter combinations from the reference table, allowing precise control over filter parameters without increasing generation difficulty. The reference table structure facilitates systematic parameter variation and optimization.
3Adaptability or versatility
If genetic programming is used to generate multiple image processing sequences, then the ability to find sequences satisfying criteria is improved, but the time required for generation and evaluation increases
Solution Approach 1:
The patent performs preliminary organization of parameters and filter types in the reference table before the actual individual generation process. By pre-structuring the available parameters and their associations with filter layers, the system reduces the computational overhead during iterative generation and evaluation, thereby reducing the overall time required while maintaining the ability to generate diverse sequences.
Data Source
AI summary
A method for generating a data processing sequence includes: (a) generating a population of one generation; (b) allocating one of an input layer, which is one node, a filter layer, which is another node, and a parameter, to each of genes forming an individual, with reference to a reference table establishing a correspondence between the gene, and the node and the parameter, and thus generating the data processing sequence corresponding to the individual; (c) comparing output data and target data to be a target associated with subject data, and calculating an evaluation value indicating a degree of similarity between the output data and the target data; and (d) specifying the data processing sequence that satisfies a criterion, using a plurality of evaluation values corresponding to a plurality of data processing sequences, respectively.