Automatic Program Generation via Model-Based Node Synthesis

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing automatic program generation systems, such as those using genetic programming, face challenges with combinatorial explosions when dealing with complex problems, requiring users to manually set basic nodes and parameters, which can lead to improper settings and inefficient processing.

Innovation Solution

An automatic program generation system that analyzes existing programs to generate a program model, then uses this model to automatically create basic nodes and constraints for evolutionary computation, eliminating the need for user-defined settings and reducing the likelihood of combinatorial explosions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If manual design of basic nodes and parameters is used for evolutionary computation, then the system can handle simple problems, but it causes combinatorial explosion for complicated problems

Engineering Contradiction:

Improveability to handle complicated problemsVSAvoidcombinatorial explosion of processing

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The system automatically generates basic nodes and parameters by analyzing existing programs, eliminating the need for manual user design. The program analysis unit extracts processing structures from stored programs, and the basic node/constraint generation unit automatically creates the necessary components for evolutionary computation, allowing the system to self-configure for different problem complexities

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system performs preliminary analysis of existing programs to generate a program model before evolutionary computation begins. By pre-extracting processing structures and automatically generating basic nodes and constraints based on this model, the system prepares an optimized search space in advance, preventing combinatorial explosion during the actual problem-solving process

Inventive Principle:

Principle #10Preliminary action

2Ease of operation

If user designs basic nodes and parameters in advance, then the system can operate with simple configuration, but it leads to improper settings for complicated problems

Engineering Contradiction:

Improveease of system configurationVSAvoidproperness of basic node and parameter settings

Core Design Contradiction:

Ease of operationVSReliability

Solution Approach 1:

The system automatically analyzes existing programs to generate appropriate basic nodes and parameters, replacing manual user design with automated program analysis. The program analysis unit extracts actual processing structures from stored programs, ensuring that the generated basic nodes and parameters are properly configured for the specific problem domain without requiring user expertise

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The system uses feedback from analyzing existing successful programs to inform the generation of basic nodes and parameters. By examining how existing programs are structured and what processing patterns work effectively, the system learns appropriate configurations and applies this knowledge to generate reliable basic nodes and parameters for new problems

Inventive Principle:

Principle #23Feedback

3Adaptability or versatility

If random transformations are used in genetic programming, then the system can explore solution space, but it causes enormous combinations without proper basic nodes

Engineering Contradiction:

Improveexploration capability of solution spaceVSAvoidconvergence speed of solution

Core Design Contradiction:

Adaptability or versatilityVSProductivity

Solution Approach 1:

The system performs preliminary program analysis to generate an appropriate program model and basic nodes before evolutionary computation begins. This pre-processing step structures the solution space in advance, so that random transformations during evolutionary computation operate within a constrained, meaningful framework rather than exploring an enormous unstructured space, thereby improving convergence speed

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system dynamically determines basic nodes and parameters based on the specific problem and existing programs rather than using fixed random transformations. By changing the parameters of evolutionary computation based on problem characteristics extracted from existing programs, the system maintains exploration capability while reducing the effective search space to improve productivity

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS10466972B2Automatic program generation system and automatic program generation method

Publication Date: 2019.11.05 HITACHI LTD

AI summary

An automatic program generation system that includes: an input unit that receives inputs of input data, target data, and design requirements for a first program to be generated; a program storage unit that stores a plurality of existing second programs; a program generation device that generates the first program; and an output unit that outputs the first program. The program generation device includes a program analysis unit that analyzes the plurality of second programs to generate a program model, a basic node/constraint generation unit that generates basic nodes and constraints for evolutionary computation based on the generated program model and the design requirements input from the input unit, and an optimization unit that generates the first program by the evolutionary computation based on the basic nodes and the constraints and the input data and the target data input from the input unit.