Incremental Data Flow Validation for Local Consistency Checks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for validating synchronous data flow systems are computationally intensive and time-consuming, especially when new tasks are added, and do not efficiently identify inconsistent elements.
Innovation Solution
A method for validating data flow systems through incremental and local verification of consistency by checking a local consistency indicator for each new connection, allowing identification of inconsistent elements and reducing the need to reconsider the entire graph.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional validation methods are used to check consistency of data flow systems, then validation thoroughness is improved, but computational time and complexity increase significantly
Solution Approach 1:
The patent segments the data flow graph into smaller subgraphs and validates consistency locally within each subgraph rather than performing global validation on the entire system. This segmentation allows the validation process to be divided into manageable parts, reducing computational complexity while maintaining validation thoroughness through systematic coverage of all connections.
Solution Approach 2:
The patent performs preliminary validation checks on individual connections and subgraphs before integrating them into the complete system. By validating local consistency indicators and repetition vectors in advance, the system identifies potential issues early in the design phase, avoiding the need for time-consuming re-validation of the entire system later.
2Reliability
If traditional validation methods are used to check consistency of data flow systems, then validation completeness is improved, but device complexity increases
Solution Approach 1:
The validation system is segmented into modular components that handle specific validation tasks independently. Each connection and subgraph can be validated by dedicated validation modules that compute local consistency indicators and repetition vectors, reducing the overall system complexity while maintaining comprehensive validation coverage.
Solution Approach 2:
The validation system is designed to automatically compute consistency indicators and repetition vectors without requiring complex external validation tools. The system performs self-validation by using its own structural information to check consistency, eliminating the need for separate complex validation apparatus.
3Productivity
If incremental validation is implemented for new connections, then validation speed is improved, but measurement precision of consistency may be reduced
Solution Approach 1:
The patent applies different validation strategies to different parts of the system based on their specific characteristics. For new connections, local consistency indicators are computed with focused precision, while established subgraphs use previously validated repetition vectors. This localized approach maintains measurement precision where needed while improving overall validation speed.
Solution Approach 2:
The validation system uses feedback from local consistency checks to guide further validation actions. When a new connection is added, the system computes local consistency indicators and uses this feedback to determine whether full system re-validation is necessary or if local validation suffices, thereby maintaining precision while improving speed.
Data Source
Figure 1A~3
Figure 4
Figure 5~6
AI summary
The present invention relates to a method and device for validating a data flow system represented by a data flow graph, comprising: - an acquisition module (9) configured to acquire data related to first and second common parts of the data flow graph, and - a supervision module (11) configured to incrementally verify the coherence of the data flow system by verifying the local coherence of a common unidirectional connection connecting the first common part of the data flow graph to the second common part of the data flow graph at each increment.