Model-to-data traceability for system model discrepancy detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional modeling environments lack the capability to reliably identify the source of discrepancies between the results of original and modified models, especially when the modified algorithms are more sensitive to input values, leading to unexpected behavior in complex systems.
Innovation Solution
The system enables model-to-data and data-to-model traceability, allowing users to trace outputs back to specific blocks within the model and compare code generated from original and modified models, facilitating the identification of discrepancies and their sources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If users modify block configurations to create custom algorithms, then model adaptability and customization are improved, but the difficulty of detecting and measuring error sources increases
Solution Approach 1:
The system performs preliminary actions by automatically generating traceability information and executing comparison tests between original and modified models before the user needs to identify errors. This pre-computation of difference data and automatic execution of diagnostic routines reduces the difficulty of error source identification when customization is performed.
Solution Approach 2:
The system implements feedback mechanisms by automatically comparing model outputs, identifying discrepancies, and providing diagnostic information back to the user. This feedback loop includes automatic execution of tests, generation of difference reports, and highlighting of modified blocks that cause unexpected behavior, making error detection easier despite increased model customization.
2Adaptability or versatility
If complex models with numerous interconnected blocks are created, then model functionality and versatility are improved, but the difficulty of detecting and measuring error sources increases
Solution Approach 1:
The system segments the complex model analysis by automatically identifying and isolating specific blocks that contribute to errors. Instead of requiring users to manually trace through numerous interconnected blocks, the system divides the diagnostic task into manageable segments by pinpointing exact locations of discrepancies and their causes within the model's blocks.
Solution Approach 2:
The system introduces an intermediary automated diagnostic mechanism that mediates between the complex model structure and the user. This intermediary automatically executes comparison routines, traces error propagation through interconnected blocks, and presents simplified diagnostic information, reducing the complexity of error detection in large-scale models.
3Measurement precision
If automatic comparison and traceability features are implemented, then error detection capability is improved, but device complexity increases
Solution Approach 1:
The system implements self-service by automatically performing error detection, comparison, and traceability tasks without requiring complex external diagnostic tools. The modeling environment itself provides these capabilities through integrated automated routines that execute when models are modified or compared, improving error detection while avoiding the need for additional complex external systems.
Data Source
AI summary
A device, method and tangible computer-readable medium are provided for detecting output discrepancies between representations of a block in two system models. For example, a first representation of a block may represent a default configuration and may execute in a first model. A second representation of the block may represent a user-modified configuration for the block and may execute in a second model. The user may execute the first and second models and may compare results using an exemplary embodiment. The embodiment may allow the user to define criteria and weightings for the criteria and to use the criteria for generating objective functions and constraints. The objective functions and constraints may be used to evaluate the performance of the two models. The embodiment may further perform trace back operations with respect to a model to determine a location in the model that produces an output discrepancy.