Software Testing via Noise Resonance to Escape Local Optima
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional automatic test case generation methods based on evolving methods often fall into local optima, leading to a small sample size of effective faults and inefficient software testing, as they excessively pursue software space coverage and emphasize mutation and crossover, resulting in test sequences unrelated to the fault phenomenon.
Innovation Solution
The software accelerated testing technology based on noise resonance integrates random resonance theory, where noise perturbations are added during test case generation to trigger noise resonance, amplifying noise and increasing the recurrence probability of software faults, using saturation and strategy noise interference modes to generate effective test sequences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If conventional evolving method is used to generate test cases, then software space coverage is improved, but testing efficiency deteriorates due to falling into local optima and generating unrelated test sequences
Solution Approach 1:
The patent applies noise resonance theory by introducing controlled noise perturbations to the test case evolution system. The noise acts as a vibration mechanism that helps the system escape from local optima by adding stochastic energy, allowing the evolutionary algorithm to explore beyond stagnant regions and discover more effective test sequences that actually trigger faults.
Solution Approach 2:
The patent dynamically adjusts noise intensity parameters during the test case generation process. By changing the noise level parameter, the system can transition between exploration phases (higher noise) and exploitation phases (lower noise), optimizing both coverage and fault detection efficiency at different stages of testing.
2Adaptability or versatility
If mutation and crossover operations are emphasized to highlight diversity, then local convergence is avoided, but sample size of effective faults deteriorates due to generating test sequences far from fault phenomenon
Solution Approach 1:
The patent implements feedback mechanisms where the effectiveness of generated test cases is evaluated based on fault triggering capability. This feedback guides the noise resonance process by providing information about which regions of the search space are productive, allowing the system to maintain diversity while concentrating efforts on areas that actually reveal faults.
Solution Approach 2:
The patent replaces the purely mechanical evolutionary operations (mutation and crossover) with a noise-resonance-based approach. Instead of relying solely on deterministic genetic operations, the system uses stochastic noise perturbations that are more effective at discovering fault conditions, substituting the mechanical evolution process with a resonance-driven search mechanism.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach accelerates the recurrence of software faults, increasing testing efficiency and reliability by generating more test cases related to the fault phenomenon, thus improving software credibility and usability.
Implementation Method 1
the random resonance theory can be described as that: the input noise intensity is increased, and meanwhile the input signal intensity of the non-linear system is kept constant; when the noise intensity is at a specified level, the output signal-to-noise ratio of the system will increase at the output terminal of the non-linear system, and a single-peak (or multi-peak) resonance curve known by people in mechanics is generated
Data Source
AI summary
A construction method of a software accelerated testing technology based on a noise resonance includes steps of: (1), collecting a source code of software to be tested; (2), automatically generating software test cases based on an intelligent evolving method, until information same or similar as a fault phenomenon is searched; (3), starting a saturation noise mode, and meanwhile exerting all controllable noise sources at a maximum intensity; (4), exerting all the controllable noise sources at a minimum intensity; (5), starting a strategy noise interference mode, and generating a noise resonance; (6), continuously exerting noises at a noise intensity able to generate the noise resonance; and (7), generating effective test sequences. Through the above steps, construction of the software accelerated testing technology based on the noise resonance is completed, which helps software testers accelerate recurrence of a software fault during testing and update the software to increase a reliability thereof.