SHIFT: Sigmoid-Based Heuristic Invertible Fitness-Landscape Transformation for Accelerating SBST
Jeongjin Han, Seunghoon Sim, Jian Lee, Seongyoon Park
TLDR
SHIFT uses sigmoid-based landscape transformation to compress dense regions, accelerating Search-Based Software Testing by improving convergence and efficiency.
Key contributions
- SHIFT compresses local fitness landscapes to overcome deceptive optima and plateaus in Search-Based Software Testing.
- Facilitates escape from stagnant search regions and dense clusters without altering global landscape semantics.
- Preserves mapping invertibility, allowing optimization algorithms to traverse more effectively with the same step size.
- Demonstrates consistent improvements in convergence speed and search efficiency for SBST over established baselines.
Why it matters
This paper introduces a lightweight yet effective method to significantly accelerate Search-Based Software Testing. By transforming fitness landscapes, SHIFT helps overcome common bottlenecks like local optima and plateaus, leading to more reliable and faster test input generation. This is crucial for automating software quality assurance.
Original Abstract
Search-Based Software Testing (SBST) automates test input generation but is frequently hindered by challenging fitness landscapes characterized by numerous deceptive local optima that impede search progress, as well as extended plateaus where informative fitness signals are scarce. To address this bottleneck, we propose SHIFT (Sigmoid-Based Heuristic Invertible Fitness-Landscape Transformation for Accelerating SBST), a method designed to compress local landscapes and facilitate escape from stagnant regions without altering global semantics. By systematically contracting dense regions where search points cluster, the approach preserves mapping invertibility while enabling optimization algorithms to traverse more effectively toward global coverage with the same step size. When evaluated against established baselines, including pure hill climbing and genetic algorithms, under a normalized experimental protocol, the proposed technique yields consistent improvements in convergence speed and search efficiency. These results demonstrate that sigmoid compression constitutes a lightweight yet effective mechanism for achieving more reliable coverage discovery in complex testing environments.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.