ICST 2026
Mon 18 - Fri 22 May 2026 Daejeon, South Korea
Mon 18 May 2026 09:15 - 09:35 at Room 104 - Morning Session 1 Chair(s): Juyeon Yoon

Automated test case generation remains challenging due to diverse implementation behaviours and complex boundary conditions. This work investigates the effectiveness of Multi-agent Large Language Model (LLM) architectures employing collaborative and competitive interaction patterns, compared against a Single-agent baseline, within the context of black-box testing. Experiments on the HumanEval benchmark show that prompt engineering is the primary driver of performance, with Rule-augmented Few-shot prompting yielding improvements up to 20-30% in both coverage and Execution Success Rate (ESR) against all the other tested strategies — i.e., Zero-shot and conventional Few-shot prompting. Although Multi-agent architectures achieve superior test coverage — peaking at 99.75% — they yield an ESR comparable to Single-agent frameworks (97.31% versus 97.25%). Crucially, this comes at the expense of a threefold to fourfold increase in token expenditure. In contrast, the Single-agent configuration paired with optimised Rule-augmented Few-shot prompting provides the most effective balance between accuracy, coverage, and computational efficiency.

Mon 18 May

Displayed time zone: Seoul change

09:00 - 10:30
Morning Session 1ASTA at Room 104
Chair(s): Juyeon Yoon Korea Advanced Institute of Science and Technology
09:00
15m
Day opening
Welcome
ASTA
Tommaso Fulcini Politecnico di Torino, Juyeon Yoon Korea Advanced Institute of Science and Technology
09:15
20m
Talk
Automated Black-Box Testing: A Comparative Study of LLM Agent Architectures and Prompt Engineering
ASTA
Anna Arnaudo Politecnico di Torino, Riccardo Coppola Politecnico di Torino, Maurizio Morisio Politecnico di Torino, Flavio Giobergia Politecnico di Torino, Van-Thanh Nguyen Politecnico di Torino, Enrico Chen Politecnico di Torino, Minh-Thai Mai , Xiaoquan Ji Politecnico di Torino, Xiaoning Ma
09:35
20m
Talk
Automatic GUI testing of Android applications based on Multi-Agent Reinforcement Learning
ASTA
kazuki dodo Keio University, Shingo Takada Keio University, Japan
09:55
20m
Talk
SLS-Fuzz: Large Language Model based Self Learned Seeder for efficient Fuzzing
ASTA
Sangharatna Godboley NIT Warangal, Darshan Lohiya , Golla Monika Rani , P. Radha Krishna National Institute of Technology Warangal, Warangal