A Memetic Algorithm Integrating Argument Acceptability for Stable Extension Solving in Abstract Argumentation Frameworks

In abstract argumentation frameworks, finding stable extensions is a crucial task for evaluating argument acceptability. Existing methods, such as brute-force enumeration, are computationally expensive; label-based tree search algorithms suffer from low efficiency, and pruning strategies have not effectively reduced computational costs. While heuristic and metaheuristic methods can find high-quality approximate solutions for large-scale problems, they often fall into local optima due to a lack of effective search guidance. To address these challenges, this paper proposes a memetic algorithm integrating argument acceptability for stable extension solving (ACMASE). The algorithm incorporates argument acceptability information and coordinates local search and crossover operations to guide the search direction, improving solution quality and the ability to solve stable extensions. The algorithm is tested on the ICCMA 2017, 2019, and 2023 datasets, as well as several random benchmarks generated by AFGen. Comparative experiments with algorithms such as Haywood, ArgTools, and Taas-Dredd show that ACMASE outperforms existing methods in both solving speed and success rate, demonstrating its advantages in stable extension solving.