Exploring Differential Evolution Algorithm with Fitness-and Position-Based Selection in Search of Best Test Pattern for Combinational ATPG
Evolutionary algorithms (EAs) have proven to be efficient in identifying the optimal test pattern that achieves the highest fault coverage within the extensive search space of digital tests for VLSI circuits. This study concentrated on creating a MATLAB-based Automatic Test Pattern Generation (ATPG) framework aimed at finding the best test pattern capable of detecting the maximum number of stuck-at faults in combinational circuits by employing Fitness-and Position-Based Selection Differential Evolution (FPS-DE). In FPS-DE, the parent selection enhances convergence and diversity. The proposed approach is assessed with ITC'99 combinational benchmark circuits and compared against the Genetic Algorithm (GA), Particle Swarm Optimization (PSO), and the widely recognized open-source ATPG tool, ATALANTA. This methodology has been applied to identify the suitable mutation strategy for each category of benchmark circuit. In summary, the experimental findings indicate that FPS-DE is capable of producing high-quality test patterns for combinational ATPG effectively.