PERFORMANCE EVALUATION OF PARALLELIZED GENETIC ALGORITHMS IN SOLVING LARGE-SCALE NP-HARD OPTIMIZATION PROBLEMS

Solving NP-hard optimization problems at scale demands both accuracy and efficiency, challenging the capabilities of conventional algorithms.Parallelized Genetic Algorithms (PGAs) offer a promising approach by exploiting concurrent computing resources to accelerate evolutionary search.This paper presents an evaluative study of PGA performance on benchmark NP-hard problems, examining scalability, speedup, and convergence behavior.Empirical results show that PGAs significantly reduce computational time and improve solution quality for problems like Job Shop Scheduling and the Vehicle Routing Problem.However, trade-offs between parallelism overhead and solution stability persist.We conclude by recommending hybrid and adaptive PGAs for future high-performance optimization tasks.