Applying Heuristic Algorithms for Solving the 0–1 Knapsack Problem: An Experimental Analysis

This study analyses and compares the performance of six heuristic algorithms: Genetic Algorithm (GA), Simulated Annealing (SA), Hybrid (SA+GA), Tabu Search (TS), Ant Colony Optimization (ACO), and Particle Swarm Optimization (PSO) in achieving the optimal solutions to combinatorial optimization problems represented by 0–1 Knapsack Problem (KP01). The algorithms were analyzed based on evaluation metrics like runtime, iteration time, performance, and complexity. The results indicate large differences in runtime efficiency and iteration speed. For runtime and iteration performance Tabu Search was the fastest with the lowest median runtime, followed by PSO and then GA. On the other hand, the average runtime for ACO is the longest and the highest standard deviation. The average time per iteration ranges from 0.027 ms for SA, to 70.8 ms per iteration for ACO which is very efficient. However, the hybrid method is competitive but provides a longer average time per iteration. Despite the variability, SA, GA, and PSO showed competitive results that can derive their robustness for combinatorial problems. These findings underline several key trade-offs and help to choose an optimal heuristic method that is robust and efficient for real-world decision-making applications.