Adaptive Simulated Annealing (ASA) and Adaptive Large Neighborhood Search (ALNS) are two metaheuristic algorithms widely applied to solve discrete optimization problems. This study employs both algorithms to address the Container Loading Problem (CLP), a critical challenge in the consolidation-based freight forwarding industry, where maximizing container utilization directly influences revenue and operational efficiency. The case company, a national freight forwarding enterprise operating consolidation services in Indonesia, currently achieves an average container utilization rate of 56.8%, indicating a substantial opportunity for improvement. By formulating the CLP as a discrete combinatorial optimization model, this research aims to enhance both container load utilization and revenue through algorithmic optimization. The novelty of this work lies in its comparative implementation of ASA and ALNS under adaptive parameter calibration, applied to a real-world freight forwarding context, which remains rarely explored in previous CLP studies. Experimental results show that ALNS consistently outperforms ASA in terms of both objective value and robustness across scenarios. Specifically, the ALNS method achieves 85.4% container utilization and an average revenue increase of 8.6% per container, demonstrating superior efficiency in freight consolidation optimization. Additionally, experiments conducted under equal iteration conditions further support that ALNS maintains higher stability and better solution consistency compared to ASA, particularly in terms of fitness and utilization efficiency across different iteration scenarios. Despite ALNS requiring longer computation time, it remains well within the acceptable time frame for freight forwarding operations, where up to 24 hours is available for shipment planning. These findings provide practical implications for logistics firms seeking to integrate metaheuristic-based decision support systems to improve capacity utilization, responsiveness, and profitability.
