Hybrid Metaheuristic Algorithms with Dynamic Clause Weight Adjustments for Efficient Optimization of NP-Hard Problems

The efficient resolution of large-scale NP-hard optimization problems remains a significant challenge due to their in-herent complexity and vast solution spaces. This paper introduces novel hybrid metaheuristic algorithms that incorporate dynamic clause weight adjustments to enhance optimization performance. The proposed approaches combine state-of-the-art techniques like Adaptive Multi-population Differential Evolution and Hybrid Genetic Algorithm with Local Search with adaptive learning mechanisms for intelligent clause weight redistribution. Extensive experiments on benchmark problems, including MaxSAT, Vehicle Routing, Job Shop Scheduling, and Quadratic Assignment, demonstrate the superior solution quality, convergence speed, and scalability of these hybrid algorithms compared to existing methods. Statistical and theoretical analyses of the results and algorithmic properties provide further insights. The findings highlight the potential of integrating dynamic weight adjustments into hybrid metaheuristics for tackling complex optimization challenges across diverse domains.