Hybrid genetic algorithm and Q-learning-based solution for the time-variant berth and quay crane allocation problem

Introduction This study addresses the joint scheduling optimization of continuous berths and quay cranes by proposing a time-variant quay crane allocation method. Methods A coordinated optimization model is constructed that considers the temporal dimension of quay crane scheduling and equipment collision factors to reduce overall port operational costs. A hybrid intelligent algorithm integrating Q-learning is innovatively designed, using a genetic algorithm as the main framework while embedding a quay crane allocation module and dynamically selecting genetic operators through Q-learning to achieve adaptive optimization of the evolutionary mechanism. Results The module with Q-learning optimization is compared to the module without Q-learning optimization, demonstrating that the Q-learning module can accelerate the convergence of the algorithm and has a better ability to find the optimal solution in large-scale cases, proving the effectiveness of the module. Discussion The results show that the proposed algorithm and CPLEX perform similarly in small-scale cases, while the solution speed and capability are better than the genetic algorithm in large-scale problems and superior to the CPLEX algorithm with time constraints in some cases, proving the effectiveness and superiority of the proposed algorithm.