The invention relates to the technical field of berth-
quay crane scheduling optimization, in particular to a berth-
quay crane scheduling optimization method. Comprising the steps of obtaining the number of ships and the number of quay cranes, setting
algorithm parameters, initializing a
population, calculating the three-target fitness of each individual in the
population, recording a
global optimal solution, entering a main loop to iteratively update the position of the individual, and updating the position based on the diving
predation behavior of the turnip and the symbiotic relationship of the simulated turnip and the Eurasian
otter. Quantifying action effects through a reward and punishment function, updating a Q table by adopting a
Q learning algorithm, storing long-term reward expectation of state-action, selecting an optimal action in a current state according to the Q table by adopting an epsilon-greedy strategy, and attenuating an exploration rate according to an index; and screening a
Pareto optimal solution in combination with non-dominated sorting and
crowding distance, and outputting an
optimal scheduling scheme after the maximum number of iterations is reached. According to the method, the berth
quay crane scheduling scheme can be optimized by combining the improved kirschner
algorithm with the quay crane
failure rate, and the reliability of quay
crane equipment is improved.