The invention relates to the technical field of supply chain management, in particular to a
medicine supply chain scheduling method and
system based on
reinforcement learning, and the method comprises the following steps: obtaining the
temperature difference and heat
conductivity coefficient inside and outside a
carriage, calculating the
heat penetration rate and
thermal safety time limit, planning a path, and correcting the speed based on a congestion index to obtain the predicted passing time; and comparing the
time limit with the deadline to generate an action
mask vector, inputting the environmental parameters into a neural network to output a
score, correcting by utilizing an action
mask, and generating a scheduling instruction through a
Softmax function. According to the method, the
thermal safety time limit is accurately deduced, the action
mask is constructed by using the
thermal safety time limit and the delivery deadline as double constraints, invalid nodes are eliminated, the multi-dimensional parameters are input into the neural network, the probability distribution is generated under the mask constraint, and it is ensured that the vehicle scheduling instruction follows the
temperature control safety threshold. The risk of
medicine thermal failure is effectively avoided, and the cold-chain logistics distribution accuracy is improved.