摘要 |
<p>A quantitative weighting system (1), wherein when the sum total of the weights of articles (S) in each of stations (2) is considered as a state, an operation of allotting the article (S) to each of the stations (2) is considered as an action, and the change of the state by the execution of the action is considered as a transition, a transition difficulty level indicating the difficulty of supply of the article (S) having a weight required for each transition is updated. In a graph structure representing a state transition model, a Q-value set for each of edges is updated such that a higher first reward is given as the transition difficulty level regarding each of the edges becomes lower, and a second reward is given if the sum total of the weights of the articles (S) regarding a node that is a transition destination and connected to each of the edges is within a target range.</p> |