Problem to be solved: to support reward design in planning a field maintenance plan by reinforcement learning.Reinforcement learning support deviceField information which contains information specifying the field and information about the assets located in the field is stored.Divide fields into multiple regionsRegion aggregation information is generated by aggregating information about the asset by using the regionBased on the information specifying the region and the area aggregation informationCreate abstraction field information that is an information about an abstraction field obtained by abstracting a fieldBased on the abstraction field information, the state distribution information is generated to indicate the state distribution of the assets in the field.Diagram
展开▼
