TASK-ORIENTED SERVICE PERSONALIZING METHOD USING REINFORCEMENT LEARNING IN SMART ENVIRONMENT

摘要

According to the present invention, as a method for automatically constructing an environment using a system including a smart object and a server controlling the smart object, provided is a service personalizing method using reinforcement learning in a smart environment which includes: a recording step of recording the environment and usage history of a user from the smart object by the server; a learning step of learning the preference of the user based on the environment record and usage history recorded by the recording step by the server; and an environment construction step of constructing a preference environment of the smart object by the server based on the preference learned through the learning step. Accordingly, the present invention can learn the task preference of the user through smart devices installed in the smart environment and automatically construct a smart environment.;COPYRIGHT KIPO 2018