System for determining operation condition of precooling operation/preheating operation of air conditioner

A precooling operation/preheating operation control apparatus comprising:

• an air conditioner that performs heat exchange between air in a target space for precooling or preheating and a heat medium;
• a ventilation apparatus that replaces part of the air in the target space with outside air that is air outside a building;
• a controller configured to:
• control the air conditioner;
• control the ventilation apparatus;
• acquire a time at which a user is expected to enter the target space, on the basis of statistical results of output of a human detecting sensor;
• and
• determine, on the basis of a set temperature of the target space at a designated time, a quantity related to an outside air temperature, and a time difference between the designated time and the time at which the user is expected to enter the target space, operation details of the air conditioner and operation details of the ventilation apparatus during a precooling operation or preheating operation,
• wherein the controller determines, by machine learning, the operation details of the air conditioner and the operation details of the ventilation apparatus during the precooling operation or preheating operation,
• wherein the machine learning is reinforcement learning,
• wherein the controller determines, in the reinforcement learning, the operation details of the air conditioner and the operation details of the ventilation apparatus, on the basis of a value function based on a state variable and a reward,
• wherein the state variable includes an outside air temperature, and
• wherein the reward is determined on the basis of
• an error between the room temperature and the set temperature, and
• an electric energy required for the precooling operation or preheating operation.

A precooling operation/preheating operation control apparatus comprising:

• an air conditioner that performs heat exchange between air in a target space for precooling or preheating and a heat medium;
• a light blocking apparatus that blocks light incident on the target space from an outside of the building;
• a controller configured to:
• control the air conditioner;
• control the light blocking apparatus;
• acquire a time at which a user is expected to enter the target space, on the basis of statistical results of output of a human detecting sensor;
• and
• determine, on the basis of a set temperature of the target space at a designated time, a quantity related to an outside air temperature, and a time difference between the designated time and the time at which the user is expected to enter the target space, operation details of the air conditioner and operation details of the light blocking apparatus during a precooling operation or preheating operation,
• wherein the controller determines, by machine learning, the operation details of the air conditioner and the operation details of the light blocking apparatus during the precooling operation or preheating operation,
• wherein the machine learning is reinforcement learning,
• wherein the controller determines, in the reinforcement learning, the operation details of the air conditioner and the operation details of the light blocking apparatus, on the basis of a value function based on a state variable and a reward,
• wherein the state variable includes an outside air temperature, and
• wherein the reward is determined on the basis of
• an error between the room temperature and the set temperature, and
• an electric energy required for the precooling operation or preheating operation.