| Keywords |
Reinforcement learning; Integrated control scheme; Indoor air quality; Ventilation; Double deep Q-network |
| Abstract |
In building industry, indoor environmental conditions can be maintained by using a simple rule based control such as setting a specific temperature or air flow rates for a HVAC system to maintain thermal comfort and IAQ.
This simple control scheme can cause inefficient building operation especially multi environmental control devices are exist in the space. Also it does not reflect the changes in occupant behavior and indoor?outdoor environmental conditions. To overcome this limitation, we suggest a new control method using a Double deep Q-network(DDQN) which utilizes a data-driven approach to find the optimal control with various indoor environmental control devices. |