In today's world where frequent traveling has become a necessity due to the fast paced business and tour needs, the hospitality industry is also experiencing a sharp rise in demand. Moreover the guest requirements for comfortable stay at a hospitality building also increased, but maintaining thermal comfort condition causes a huge rise in energy consumption for Heating, ventilation and Air Conditioning system. However, the percentage of occupied duration in a room is low due to the varying schedule and needs (business, shopping, sightseeing etc.) of individual guest. Since the comfortable stay is the top most priority so comfort temperature is maintained by keeping the HVAC running even though guest rooms are empty. However, the energy consumption can be reduced during the guest's absence by turning the HVAC off for the room, but it causes dis-comfort of the guest on coming back to the guest room. So it poses the trade-off problem to maintain the comfort for the guest and reduce the energy consumption at the same time in a hospitality building.

In this paper, we proposed an energy saving algorithm by solving the trade-off problems, i.e., reduce the energy consumption optimally while guarantee the guest comfort. Guest comfort is defined by Predicted Percentage Dissatisfied 10 with 5% deviation but it is hard to maintain the minimal comfort because thermal dynamics is influenced by diverse building/HVAC factors. To optimize energy saving with comfort guarantee, we develop two Artificial Neural Network models to predict the thermal dynamics and the energy consumption and propose an algorithm to find the optimal control parameter. Our proposed algorithm is evaluated by using both the test bed real data and the EnergyPlus model with reference climates, and results show 32%~35% energy saving rate with comfort guarantee.