Occupancy modelling using Markov Chain

In recent studies, the energy consumption of buildings takes up a staggering 40% of the total energy consumption of the world, of which half of this energy is used by the heating, ventilation and air conditioning (HVAC). Thus, it would be beneficial for us to learn of the occupancy of a building as it is a significant factor affecting the energy consumption of buildings. However, it is hard to obtain accurate estimation of occupancy in a building due to its random nature. As of recent, there has been an increasingly high interest in modeling occupancy, especially in buildings and many methods have been utilized to forecast occupancy in both single and multiple zone scenarios. Besides Markov chain, several other techniques have also been employed with the aim to improve occupancy modelling, mainly with the aim to improve energy usage in buildings. The techniques employed include random sampling, machine learning, logistic regression, decision tree and agent-based techniques. The aim of this report is to compare the existing homogeneous and inhomogeneous Markov chain models that focuses primarily on occupancy modelling prediction, and to provide insights on the multiple advantages and disadvantages of these different techniques. After evaluation, improvements to the prevailing issues would be recommended optimizing future research for the current methodology of existing Markov chain models.Bachelor of Engineering (Electrical and Electronic Engineering