Hybrid Energy Optimization Algorithms Based on Energy Consumption Prediction in IoT Environment

Abstract

Intelligent optimized energy management and prediction model in residential buildings received attraction of the researchers in last couple of years. Various techniques and models have been proposed in the literature for optimized energy management and prediction, but the trade-off between occupant comfort index and energy consumption is still a great challenge to the research community. Previously we have proposed power consumption optimization and prediction models based on particle swarm optimization (PSO) and genetic algorithm (GA). Our proposed models accomplished good performance results up to some extent, but still there is room for more improvements. In this thesis we proposed hybrid optimization of energy management and power control models based on preprocessing mechanisms for occupants comfort index, energy saving and energy consumption. The focus of our proposed hybrid optimized and prediction models is to increase occupant's comfort index and reduce energy consumption using hybrid optimization, power prediction and preprocessing of power consumption data. The proposed single and multi-preprocessing hybrid optimization based power control models provides energy efficient environment by reducing power consumption and improving occupant's comfort index as compared to GA and PSO based power prediction models. Our proposed hybrid energy optimization based prediction models are simple and maintains better user's comfort index and minimized the energy consumption without compromising the occupants comfort index. User set parameters plays a vital role in deciding the occupants comfort index. In [23, 26-30], user is not involved to determine the occupants comfort index, while our proposed models consider user set parameters to decide the occupants comfort index. So our proposed models are user friendly. In [23, 31, 32], the energy efficiency is not addressed, while our proposed models gives attention to energy savings and our models are energy efficient by reducing energy consumption. In [29, 30], the occupants comfort index is not considered while our proposed approach addressed occupants comfort index. So the bottom line is, our proposed hybrid energy optimized models based on prediction and preprocessing addressed energy efficiency, occupants comfort index and user set parameters, while other approaches mentioned above either provides energy efficiency or occupants comfort index without considering user set parameters.