Smart building energy management with renewables and storage systems using a modified weighted mean of vectors algorithm.

With the advancement of automation technologies in household appliances, the flexibility of smart home energy management (EM) systems has increased. However, this progress has brought about a new challenge for smart homes: the EM has become more complex with the integration of multiple conventional, renewable, and energy storage systems. To address this challenge, a novel modified Weighted Mean of Vectors algorithm (MINFO) is proposed. This algorithm aims to enhance the performance of smart building EM by overcoming the limitations of conventional approaches, such as low solution accuracy and inadequacy in handling complex problems. MINFO operates on two key principles. Firstly, it employs the Elite Centroid Quasi-Oppositional Base Learning (ECQOBL) approach to improve the exploitation capabilities of conventional algorithms. Secondly, it utilizes an Adaptive Levy Flight Motion (ALFM) technique to enhance exploration. The EM problem tackled involves optimizing the scheduling of multiple energy sources, including diesel generators, PV units, and batteries, within a smart building context. Additionally, it incorporates time-of-use-based demand-side response (DSR) to manage shiftable loads, thereby reducing electricity costs and peak-to-average ratio (PAR) simultaneously and independently. The effectiveness of MINFO is demonstrated through comprehensive evaluations, comparing its performance with other optimization methods across 33 benchmark functions from basic and CEC-2019 test suites. Results indicate that MINFO significantly improves smart building EM, achieving a reduction of 53.20% in electricity costs (cost only), 53.19% in PAR (PAR only), and 50.84% in combined cost and PAR compared to the base case. These findings underscore the robustness of MINFO as an optimizer for smart building energy management.