A novel improved optimization algorithm for energy-aware service placement in industrial Fog-IoT systems

As the Internet of Things (IoT) continues to expand, industrial fog computing has become a critical method for handling distributed computations. Efficiently allocating resources to application tasks or requested services to meet requirements, such as minimal completion time, energy consumption, and controlling carbon dioxide emissions, is crucial in these systems. This service placement is recognized as a non-deterministic polynomial-time hard (NP-hard) problem. Various optimization algorithms have been proposed for service placement in fog computing, but their performance often falls short, prompting ongoing research into more effective algorithms. In this study, we introduce an enhanced algorithm,the Improved Satin Bowerbird Optimization (SBO) algorithm, to tackle the service allocation challenge in fog computing environments. Additionally, it incorporates modified exploitation strategies from the Harris Hawks Optimization (HHO) to enhance search efficiency. The proposed algorithm was first evaluated on standard classical benchmark functions (unimodal and multimodal) and then on the CEC2017 benchmark suite to validate its exploration and exploitation capabilities. Finally, when evaluated on the service placement problem, the proposed algorithm demonstrated average improvements of 7% in makespan, 5% in energy consumption, and 17% in carbon dioxide emission rate compared to other algorithms.