Evolutionary Approaches to Fog Node Placement in LV Distribution Networks

Abstract

Smart grids merge intelligent computing technologies and electrical grid networks for better monitoring, control and management of electrical energy and facilities. The maturity of cloud computing has been the major driving factor for its adoption in smart grid deployments. Despite the elasticity of cloud resources, centrality and long distances to remote data centers cause high latency, high bandwidth consumptions and unstable connectivity, which are undesirable for IoT-based smart grid applications. Fog computing as an extension of cloud computing services to the edges of the network overcomes these challenges and perfectly suit the distributed nature of the low voltage (LV) electrical distribution networks as part of smart grid. The pressing issues with the adoption of fog computing for smart grid applications are finding the best placement plan for fog node locations in LV distribution networks to enhance monitoring and control. The main goal of this work is to present a mathematical model to address the aforementioned issues focusing on minimizing deployment cost and network delay. In addressing this multi-objective problem, a new algorithm, namely Future Search Particle Swarm Non-dominated Sorting Genetic Algorithm (FPNSGA), is proposed based on the combination of the best features of the NSGA-II, SMPSO, and a recently formed algorithm, Future Search. The effectiveness of the algorithm is evaluated based on the benchmarking technique (Weighted Sum approach), the convergence and diversification of the solutions using HV indicators and CPU time. The results from simulations show that the proposed mechanism is very competitive and outperforms other fog planning network schemes.