Three-stage fuzzy-metaheuristic algorithm for smart cities: Scheduling mobile charging and automatic rule tuning in WRSNs

Abstract

Recent breakthroughs in wireless power transfer technology have been promising for empowering and enabling seamless operation of wireless sensors and developing sustainable systems such as wireless rechargeable sensor networks (WRSNs). Relying on this technology and focusing on WRSNs, this paper proposes an adaptive charging scheduling algorithm called the three-stage fuzzy metaheuristic algorithm (TSFM) by using the whale optimization algorithm (WOA) and a multiobjective function designed to meet the application requirements of WRSNs in smart cities. The TSFM algorithm benefits from the WOA for the automated configuration and optimization of the fuzzy rule-base table in a three-stage fuzzy inference system for clustering, routing, and scheduling mobile charging simultaneously. Unlike the other methods, the TSFM algorithm determines the ratio of effect of fuzzy input parameters of every stage based on application and environment in addition to considering effective parameters for every stage, especially for charging scheduling (e.g., the energy of a charge-request node, the distance between a charge-request node and a mobile charging (MC), the density of a charge-request node, the average energy consumption rate of a charge-request node, and the time elapsed after a charge request). Furthermore, for the proper distribution of energy in a network, better responsiveness, and optimal use of a charging robot, two energy thresholds are defined in the TSFM algorithm, namely as the charging request issue threshold and the charging operation completion threshold. The TSFM algorithm is adaptable to applications, environments, and requirements of WRSNs not only by using the features of nodes (as the fuzzy input parameters) but also by considering the application-specific features, environments, and infrastructure (e.g., network size, number of nodes, base station (BS) positions, and designer’s goals) within the process of optimizing tunable parameters. The TSFM algorithm was evaluated and compared with some common methods such as FCFS, NJNP, and ESS in terms of functional metrics such as energy utility, charging latency, survival rate, death rate of nodes, system stability, number of packets received by BS, number of request nodes, number of alive nodes, average residual energy, adaptability analysis, and statistical validation in two scenarios and six applications. According to the simulation results, the proposed TSFM algorithm improved the performance criteria and outperformed the other methods by far in terms of application. For instance, the TSFM algorithm improved the minimum system stability rate by 167%, 0.73%, and 123% compared with FCFS, NJNP, and ESS methods, respectively.