Edge-computing-enabled hybrid and multi-objective geographic routing for mesh IoT networks: An IMOGWO-based approach

Abstract

Due to their robustness and resource limitations, IoT objects pose several multi-objective optimization challenges, making routing in mesh IoT networks a critical issue. Meanwhile, meta-heuristic and multi-objective optimization approaches provide promising results. This paper proposes a novel geographic, hybrid and multi-objective routing method for IoT mesh networks. Routing is formulated as an optimization problem with multiple objective functions. The proposed Improved Multi-Objective Gray Wolf Optimizer (IMOGWO) meta-heuristic is applied between communicating objects in a distributed manner to solve and optimize routing decisions. The work presents the first application and evaluation of IMOGWO to a real-world problem, specifically routing in IoT mesh networks. Combined IoT simulations (using both real and simulated nodes) are performed to evaluate the introduced approach and show its effectiveness in comparison to other existing routing methods, including MOGWO-based routing, an AcNSGA-III-based QoS routing and a BFOA-based geographic routing algorithm. Results indicate that the proposed approach enhances network performance. Particularly, IMOGWO increases the stability period by 9.30% compared to MOGWO, 20.51% compared to AcNSGA-III and 38.24% compared to BFOA. In addition, it ensures a better packet delivery ratio (92.2%). Furthermore, it maintains a lower average transmission latency (1.93s) than AcNSGA-III and BFOA. These improvements, validated through inferential statistical tests, demonstrate that IMOGWO optimizes routing for IoT mesh networks effectively.