Optimal deploying IoT services on the fog computing: A metaheuristic-based multi-objective approach
Banghua Wu, Xuebin Lv, Wameed Deyah Shamsi, Ebrahim Gholami Dizicheh
Abstract
This study deals with solving the Internet of Things (IoT) Service Placement Problem (SPP) in fog computing environment using metaheuristic approaches. Basically, SPP is a non-deterministic polynomial-time hard (NP-hard) with huge discrete search spaces that is often processed by heuristic and metaheuristic approaches. We proposed an Improved Parallel Genetic Algorithm (IPGA) to solve SPP and named it IPGA-SPP. Since the genetic algorithm may get stuck in local optima, we configure it in parallel with a shared memory along with several elitist operators. IPGA-SPP considers resource distribution for load balancing and prioritizes service execution to reduce latency. Also, IPGA-SPP solves the problem as a multi-objective problem by maintaining a set of Pareto solutions by making compromises between service latency, service cost, resource utilization and service time. Although many metaheuristic approaches have been developed for SPP, but satisfying the quality of service (QoS) and simultaneously guaranteeing security to overcome the constraints of fog computing has been less considered. In this regard, we equip IPGA-SPP with a two-way trust management mechanism so that clients and service providers can verify each other's trustworthiness. Therefore, the proposed scheme is a latency-aware, cost-aware and trust-aware approach to improve the deployment process in fog computing. Through simulation on a synthetic fog environment, IPGA-SPP has shown an average of 8.4% better performance compared to state-of-the-art methods such as CSA-FSPP, GA-PSO, EGA and WOA-FSP.