Smart Service Placement Model for IoT Applications in Fog Computing Using an Improved Imperialist Competitive Algorithm Service Placement (IICASP)

With the rapid growth of the Internet of Things (IoT), traditional infrastructures have evolved into intelligent systems that generate large volumes of dynamic and heterogeneous data, significantly increasing the need for efficient processing, analysis, and storage. Transmitting such data to traditional cloud data centers has led to challenges such as network congestion, increased latency, and reduced Quality of Service (QoS). In response, fog computing has emerged as a novel distributed architecture that brings computational resources closer to the network edge and end‐users. One of the key challenges in this domain is the optimal placement of services within the fog environment in a way that optimizes execution time, QoS, and reliability. In this study, we propose a smart improved version of the imperialist competitive algorithm (ICA), named improved imperialist competitive algorithm service placement (IICASP), which leverages adaptive search mechanisms, dynamic allocation strategies, and revolution rate control based on problem‐specific conditions to optimize the service placement process in fog computing environments. The performance of this algorithm is evaluated against various versions of ICA, hybrid algorithms such as genetic algorithm (GA)–PSO, and other metaheuristic approaches. Evaluations conducted across scenarios with varying numbers of tasks and fog nodes demonstrate that IICASP significantly outperforms other methods in most performance indicators. Experimental results show that compared with the standard ICA, IICASP reduces total execution time by 35.2% and decreases the number of failed services by 50%. Additionally, compared with the GA, it achieves a 46.6% improvement in execution time and a 66.6% reduction in the service failure rate. The proposed algorithm not only executes the highest number of successful services but also proves robust in maintaining performance stability under high‐load conditions. Finally, despite certain limitations—such as the assumption of static resources and the omission of task migration—the paper proposes future research directions aimed at enhancing intelligence and adaptability in resource allocation for fog computing environments.