Multi-Node Allocation Heuristics for IoT Applications

The rapid and increasing technological advancement highlights the importance of the Internet of Things (IoT) in the digital transformation era, establishing an intelligent network that efficiently connects devices. In IoT networks, the nodes devices have heterogeneous configurations, covering various characteristics, such as processing and storage capacity, network bandwidth, and energy consumption. This complexity of configurations makes it a significant challenge for users interested in allocating the devices according to the applications' demands. This paper presents a Multi-Node greedy heuristic (MNA-IoT) for allocating node resources to meet application demands in IoT systems. The proposed heuristic considers that the applications have multiple objectives, and it is applied to scenarios in which application requirements may exceed the capabilities of any individual node in the IoT network. This work presents scenarios and comparisons with other heuristics: a single-node allocation heuristic, gamultobj (a genetic-based algorithm), and paretosearch (direct search-based in the solutions space). Our results show that MNA-IoT can achieve the same or even better solutions than the heuristics and attains a speedup of around 1815 and 21461 compared to gamultobj and paretosearch.