Surrogate models for IoT task allocation optimization
Dominik Weikert, Christoph Steup, Sanaz Mostaghim · Proceedings of the Genetic and Evolutionary Computation Conference Companion · 2022
The optimization of Task Allocation is an important and ongoing problem in the field of Internet-of-Things (IoT) networks. A sometimes overlooked aspect is the computational overhead of the employed optimization algorithms, which may require expensive simulations. Such simulations may be too taxing for the limited hardware found in such networks. To alleviate this, this work proposes two distinct surrogate models suited to replace the simluation-based evaluation in a task allocation optimization. Experiments show that both models provide comparable results to a simulation-based optimization while requiring far less computational resources during the evaluation.