Evolutionary multi-objective set cover problem for task allocation in the Internet of Things

Abstract

Efficient distribution of tasks in an Internet of Things (IoT) network ensures the fulfillment for all objects to dynamically cooperate with their limited energy, processing and memory capabilities. The main contribution of this paper is threefold. Firstly, we address the task allocation in the IoT as an optimization problem with a new formulation derived from the context of set cover problem. To the best of our knowledge, no such study has been considered in the literature. Secondly, we extend the set cover problem to further express the conflict that meets with both operational period and stability. Thirdly, an evolutionary single objective and multi-objective algorithms are developed to tackle the formulated problem. Two heuristic operators are also introduced and injected within the framework of the evolutionary algorithms where the need arises to harness their strength in terms of both operational period and network stability. Performance evaluation is reported while different problem dimensions are experimented with in the simulations. The results show that the proposed multi-objective evolutionary algorithm is quite appropriate to converge to more accurate solutions than the counterpart single objective evolutionary algorithm. Further, the results give plausible evidence supporting the importance of the proposed heuristic operators to mitigate against the contradictory nature of the network lifetime in terms of operational period and stability.