An efficient particle swarm optimization with evolutionary multitasking for stochastic area coverage of heterogeneous sensors

This paper investigates the stochastic area coverage problem of sensors with uncertain detection probability. The risk associated with uncertain parameters is managed using the conditional value-at-risk (CVaR) risk measure. The loss function is represented by the uncovered area coverage rate. We then formulate the minimum CVaR-based uncovered area coverage (CVaR-UAC) problem and provide some theoretical guarantees for the problem. Unlike previous research that treats area coverage as a single problem, we propose an efficient particle swarm optimization (PSO) with evolutionary multitasking to solve the stochastic area coverage problem along with multiple simplified problem forms. These simplified problems act as the auxiliary tasks for the original CVaR-UAC to enhance the evolutionary search. We have improved the proposed PSO algorithm from the framework of disturbance PSO and virtual force directed co-evolutionary particle swarm optimization, using a hybrid method in population initialization and an adaptive perturbation in individual updating. As a result, the exploration ability of the algorithm is significantly enhanced. The experiment results have demonstrated the effectiveness of the proposed algorithm compared with state-of-the-art algorithms in terms of solution quality.