A Dynamic Heuristic Set-based Particle Swarm Optimizer for the Multi-Objective Task Assigning Problem in Wireless Sensor Networks

The task assigning problem in wireless sensor networks is a challenging NP-hard combinatorial optimization problem that has caught much attention. Although binary particle swarm optimizers have been proven to be effective for this problem, the prematurity remains severe. To fix the prematurity, this paper proposes a set-based particle swarm optimizer with dynamic heuristic substitution techniques. Differing from the original binary particle swarm optimizer, the representation scheme is based on the concept of set. The position of each particle is a set whose elements are defined by the problem. A particle position element denotes the allocation of a task to a sensor node. The proposed algorithm tries to select an optimal subset from the universal set based on the particle swarm optimization framework. Moreover, the comprehensive learning strategy that utilizes beneficial information from the best-ever positions of all particles is employed, which helps keep the swarm away from local optima. Considering the constraints in real-world situations, the position updating adopts a substitution operation, guided by heuristic information that dynamically alters. To improve particle diversity, mutation operations are carried out after renewing the particle positions. Simulation results demonstrate the superiority of the proposed approach over others.