PSO-based Hybrid Algorithm for Multi-objective TDMA Scheduling in Wireless Sensor Networks

Review of nature-inspired methods for wake-up scheduling in wireless sensor networks

“…Examples of such hybrid approaches include the use of global pheromone updates in otherwise local ACO [82,85], combination of PSO with simulated annealing [87] or Q-learning [91], and the use of domain-specific operators in GA [100].…”

“…The two major groups of nature-inspired methods utilized to address wake-up scheduling are swarm intelligence and evolutionary computing. The algorithms based on swarm intelligence include methods inspired by Ant Colony Optimization [57,[79][80][81][82][83][84][85][86], Particle Swarm Optimization [58,64,83,[87][88][89][90][91], and Pulse-coupled Biological Oscillators [19,59,92,93]. Evolutionary methods for WSN scheduling are mostly based on Genetic Algorithms [94][95][96][97][98][99][100][101] but other evolutionary approaches have been considered as well [102].…”

“…Simulation experiments show that the hybrid natureinspired meta-heuristic outperforms two methods based on graph colouring and a hybrid between PSO and GA. Another PSO-based TDMA scheduling method has been proposed in [88]. In comparison to [87], the new approach can accommodate conflicting optimization objectives. The PSO finds a set of Pareto-optimal schedules with respect to average energy consumption and average packet delay, which are presented to network designer to support decisionmaking and network control.…”

“…TDMA scheduling has been modelled as a multi-objective optimization problem and solved using a hybrid PSO-based algorithm in [87]. The proposed approach aims at saving energy by eliminating transmission collisions.…”

“…The main reasons for using bio-inspired methods to solve wake-up scheduling problems have been their optimization capability [143,80,[87][88][89][90]94,20,14,17,105,95,129,104,98,102,100,84,79,58,74,57] that facilitates the discovery of efficient wake-up schedules; distributed nature [113,18,92,80,[87][88][89]94,14,[103][104][105]136,138,17,66,69,74,78,93,57] that well matches the topology of WSNs and allows optimization based on readily available local information; and various self-n properties, that endow the networks with greater autonomy; they include ○ self-organization [ ○ self-assembly (of services and schedules) [138,14], ○ self-healing …”