A Novel, Low Computational Complexity, Parallel Swarm Algorithm for Application in Low-Energy Devices

Abstract

In this work, we propose a novel metaheuristic algorithm that evolved from a conventional particle swarm optimization (PSO) algorithm for application in miniaturized devices and systems that require low energy consumption. The modifications allowed us to substantially reduce the computational complexity of the PSO algorithm, translating to reduced energy consumption in hardware implementation. This is a paramount feature in the devices used, for example, in wireless sensor networks (WSNs) or wireless body area sensors (WBANs), in which particular devices have limited access to a power source. Various swarm algorithms are widely used in solving problems that require searching for an optimal solution, with simultaneous occurrence of a different number of sub-optimal solutions. This makes the hardware implementation worthy of consideration. However, hardware implementation of the conventional PSO algorithm is challenging task. One of the issues is an efficient implementation of the randomization function. In this work, we propose novel methods to work around this problem. In the proposed approach, we replaced the block responsible for generating random values using deterministic methods, which differentiate the trajectories of particular particles in the swarm. Comprehensive investigations in the software model of the modified algorithm have shown that its performance is comparable with or even surpasses the conventional PSO algorithm in a multitude of scenarios. The proposed algorithm was tested with numerous fitness functions to verify its flexibility and adaptiveness to different problems. The paper also presents the hardware implementation of the selected blocks that modify the algorithm. In particular, we focused on reducing the hardware complexity, achieving high-speed operation, while reducing energy consumption.

Multi-Sensor Fusion by CWT-PARAFAC-IPSO-SVM for Intelligent Mechanical Fault Diagnosis

“…Following a boost in two aspects of the above general particle swarm algorithm [ 16 ], the IPSO algorithm is constructed.…”

“…The big challenge in the conventional particle swarm optimization algorithm for application in miniatured devices and systems is the reduction of the energy consumption by the running hardware. The flexible and adaptive algorithm is proposed to reduce the structural complexity of hardware, obtain high-speed operation and reduce energy consumption [ 16 ]. The particle filter is used as the optimization method for the applications in intelligent detection systems and is normally combined with multidimensional data analysis [ 17 , 18 ].…”

An Investigation on Hybrid Particle Swarm Optimization Algorithms for Parameter Optimization of PV Cells

“…This directs researchers and scientists to further develop new algorithms or modify previous algorithms to resolve optimization problems. PSO and modified PSO algorithms are studied for various engineering applications e.g., speech emotion recognition, railway controls, job shop scheduling problems, geotechnical engineering, load flow control, source seeking problems, elevator door systems, quad assignment problems, equipment possession quantity, optimal designs of PID controllers, parameter estimation of photovoltaic cells or modules, and the beam-slab layout design of rectangular floors [40][41][42][43][44][45][46][47][48][49]. Hajihassani et al studied PSO algorithms for applications in geotechnical engineering [43].…”