Hardware implementation of metaheuristics through LabVIEW FPGA

© 2021 Elsevier B.V.Metaheuristic optimization methods have been implemented for solving several problems. However, when it is required to implement those algorithms in hardware to run online, there is not enough information. This paper describes how could be programmed and implemented those optimization algorithms. Moreover, a complete evaluation is shown as well as a comparative study regarding the most important metaheuristic optimization algorithms. Thus, this paper presents a comparison between five optimization algorithms implemented into a cRIO field-programmable gate array (LabVIEW FPGA) NI-9030 of National InstrumentsTM(NI). The algorithms implemented were particle swarm optimization (PSO), bat algorithm (BA), grey wolf optimizer (GWO), earthquake algorithm (EA), and Nelder¿Mead algorithm (NM). To analyze hardware device utilization and execution time by each algorithm, synthesis results were presented. In addition, a set of ten benchmark functions was selected to compare performance between algorithms. Results show the feasibility of this approach for NI FPGA hardware. From device utilization results, GWO presents the lowest placed usage (29%) while NM shows the fastest execution time (0.683 ms). Nevertheless, PSO, GWO and EA show better performance between benchmark functions due their exploration characteristics which make possible to find a better solution.

Hardware implementation of metaheuristics through LabVIEW FPGA Academic Article in Scopus