Title: A new metaheuristic lung-inspired algorithm for continuous optimization problems
Authors: Mustafa Kadhim Taqi and Mueen Mohsin Abbood
Volume: 9
Issue: 2
Pages: 7-19
Publication Date: 2025/02/28
Abstract:
This paper introduces a metaheuristic algorithm called lung-inspired algorithms (LA). LA is mainly inspired by the breath cycle of the human lung. Three phases namely inhalation, gas exchange, and exhalation are conducted during this cycle, to maximize the saturation of oxygen in the blood as carbon dioxide is removed. This algorithm begins by random initial solutions. The saturation rate would then be determined depending on the objective functions average valuation of all proposals within existing population at each iteration. The inhalation stage explores positions opposite to low saturation solutions in order to identify promising places. However, once they exceed the rate of saturation, solutions are reconsidered. In comparison, the gas exchange process takes advantage of movement operators dependent on the oxygen flux law. Then, by simulating lung exhalation function, weak solutions are eliminated from the population. The best is identified and the saturation rate adjusted when all alternatives were placed in the population. The algorithm is assessed by using twenty-three identified benchmark datasets and corresponding output to many other metaheuristics. This involves the most well-studied metaheuristics GA and PSO. In addition, the newly-designed metaheuristics that includes BAT, GSA, WOA, MPA, and the GWO have also been involved. In almost all the test functions, the output of LA is outperformed.