Application of artificial neural network coupling multiobjective particle swarm optimization algorithm to optimize Pleurotus ostreatus extraction parameters

Abstract

In this research, extraction parameters were optimized for the maximization of the antioxidants of mushrooms. Firstly, the total antioxidant status (TAS) and total oxidant status (TOS) of mushrooms under different extraction parameters were determined. Then, artificial neural network (ANN) and multiobjective particle swarm optimization (MOPSO) were applied for modeling and optimization of the extraction process, respectively. The three variables affecting this process were extraction temperature (40 to 70 degrees C), extraction time (4 to 10 h), and extraction concentration (0.25 to 2 mg/ml). The developed ANN-based model could predict with a correlation coefficient of 0.986 and 0.991 for TAS and TOS values, respectively. The optimum extraction parameters were extraction temperature of 40.721 degrees C, extraction time of 6.267 h, and extraction concentration of 2 mg/ml. These results indicated that the ANN-MOPSO hybrid model provides an accurate prediction and optimization method for TAS and TOS values of Pleurotus ostreatus extracts. Novelty impact statement This research has three stages: experimental study, modeling, and optimization. Artificial neural network (ANN) was used for modeling stage and multiobjective particle swarm optimization algorithm (MOPSO) was used for the optimization process. It is the first time two methodologies have been combined for Pleurotus ostreatus extraction optimization. ANN-MOPSO hybrid model can be used to optimize mushroom extraction to save time, labor, and money.