Modeling and Global Optimization of Biosurfactant Production from Bacteria Utilizing Frying Oil Waste Via Sequential Statistical and Crow Search Algorithm

Rhamnolipids are green surfactants and suitable alternatives to their chemical counterparts because of their biodegradability, nontoxicity, and environmental compatibility. This study investigated the modeling and global optimization of biosurfactant production from newly isolated Pseudomonas aeruginosa MYSAG using a sequential statistical and crow search algorithm (CSA). The effects of six variables: frying oil waste (FOW), glucose, NH₄Cl, urea, salt, and media pH, were evaluated first using the Plackett–Burman Design (PBD). It was found that FOW and urea had a higher impact than others. Hybridizing Central Composite Design (CCD) with CSA was employed for multi-objective global optimization. The optimal set of 17% FOW and 1% urea gave the maximum biosurfactant and biomass yields of 5.66 g/L and 2.81 g/L, respectively. The models were assessed via several performance indicators: R², Relative Error (RE), Mean Absolute Error (MAE), and Root Mean Square Error (RMSE). The R² values were > 82%, and all error values appeared small. The produced biosurfactant was characterized using oil displacement (OD), surface tension (ST), thin-layer chromatography (TLC), and Fourier transform infrared (FTIR) assays. The OD indicated that the biosurfactant was a growth-associated product. The measured ST value was less than 30 mN/m. The retention factors (R_f) in TLC for the mono- and di-rhamnolipids were calculated to be 0.88 and 0.17, respectively. The FTIR spectrum showed major peaks almost identical to those of pure rhamnolipid. The results were consistent with those of the literature. This article demonstrated the utilization of FOW for biosurfactant production with less cost and environmental hazards.