Development and Multi-Objective Optimization of Cosmetic Cream Formulations Using Artificial Neural Networks and Genetic Algorithms

https://doi.org/10.62239/jca.2025.054

Abstract

In recent years, demand for herbal cosmetics has grown due to their safety, efficacy, and eco-friendliness. Yet, herbal skincare formulations often lack standardization, leading to variable quality. This study focuses on developing and optimizing a herbal moisturizing cream using artificial intelligence (AI) in formulation design. Fourteen natural ingredients, including aloe vera, shea butter, almond oil, hyaluronic acid, and essential oils, were combined into 30 experimental formulations through a Simplex Lattice mixture design. Key properties—pH, viscosity, permeation, and moisturizing ability—were measured and modelled using a fuzzy logic neural network integrated with a genetic algorithm. The optimal formula was then evaluated for texture, absorption, spread ability, and stability. Integrating I(14)-HL(2)-O(4) with genetic algorithms highlights an effective approach for standardizing and optimizing herbal cosmetic formulations, offering a scientific basis for creating high-quality, consistent skincare products.