Optimization of native plant selection for phytomining of strategic/critical raw materials using Artificial Bee Colony algorithm

This study investigates the potential of native plant species (NPs) as secondary sources for strategic/critical raw materials (SCRMs) through phytomining, aiming to identify the most efficient accumulator for sustainable element recovery. To achieve high-precision results, the Artificial Bee Colony (ABC) algorithm, a robust meta-heuristic optimization technique inspired by the foraging behavior of honeybees, was utilized to minimize the error between experimental data and predicted models to optimize the accumulation of key elements including cobalt (Co), strontium (Sr), antimony (Sb), vanadium (V), phosphorus (P), lanthanum (La), magnesium (Mg), titanium (Ti), aluminum (Al), and scandium (Sc). The research focused on three NP species—Musk thistle (MT), curly dock (CD), and Juncus (JS)—collected from a geothermal hot spring area. Methodological validation was performed through comparative statistical analysis and the calculation of bioconcentration factor (BCF) and translocation factor (TF) to assess phytoremediation efficiency. The ABC optimization results revealed that MT exhibited the most significant accumulation capacity, particularly for Mg, reaching a peak concentration of 3,910 mg/kgdw in the leaves. The concentration range (lowest to highest value) was between 0.01 (for La) and 3,910 mg/kgdw for the studied plants. Principal component analysis (PCA) corroborated these findings, explaining cumulative variances of 80.4% (MT), 89.4% (CD), and 76.4% (JS). The innovation of this study lies in the application of the ABC algorithm to the field of phytomining, which offers a more dynamic and flexible alternative to traditional regression-based models. The primary advantage of this proposed method is its superior global search capability and ability to handle multi-elemental interactions with high predictive accuracy. However, a potential disadvantage remains the computational complexity and the requirement for fine-tuning control parameters to ensure consistent convergence. Overall, MT is identified as a highly promising candidate for the phytomining of SCRMs, with Mg emerging as the most effectively accumulated element under the optimized ABC model.