A novel bio-inspired algorithm based on plant root growth model for data clustering

Inspired by the behaviors of plant root growth, an Artificial Root Mass (ARM) optimization algorithm based on an artificial root model is proposed. This ARM algorithm simulates the plant's root growth strategies including proliferation and `intelligent' decisions about growth directions. Ten well-known benchmark functions are employed to validate its optimization effect. ARM is compared with other existing algorithms, including genetic algorithm (GA), particle swarm optimization (PSO) and differential evolution (DE). The experimental results show that ARM seems much superior to other algorithms on the selected benchmark functions in multidimensional cases. ARM algorithm is used for data clustering on several benchmark datasets. The performance of the ARM algorithm is compared with GA, PSO and DE on clustering problems. The simulation results show that the proposed ARM outperforms the other three algorithms in terms of accuracy and robustness on most of selected datasets. The proposed algorithm ARM provides a new reference for solving data clustering problems.