The improved slime mould algorithm with exponential functions

Applying the Slime Mould Algorithm (SMA) in optimization and coding it with Python programming language, we found that runtime warnings occurred because of the inherent limits of arctanh functions. Considering the circumstances that the maximum residual errors would be constrained and the maximum allowed iteration number would be bothering, we therefore proposed an improvement to the SMA with exponential functions. Three kinds of benchmark functions were introduced to verify the suitability, accessibility, and capability. Experiment results demonstrated that the improved SMA with exponential functions would be suitable and accessible. Comparing with the standard SMA, the improved algorithm would result in better performance with exception that they perform similar when optimizing the Quintic function, the representative of those kinds of benchmark functions with basins in their three-dimensional profiles.