Slime Mould Reproduction: A New Optimization Algorithm for Constrained Engineering Problems

Abstract

In recent explorations of biologically inspired optimization strategies, the Slime Mould Reproduction (SMR) algorithm emerges as an innovative meta-heuristic optimization technique. This algorithm is deeply rooted in the reproductive dynamics observed in slime molds, particularly the intricate balance these organisms strike between local and global spore dispersal. By replicating this balance, the SMR algorithm deftly navigates between exploration and exploitation phases, aiming to pinpoint optimal solutions across diverse problem domains. For the purpose of evaluation, the SMR algorithm was diligently tested on three engineering problems with inherent constraints: Gear train design, three-bar truss design, and welded beam design. A comprehensive comparative study indicated that the SMR algorithm outperformed esteemed optimization techniques such as Particle Swarm Optimization (PSO), Artificial Bee Colony (ABC), Differential Evolution (DE), Grasshopper Optimization Algorithm (GOA), and Whale Optimization Algorithm (WOA) in these domains. While the exemplary performance of the SMR algorithm is worth noting, it is essential, in line with the No Free Lunch (NFL) theorem, to underscore that the performance of any optimization algorithm invariably depends on the particular problem it addresses. Nevertheless, the SMR algorithm's consistent triumph in benchmark tests underscores its potential as a formidable contender in the vast realm of optimization algorithms. The current exploration not only emphasizes the ever-expanding horizon of bio-inspired algorithms but also positions the SMR algorithm as a pivotal addition to the arsenal of optimization tools. Future implications and the potential scope of the SMR algorithm extend to various domains, from computational biology to intricate industrial designs. Envisioning its broader applicability, upcoming research avenues may delve into refining SMR's core procedures, borrowing insights from a broader range of biological behaviors for algorithmic ideation, and contemplating a binary version of the SMR algorithm, thereby amplifying its versatility in diverse optimization landscapes.