Green Mamba Optimization Algorithm

In recent years, swarm intelligence optimization algorithms have been widely applied in continuous optimization problems. However, most existing methods are based on the implicit assumptions of "reversible search" and "fixed search space," meaning that individual algorithms can move repeatedly throughout the search space, and their search behavior is mainly adjusted through random perturbations or parameter adaptation. These assumptions often lead to low search efficiency, severe oscillations, and over-reliance on globally guided solutions in high-dimensional, strongly correlated, energy-constrained, or black-box optimization problems. To overcome these limitations, this paper proposes a novel continuous optimization framework—Green Mamba Optimization (GMO). This method no longer relies on traditional biomimetic metaphors but re-characterizes the optimization process from the perspectives of irreversible dynamical systems, energy-constrained reachability, and self-induced optimization geometry. This paper is the first to systematically introduce irreversible search operators into swarm optimization, propose the minimum reachability principle, and construct a low-dimensional search geometry induced by the optimization process itself. Theoretical analysis shows that this method effectively suppresses invalid oscillations and accelerates convergence by progressively compressing the search degrees of freedom and restricting the reachable solution set. Experimental results show that this method has significant advantages in high-dimensional continuous function and complex structure optimization problems.