Forest Turtle Optimization Algorithm

Traditional swarm intelligence optimization algorithms generally aim to rapidly reduce the objective function value within a finite number of iterations, implicitly assuming that erroneous decisions during the search process are reversible. However, in real-world engineering problems characterized by high noise, strong non-convexity, or high evaluation costs, search behavior often carries the risk of irreversibility, and overly aggressive update strategies can lead the algorithm into an unrecoverable state of degradation. To address this, this paper proposes a forest turtle optimization algorithm based on long-term survival rationality. This algorithm no longer relies solely on the objective function value as the sole decision-making criterion but introduces a survival rationality function to jointly evaluate future gains, risk exposure, and irreversibility, thereby achieving conservative and stable optimization search in uncertain environments. Through slow-drift motion mechanics, a risk-aware memory mechanism, and a shell-shrinking information internalization strategy, the forest turtle optimization algorithm maintains long-term effective search capabilities in complex search spaces. Theoretical analysis shows that this method possesses stronger stability and robustness in high-risk optimization problems.