The Fungal Growth Optimizer (FGO) is a nature-inspired metaheuristic that simulates fungal colony behaviors, but its exploitation phase can lose search diversity when guidance is dominated by limited peer or global-best information. In this paper, we propose an enhanced variant called the Comprehensive Learning Fungal Growth Optimizer (CLFGO). We integrate a conditionally activated Comprehensive Learning (CL) strategy into the FGO framework. When a candidate solution stagnates, the strategy constructs a dimension-specific learning exemplar. This mechanism allows each dimension to learn from the personal best of a different peer, extending the original fungal growth model. CLFGO is therefore intended for high-dimensional, multimodal, hybrid, and composition landscapes in which the original FGO is prone to diversity loss, rather than as a universal replacement for all problem classes. This approach improves population diversity and reduces the risk of premature convergence. We evaluated CLFGO on 29 CEC2017 benchmark functions at 30 dimensions against nine metaheuristics under the same maximum-function-evaluation budget. CLFGO achieved the lowest mean error on 21 of 29 functions and attained a Friedman average rank of 1.5517. Furthermore, we applied CLFGO to a reservoir production optimization problem, where it obtained a mean Net Present Value of 9.97×108 USD, outperforming the compared algorithms in both solution accuracy and convergence stability.
Gong et al. (2026) studied this question.