Bio-inspired optimization techniques, such as Particle Swarm Optimization (PSO), Cuckoo Search (CS), and Roach Infestation Optimization (RIO), have been extensively explored to find solutions of complex optimization problems that would be computationally expensive with traditional approaches. In particular, RIO has been widely used due to its social interaction mechanisms, which promote exploration through collective behaviors inspired by cockroach infestation dynamics. Despite these advantages, the effectiveness of RIO, as with many metaheuristic approaches, depends heavily on the appropriate tuning of its control parameters, a challenge that becomes more critical when attending with large scale optimization problems. This work focuses on improving the performance of RIO by integrating a Generalized Type-2 Fuzzy Logic System (GT2FLS) for dynamic parameter tuning. Unlike conventional approaches that rely on fixed parameter configurations or Type-1 Fuzzy Systems, the proposed method explicitly considers uncertainty in the adaptation process using Generalized Type-2 fuzzy reasoning. GT2FLS continuously adjusts the RIO parameters based on indicators, such as diversity and the current iteration number, both observed during the search process. Experiments for 500, 1000, and 2000 dimensions were conducted using benchmark mathematical functions, and these showed improved convergence stability and competitive solutions with reliable outcomes compared to RIO and other known optimization algorithms, such as PSO and CS. The results confirm that using GT2FLS in the parameter adaptation mechanism significantly improves robustness and scalability, making the proposed approach a promising alternative for large-scale optimization tasks.
Lizárraga et al. (Sat,) studied this question.