Systems, Man, and Cybernetics

 

Home

• [P505] Reinforcement Self-Organizing Interval Type-2 Fuzzy System with Ant Colony Optimization
  Author(s): Chia-Feng Juang, Chia Hung Hsu, Chia-Feng Chuang
  
  

  Abstract
  This paper proposes a Reinforcement Self-Organizing Interval Type-2 Fuzzy System with Ant Colony Optimization (RSOIT2FS-ACO) method. The antecedent part in each fuzzy rule of the RSOIT2FS-ACO uses interval type-2 fuzzy sets in order to improve system robustness to noise. There are no fuzzy rules initially. The RSOIT2FS-ACO generates all rules online. The consequent part of each fuzzy rule is designed using Ant Colony Optimization (ACO). The ACO approach selects the consequent part from a set of candidate actions according to ant pheromone trails. The RSOIT2FS-ACO method is applied to a truck backing control. The proposed RSOIT2FS-ACO is compared with other reinforcement fuzzy systems to verify its efficiency and effectiveness. A comparison with type-1 fuzzy systems verifies the robustness of using type-2 fuzzy systems to noise.

  
   

Full Text (PDF)