Optimization for Interval Type-2 Polynomial Fuzzy Systems: A Deep Reinforcement Learning Approach

It is known that the interval type-2 (IT2) fuzzy controllers are superior compared to their type-1 counterparts in terms of robustness, flexibility, etc. However, how conduct type reduction optimally with consideration system stability under fuzzy-model-based (FMB) control framework still an open problem. To address this issue, we present a new approach through membership-function-dependent (MFD) and deep reinforcement learning (DRL) approaches. In proposed approach, IT2 membership functions controller completing during optimizing performance. Another fundamental issue conditions must hold subject different type-reduction methods. tedious impractical resolve according methods, which could lead infinite possibility. more practical guarantee holding rather than resolving conditions, MFD imperfect premise matching (IPM) concept. Thanks unique merit all embedded within footprint uncertainty (FOU) guaranteed be valid. During processes, state transitions associated properly engineered cost/reward function can used approximately calculate deterministic policy gradient optimize acting then improve performance determining grade controller. The detailed simulation example provided verify merits approach.