The aim of this paper is to develop a productive numerical technique to deal with a class of time partial ideal AI control issues. The classical fuzzy inference methods cannot work to their full potential in such circumstances, because the given knowledge does not cover the entire problem domain. In addition, the requirements of fuzzy systems may change over time. The use of a static rule base may affect the effectiveness of fuzzy rule interpolation due to the absence of the most concurrent (dynamic) rules. The experimental result indicates that evolved bat algorithm with our proposed fitness function presents a 93.77% success rate in average for finding the feasible solutions. The contribution of this study is that near outcomes likewise confirm that the partial administrator for a Mittag–Leffler circuit in the Caputo sense improves the execution of the AI controlled framework as far as the transient reaction, in contrast with the other fragmentary and whole number subordinate administrators.

本文的目的是开发一种有效的数值技术来处理一类时间局部理想AI控制问题。在这种情况下，经典的模糊推理方法无法充分发挥其潜力，因为给定的知识无法涵盖整个问题领域。另外，模糊系统的要求可能会随时间变化。静态规则库的使用可能会由于缺少大多数并发（动态）规则而影响模糊规则插值的有效性。实验结果表明，采用本文提出的适应度函数的进化蝙蝠算法平均能找到93.77％的成功率。