Performance Optimization of Two Stage OTA Using Novel Modified Teaching Learning Based Optimization Algorithm

The Evolutionary algorithms have become an essential tool in computer-aided design for addressing complex optimization problems. However, their application in analog VLSI design is often constrained by high computational cost, large memory demand, and the necessity for fine tuning of parameters. To address these limitations, this work presents the design optimization of a two-stage operational transconductance amplifier (OTA) using PTM 45 nm CMOS technology. The Modified Teaching Learning Based Optimization (MTLBO) algorithm is employed, which eliminates the need for additional control parameters by mimicking the natural learning interaction between teachers and learners. The algorithm was implemented in Python and executed on an AMD Ryzen™ processor with 16 GB RAM under Ubuntu OS. Simulation results show that the optimized two stage OTA achieves the of 83.17 dB, higher unity gain bandwidth of 1.71 MHz with least power consumption of 3.24 µW. Comparative analysis demonstrates that MTLBO provides faster convergence with fewer iterations and outperforms other metaheuristic approaches, establishing it as a strong candidate for efficient, low power, and high performance analog CMOS circuit design automation. Future work may include extending the methodology to more complex analog and mixed-signal circuits, exploring multi-objective optimization.