rl-reward-circuit-mixed-optimization

Implements a reinforcement learning reward function for analog circuit design that handles mixed minimization/maximization objectives using normalized differences, while incorporating transistor saturation tracking to ensure stability.