Research efforts in neuromuscular therapy and assistive devices have been in constant progress in the last 20 years. One of the most sought-after assistive devices includes the rehabilitation of upper limb exoskeletons. This paper presents the implementation of an optimization-based optimal controller technique for the "SAMA" three-DOF active upper limb exoskeleton for rehabilitation. Proportional-integral-derivative computed torque control (PID-CTC) is the basis of the proposed controller. The hybrid genetic pattern search method is then used to optimize the ideal controller settings (HGPSA). Using HGPSA, four objective functions are selected to find the best control values. The integral absolute error (IAE), integral time absolute error (ITAE), integral square error (ISE), and integral time square error (ITSE) are these functions. The obtained results demonstrate that the optimal dynamic responses are optimized by an ideal CTC based on HGPSA, with a highest steady-state error of 0.18° and 0% overshoot. Additionally, a reaction safe interaction strategy is added to the robot model to provide compliance during trajectory planning.