This work proposes an optimal energy-efficient control strategy for a rigid articulated pendulum subjected to nonlinear dynamics, with the aim of inducing a complete rotation from \(+90^\circ\) to \(-90^\circ\) while minimizing the total mechanical effort applied. The effort is quantified as the absolute sum of the motor torques acting on the system. Temporal dynamics are simulated using the finite element model in ANSYS Mechanical; the nonlinear effects are, for accuracy, captured by the HHT (Hilber-Hughes-Taylor) implicit integration scheme. In summary, the optimization problem represents finding a discrete sequence of motor torques. The solution to the problem is implemented on the OptiSLang platform by comparing two approaches: a direct, deterministic method (Downhill Simplex) and an evolutionary method (EA). Results indicate that the Simplex method leads to more energy-efficient minimization of the overall effort, whereas the evolutionary algorithm produces a far smoother trajectory and resultantly closer to optimum, although at a slightly higher effort. Sensitivity analysis was performed to identify the most influential control torques, which contributed to the reduction of the problem’s complexity. This paper illustrates how numerical simulation, metamodeling, and advanced optimization are effective in creating robust control laws within dynamic systems. Applications may range over subjects such as robotics, embedded mechatronics, and intelligent energy-efficient structure control.
