The optical switch based on the dimer plasmonic nano-tubes on the silicon waveguide has been numerically analysed. In the proposed switch, the optical switch has been excited by two monochromatic incident plan-waves with the same frequency and two polar angles “θ = 0” and “θ = π/2”. If the signal with θ = 0 applies, the incident wave will be transmitted, and on the other hand coherent perfect absorption (CPA) will occur and the two incident waves will be suppressed, if we apply both signals to the switch at the same time. Therefore, the signal with θ = π/2 acts as control signal. Since the CPA efficiency depends strongly on the number of plasmonic nano-tubes and its location, a new efficient binary optimization method based on the Teaching–Learning-Based Optimization (TLBO) algorithm is proposed to design an optimized array of the plasmonic nano-tubes in order to achieve the maximum absorption coefficient in the ‘off’ state and the minimum absorption coefficient in the ‘on’ state. In Binary TLBO, a group of learners consists of a matrix with binary entries, control the presence (‘1’) or the absence (‘0’) of nano tubes in the array.